Deciphering the Heterogeneity of Pulmonary Macrophages in Response to Fine Particles

ABSTRACT

The heterogeneity of pulmonary cells affects their response to inhaled fine particles. However, the most critical responder cells in particle exposure still remain controversial, and their functional heterogeneity warrants detailed exploration. Herein, a gold nanoparticle model dual‐labeled with sulfo‐cyanine3 and Tag RNAs is developed, aiming to capture prominent fine particle–responsive cell subpopulations using fluorescence‐activated cell sorting and single‐cell RNA sequencing. Alveolar macrophages (AMs), recruited macrophages (recMacs), and interstitial macrophages (IMs) exhibited the strongest responsiveness to particles, among which 14 subsets are identified with partially overlapping yet distinct functions. Notably, AMs_3, AMs_7, and recMacs_4 subsets are highlighted in the particle–responsive single‐cell atlas, as evinced from their functional scoring and Tag RNA levels. AMs_3 and AMs_7 showed function enrichment for acute‐phase responses to inhaled particles, while recMacs_4 exhibited enrichment for cell chemotaxis and phagocytosis. Fine particle engulfment led to enhanced outgoing interactions between these three macrophage subsets and other cell populations within particle–responsive cellular communication networks. This study reveals that specific macrophage subpopulations act as the primary responsive cell subpopulations in engulfing inhaled particles and link their transcriptional heterogeneity to functional diversity, thereby opening a new avenue to explore the heterogeneity of immune cells in pulmonary disorders.

This study introduces a dual‐labeled gold nanoparticle model to capture particle–responsive cells via integrating fluorescence‐activated cell sorting and single‐cell RNA sequencing. Fourteen macrophage subsets are identified with response and function heterogeneity, among which AMs_3, AMs_7, and recMacs_4 subsets show stronger phagocytic activation against fine particles. These specific macrophage subsets trigger extensive intercellular communication with other pulmonary cell populations.

Abbreviations

AMs

alveolar macrophages

AT1

type 1 alveolar epithelial cells

AT2

type 2 alveolar epithelial cells

AuNPs

gold nanoparticles

BMDCs

bone marrow‐derived dendritic cells

BMDMs

bone marrow‐derived macrophages

CXCL

C‐X‐C motif chemokine ligand

Cy3

sulfo‐cyanine3

CyTOF

cytometry by time ‐ of ‐ flight

DCs

dendritic cells

DEGs

differentially expressed genes

EDS

energy‐dispersive X‐ray spectroscopy

FACS

fluorescence‐activated cell sorting

GFP

green fluorescent protein

GO

gene ontology

GTF

gene transfer format

IL‐1

interleukin‐1

IMs

interstitial macrophages

LbL

layer‐by‐layer

LbL‐AuNP

layer‐by‐layer assembled gold nanoparticle

MFI

mean fluorescence intensity

MIF

macrophage migration inhibitory factor

Mtb

Mycobacterium tuberculosis

MUA: 11

mercaptoundecanoic acid

MuV

mumps virus

NK

natural killer

PBS

phosphate‐buffered saline

PD‐L1

programmed death‐ligand 1

PEI

poly(ethylene imine)

PRR

pattern recognition receptor

RBCs

red blood cells

recMacs

recruited macrophages

scRNA‐seq

single‐cell RNA sequencing

TEM

transmission electron microscopy

UMAP

uniform manifold approximation and projection

1. Introduction

The lung is not only the primary target organ for inhaled fine particles but also an important site for the proliferation, recruitment and activation of inflammatory cells [1, 2]. The inhaled particles may interact with different cell types, such as alveolar macrophages (AMs) and neutrophils, prompting the recruitment of other immune cells and leading to the remodeling of immune niches in the lung [3, 4, 5]. The interaction between exogenous particles and immune cells triggers cellular phagocytosis, which is significantly influenced by the physicochemical properties of the particles and the responsive cell types [6, 7]. However, most existing studies delineate this interaction based on the overall cell population responses [8, 9, 10, 11, 12]. Although these bulk‐cell‐based detection methods are simple, fast, and cost‐effective, they do not account for cell‐to‐cell heterogeneity, which can lead to bias or misleading results [13]. Cell‐to‐cell heterogeneity arises from intrinsic diversity within cell populations, and is also affected by differences in cellular exposure outcomes, where only a small fraction of cells phagocytose exogenous particles while most act as bystanders (exposed without engulfment). Therefore, understanding the heterogeneity of cellular reactions and functions is necessary to accurately identify target cells and develop precision therapies that optimize therapeutic outcomes for various lung diseases.

Various immune cell types in the lung participate in the phagocytosis and clearance of exogenous particles [3], each displaying distinct fates and functions during particle exposure [14]. Taking pulmonary macrophages as an example, single‐cell transcriptomics has transformed the traditional view that interstitial macrophages (IMs) are a homogeneous population, the heterogeneity of which is evident in both developmental origins and extends to dynamic processes such as differentiation and reprogramming [15]. During immune quiescence or activation, IMs can be categorized into 10 functional subsets based on their chemokine expression patterns, contributing to the formation of a microenvironmental regulatory network [15]. Among these identified macrophage subsets, Ly6G⁺ macrophages can infiltrate the alveolar spaces surrounding perilesional areas in the early stage of restoration after influenza A (H1N1) virus infection [16]. These cells exhibit unique metabolic, phagocytic and efferocytic abilities and participate in type 2 alveolar epithelial cell (AT2)‐dependent alveolar regeneration [16]. Therefore, immune cell subpopulations are now defined by their specific functions and influenced by dynamic microenvironments and subset‐specific molecular signatures. Although single‐cell RNA sequencing (scRNA‐seq) facilitates the in‐depth investigation of cellular heterogeneity in physiological regulation and pathological responses [17], the monitoring of particle engulfment at single‐cell resolution still presents technical challenges, particularly in the absence of unique identifiers for these inhaled particles. A recent scRNA‐seq study dissected pulmonary cell heterogeneity in influenza infection by distinguishing infected and bystander cells through viral nucleic acid detection [18], which brings inspiration that Tag RNAs could be used as particle‐recognizable labels.

This study is motivated by the central hypothesis that the lung harbors transcriptionally distinct sentinel cell subpopulations, which are highly efficient in particle engulfment but remain unidentified by conventional typing methods due to their rarity or unknown characteristics. Our objective was to identify the primary fine particle‐responsive cell populations in the lung and examine their decisive transcriptional atlas in relation to cellular heterogeneity. To this end, we constructed a sulfo‐cyanine3 (Cy3)‐labeled layer‐by‐layer assembled gold nanoparticle (LbL‐AuNP) model loaded with poly(A) tail‐modified Tag RNAs to achieve dual labeling for monitoring particle phagocytosis. Subsequently, we combined fluorescence‐activated cell sorting (FACS) with scRNA‐seq to gain a more holistic view of cellular responses. Notably, the fundamental advantage of our methodology over existing techniques is its capacity to concurrently link “phagocytic activity” with the “full transcriptomic landscape” within individual cells. Here, our data unveiled that macrophages (AMs, IMs, and recruited macrophages [recMacs]) exhibited the highest responsiveness to fine particles. Functional scoring and Tag RNA detection further revealed three macrophage subsets (AMs_3, AMs_7 and recMacs_4) that showed the highest engulfment and activation response to exogenous particles. AMs_3&7 subsets, characterized by high Cd274 expression, manifested by upregulation of phagocytosis‐related genes and enrichment of pathway genes in cellular response to environmental stimulus and chemotaxis. Cell–cell communication analysis indicated that these three subsets displayed enhanced fine particle‐responsive capacity via C‐X‐C motif chemokine ligand (CXCL) and macrophage migration inhibitory factor (MIF) signaling pathway networks. Overall, our findings demonstrate the heterogeneity of pulmonary macrophage subpopulations in their response to fine particles, providing novel insights into inhalable particle–induced lung diseases and potentially enhancing clinical practices for inhalable drugs.

2. Results

2.1. LbL‐AuNPs Promote Obvious Immune Responses in the Lung

To characterize the heterogeneous responses of pulmonary cell populations to inhaled fine particles, we synthesized a particle model utilizing gold nanoparticles (AuNPs) as the core, which we fabricated through a layer ‐ by ‐ layer (LbL) assembly technique, following established procedures from previous research [19]. Figure 1A depicts the detailed preparation process of LbL‐AuNPs. Gold core size distribution was determined to be approximately 14.5 nm using transmission electron microscopy (TEM; Figure S1A). An increase in particle hydrodynamic diameter from 21.7 to 220.6 nm during the coating process was found using dynamic light scattering (DLS; Figure S1B). Charge reversal in zeta potential indicated the successful deposition of poly(ethylene imine) (PEI) and Tag RNAs on the surface of 11‐mercaptoundecanoic acid (MUA)‐AuNPs (Figure S1C). TEM observation revealed the formation of monodisperse particles, which further confirmed their morphology during the LbL coating process (Figure S1D). Additionally, we evaluated the Tag RNA loading efficiency (Figure S1E) and the particle dispersity (Figure S1F) across various weight ratios of carrier (Cy3‐PEI‐AuNP) to Tag RNA. Based on these evaluations, we selected a weight ratio of 5:1 (carrier to Tag RNA) for Tag RNA loading. Together, these data suggested that the prepared LbL‐AuNPs maintained a monodisperse state, with uniform size and no aggregation.

FIGURE 1.

Schematic illustration of design of model particle, cellular uptake analysis and single‐cell transcriptome sequencing of mouse lung tissue following fine particle treatment. (A) Flowchart illustrating construction of fluorescent layer‐by‐layer assembled gold nanoparticles (LbL‐AuNPs, upper panel) and cellular uptake analysis based on intracellular Au content, Tag RNA content and fluorescence intensity (lower panel). (B) Workflow of single‐cell RNA sequencing (scRNA‐seq) to characterize particle–induced cellular response heterogeneity in mouse lung tissue. Lung tissues were collected from mice treated with PBS (blank control) or LbL‐AuNPs, and dissociated into single‐cell suspensions. After sorting by fluorescent activated cell sorting, single‐cell suspensions were divided into three groups: blank control group (n = 3), Cy3‐negative group (n = 3) and Cy3‐positive group (n = 3). These groups were then subjected to scRNA‐seq and further analysis.

Phagocytosis is a key cellular process for recognizing and removing exogenous particles, subsequently triggering appropriate immune responses [7, 20, 21]. Next, we conducted a time‐course analysis of LbL‐AuNP uptake at the cellular level. We assessed the mean fluorescence intensity (MFI), Au content and Tag RNA content in bone marrow‐derived macrophages (BMDMs), bone marrow‐derived dendritic cells (BMDCs), J774A.1 and RAW 264.7 cells upon treatment with LbL‐AuNPs at different time points. Taking BMDMs as an example, MFI at 6, 12, 18 and 24 h after particle treatment was 1.8‐, 3.1‐, 3.6‐, and 4.2‐fold higher than that at 2 h, respectively (Figure S2A, p < 0.001 for all). Similarly, the Au content significantly increased over time, showing 2.0‐, 3.0‐, 3.7‐ and 5.4‐fold higher levels than that at 2 h (Figure S2B; p < 0.05, p < 0.001, p < 0.001 and p < 0.001, respectively). Furthermore, Tag RNA content followed the same trend, with the values at 6, 12, 18, and 24 h being 2.4‐, 5.0‐, 6.5‐, and 7.7‐fold higher than that at 2 h, respectively (Figure S2C, p < 0.001 for all). Similar patterns of increased MFI, Au content, and RNA content over time were observed in the other cells tested, including BMDCs, J774A.1, and RAW 264.7 cells, substantiating the time‐dependent uptake of fine particles by cells (Figure S2A–C). Furthermore, correlation matrix analysis revealed a strong positive correlation between treatment time and MFI, Tag RNA content, and Au content in BMDMs, BMDCs, J774A.1, and RAW 264.7 cells (Figure S2D, Kendall's tau > 0.8 for all). Notably, a strong positive correlation was observed between Au content and both MFI and Tag RNA content (Figure S2D, Kendall's tau > 0.8 for all). These findings indicate that both Tag RNA content and MFI as reliable indicators for assessing the overall internalization behavior of LbL‐AuNPs.

We conducted in vivo experiments to validate lung immune responses following treatment with LbL‐AuNPs over different time periods. Hematoxylin and eosin staining indicated alveolar collapse accompanied by obvious infiltration of inflammatory cells at 6 h post‐treatment, with greater phenotypes at 12 h post‐treatment (Figure S3A). Meanwhile, particle aggregates were identified in lung tissue sections, as indicated by red arrows (Figure S3A). Further examination of the fluorescence distribution in the lung revealed that LbL‐AuNPs were predominantly localized in the alveoli or around the alveolar walls at 2 h post‐treatment (Figure S3B). Notably, increased fluorescence signals were observed in the lung interstitium at 6 and 12 h post‐treatment (Figure S3B). Consistent with LbL‐AuNP uptake at the cellular level, TEM analysis and energy‐dispersive X‐ray spectroscopy (EDS) analysis of pulmonary cell sections further confirmed the phagocytized particles (Figure S4A,B). These findings supported the data obtained from histological and fluorescence analyses. In addition, flow cytometry analysis revealed that the proportion of Cy3⁺ cells in neutrophils increased from 5.7% at 2 h to 18.2% at 12 h (p < 0.001), while in macrophages, the proportion increased to 31.8% at only 2 h and peaked at 36.6% at 6 h (Figures S5 and S6). Collectively, these results reveal that LbL‐AuNPs interact with pulmonary immune cells, considerably modulating cellular immune responses, indicating their potential to influence shaping the pulmonary immune environment.

2.2. Single‐Cell Landscape of Pulmonary Cell Populations in Fine Particle–Treated Mice

To investigate the differential reaction of lung cell populations to fine particles, we combined scRNA‐seq with LbL‐AuNPs to assess transcriptional changes and Tag RNA detection levels at the single‐cell level within lung tissue samples. ScRNA‐seq is an increasingly recognized method for sorting out the heterogeneity of immune responses in pulmonary cell populations [22, 23]. Importantly, exogenous Tag RNAs loaded on designed LbL‐AuNPs were modified with a poly(A) tail, enabling their capture during 3ʹ end scRNA‐seq [24]. This method enables the detection of endogenous transcriptomes and the identification and quantification of poly(A)‐tailed exogenous Tag RNAs, facilitating the evaluation of particle internalization during sequencing.

This study included the following three groups based on the detection results of LbL‐AuNP fluorescence signals: (1) the blank control group, containing cells derived from the lung of phosphate‐buffered saline (PBS)–treated mice; (2) the Cy3‐negative group, containing cells obtained from the lung of fine particle–treated mice without internalizing the particles; and (3) the Cy3‐positive group, containing cells collected from the lung of particle–treated mice with internalized particles. Furthermore, scRNA‐seq was performed on these groups to construct a single‐cell transcriptional atlas of lung cell subpopulations with LbL‐AuNP treatment (Figure 1B). Thus, we aimed to gain insights into the heterogeneous immune responses in the lung by distinguishing and analyzing pulmonary cell populations under various treatment conditions, including untreated cells, particle‐negative cells and particle‐positive cells.

A total of 72,558 cells were analyzed, of which 23,080 were extracted from the blank control group, 24,566 were derived from the Cy3‐negative group, and 24,912 were obtained from the Cy3‐positive group. We measured the expression of canonical markers obtained from previous studies [25, 26, 27] and the CellMarker database. Therefore, we classified cell clusters into neutrophils, AMs, IMs, recMacs, dendritic cells (DCs), plasma cells, B cells, T cells, type 1 alveolar epithelial cells (AT1), AT2, natural killer (NK) cells, fibroblasts and endothelial cells (Figure 2A,B). Although the 13 cell clusters identified were present in all the three groups, a notable difference was noted in the count and percentage of prominent cell types between the Cy3‐negative group and the Cy3‐positive group (Figure 2C,D). Figure 2D illustrated the uniform manifold approximation and projection (UMAP) plots of various cell clusters across different groups. In detail, the Cy3‐negative group predominantly comprised T cells (50.5%) and B cells (13.5%), while the Cy3‐positive group was characterized by a higher proportion of macrophages (27.2%, comprising AMs, IMs, and recMacs) and neutrophils (33.5%; Figure 2C). Consistent with the scRNA‐seq results, flow cytometry analysis revealed that various lung cell populations (including macrophages, neutrophils, DCs, T cells, B cells, epithelial cells, and endothelial cells) could internalize fine particles (Figure S6). The quantitative analysis of MFI values of lung cell populations at 2 and 6 h post‐particle treatment indicated that macrophages exhibited the highest MFI values at both time points among the evaluated pulmonary cell types, with the highest MFI value at 6 h, signifying the strongest phagocytic activity at this time point (Figure 2E). The heterogeneity in fine particle–related response may be attributed to the distinct functions of different cell types in maintaining pulmonary homeostasis, and the predominance of macrophages and neutrophils in the Cy3‐positive group suggested their crucial roles in particle phagocytosis, clearance, and inflammatory responses [28, 29, 30].

FIGURE 2.

Macrophages dominate pulmonary immune response upon LbL‐AuNP administration. (A) A uniform manifold approximation and projection (UMAP) plot was generated for 72,558 profiled cells derived from lung tissue of PBS‐treated control mice and LbL‐AuNP‐treated mice, including cells from blank control group (n = 3), Cy3‐negative group (n = 3) and Cy3‐positive group (n = 3). (B) Dot plot representing the expression of selected marker genes of individual cell types. (C) Proportions and (D) UMAP plots of cell types in the lung of blank control group, Cy3‐negative group and Cy3‐positive group, delineating cell types through differential coloring. (E) Pyramidal plot of mean fluorescence intensity (MFI) values in pulmonary cell populations of mice treated with LbL‐AuNPs after 2 and 6 h. Flow cytometry was employed to analyze and quantify MFI values of various pulmonary cell populations in mice treated with LbL‐AuNPs at 2 and 6 h post‐treatment. Data were then visualized using pyramidal plot to illustrate changes in MFI values over time across different cell populations. (F) Detection levels of Tag RNAs across different groups (left panel) and cell types (right panel). (G and H) Heatmaps illustrate differential enrichment of selected gene sets in (G) phagocytosis‐related pathways and (H) inflammatory response‐related pathway across distinct cell subsets. Color scale reflects z‐score of normalized enrichment score for each pathway. (I) GO enrichment analysis of upregulated differentially expressed genes (DEGs) in AMs (left panel), IMs (middle panel) and recMacs (right panel) between Cy3‐negative and Cy3‐positive groups. P_adjust was calculated by the Benjamini‐Hochberg method. (J) Schematic illustration of scRNA‐seq workflow to reveal cell type composition in the lung of LbL‐AuNP ‐ treated mice, including cell sample processing, dimensionality reduction and clustering analysis, as well as functional analysis of cell populations.

Next, we utilized Tag RNAs loaded on LbL‐AuNPs to quantify the cellular phagocytosis of fine particles in distinct cell types. The analysis of Tag RNA detection levels showed that the Cy3‐positive group had the highest average detection level (Figure 2F). Further analysis across different cell types showed that AMs, IMs, and recMacs exhibited the highest percent of Tag RNA detection, followed by AT2, neutrophils, endothelial cells, and DCs (Figure 2F). Additionally, gene set enrichment analysis of phagocytosis‐associated genes revealed relatively high enrichment scores for AMs, IMs and recMacs, with recMacs demonstrating the highest score (Figure 2G). Similarly, AMs, IMs and recMacs exhibited relatively high enrichment scores in the analysis of inflammation‐related gene sets (Figure 2H). These findings indicate AMs, IMs and recMacs as the predominant cell types responding to particle administration.

To further explore the internalization of LbL‐AuNPs in macrophages, we performed in vitro experiments using mouse primary BMDMs and mouse macrophage cell line J774A.1 cells. Confocal microscopy results showed that fine particle internalization, rather than cell adhesion, increased at 6 h post‐treatment and became more pronounced between 12 and 24 h (Figure S7A,B). Furthermore, TEM and EDS analysis of LbL‐AuNP‐treated BMDMs revealed aggregated spherical particles within the cytoplasm (Figure S7C,D). Similarly, the TEM images of LbL‐AuNP‐treated J774A.1 cells corroborated the presence of these particles (Figure S7E).

The heterogeneity among distinct subpopulations of the same cell type reflects their differentiation states and effector functions [31]. However, the impact of such heterogeneity on the rapid activation and signal transmission of fine particle–mediated immune responses remains unclear. Therefore, we compared the transcription characteristics of AMs, IMs and recMacs after LbL‐AuNP treatment. Compared to the Cy3‐negative group, AMs, IMs and recMacs exhibited distinct response patterns in the Cy3‐positive group (Figure 2I). Specifically, upregulated differentially expressed genes (DEGs) in AMs of the Cy3‐positive group were enriched in oxidative phosphorylation, cellular respiration and mitochondrial translation, implying an increased demand for energy supply in AMs in response to fine particles (Figure 2I). The upregulated DEGs of IMs were primarily enriched in the regulation of the JNK cascade and the regulation of stress‐activated MAPK cascade, revealing enhanced signal transduction in IMs of the Cy3‐positive group (Figure 2I). Furthermore, the upregulated DEGs of recMacs were enriched in leukocyte cell–cell adhesion, regulation of adaptive immune response and positive regulation of leukocyte activation, suggesting reinforced immune responses in these cells in response to particles (Figure 2I). Together, our findings demonstrate that AMs, IMs and recMacs are involved in the recognition and phagocytosis of exogenous particles, play a predominant role in recognizing and clearing the pulmonary immune response following fine particle treatment (Figure 2J).

2.3. Distinct Expression Signatures Reveal Pulmonary Macrophage Heterogeneity

To further investigate the functional heterogeneity of the pulmonary macrophage subsets, we performed dimensionality reduction and clustering analysis on the three macrophage populations and identified 14 subclusters (Figure 3A). The blank control group contained 14 identified macrophage subsets, among which AMs_1 (31.5%), AMs_5 (18.0%), and recMacs_1 (13.4%) ranked as the top three in terms of percentage (Figure 3B,C), suggesting their predominant roles in tissue homeostasis, basal immunosurveillance and repair under normal physiological conditions [32, 33]. In addition, the Cy3‐negative group exhibited the lowest number of cells and reduced subset diversity among the three groups, comprising 10 macrophage subtypes with the absence of AMs_1, AMs_2, AMs_3 and AMs_7. The Cy3‐positive group shared the same macrophage subset composition as the blank control group while exhibiting distinct proportional differences (Figure 3B,C). Notably, the Cy3‐positive group comprised the largest proportion of cells within the AMs_3 (75.5%), AMs_7 (91.3%), AMs_8 (98.6%), and recMacs_4 subsets (78.4%), suggesting that these macrophage subclusters are more sensitive to fine particles (Figure S8). Overall, these findings imply that the complexity and diversity of cell subpopulations should be considered, rather than regarding cell types as single response units, when studying the biological effects of fine particles.

FIGURE 3.

Diversity of cell lineage and functionality among alveolar macrophages (AMs), interstitial macrophages (IMs) and recruited macrophages (recMacs). (A) UMAP projection of 11,301 macrophages showing the composition of 14 main subtypes. (B) Proportions and (C) UMAP visualization of different macrophage subtypes derived from lungs in blank control group, Cy3‐negative group and Cy3‐positive group. Cells are colored according to their subtype, with each dot representing individual cell. (D) Heatmap showing expression of representative differentially expressed genes (DEGs, left panel) and gene signature elucidation based on gene ontology (GO) enrichment analysis (right panel) across each macrophage cluster.

Next, we investigated the differences in the biological pathways of the 14 identified macrophage subclusters by conducting functional enrichment analysis of their DEGs. DEGs of the AM subsets were enriched in pathways associated with immune responses, tissue homeostasis, cell development and differentiation, metabolic regulation and antiviral defense (Figure 3D). Among them, DEGs of the AMs_3 and AMs_7 subsets were enriched for cell response to biotic stimulus and acute‐phase response, reflecting that these subsets were likely involved in the rapid detection and reaction to fine particles (Figure 3D). By contrast, DEGs of the AMs_8 and AMs_9 subsets were enriched in the pathways associated with tissue homeostasis and tissue remodeling, highlighting the importance of these subsets in maintaining tissue integrity and facilitating repair processes (Figure 3D). Different subsets of recMacs harbored diverse functions such as immune defense, signaling regulation and stress response (Figure 3D). Specifically, DEGs of the recMacs_4 subset were enriched for cell chemotaxis and phagocytosis, implying this subset potentially participated in particle responses through immune cell recruitment and fine particle engulfment (Figure 3D). DEGs of the IM subcluster were enriched for Rho protein signal transduction and small GTPase mediated signal transduction, indicating IM subcluster's potential in enhancing particle responses through signal transduction (Figure 3D). Therefore, the macrophage subsets exhibited partially overlapping yet distinct functions in fine particle response, highlighting the complex functions and heterogeneity of the different pulmonary macrophage subsets in maintaining lung homeostasis.

2.4. Specific Pulmonary Macrophage Subsets Harbor Elevated Responsibility for the Rapid Engulfment of Exogenous Fine Particles

The study of gene expression dynamics can organize single cells in pseudotime order, thereby clarifying the principal factors influencing cellular programs that regulate fate transitions [34, 35]. We performed pseudotime trajectory analysis on the AM subclusters, resulting in the identification of three distinct developmental trajectories (Figure 4A). Three trajectories started from AMs_9, transiting through AMs_2 to give rise to AMs_5 (trajectory 1), AMs_7 (trajectory 2) or AMs_9 (trajectory 3), with trajectory 2 encompassing the greatest cell subtypes. The pivotal role of AMs_2 as a branching point was further supported by RNA velocity analysis (Figure S9A). Moreover, we examined the expression levels of representative immunomodulatory genes (Trem1 and Isg15) during the development of AM subclusters. The expression levels of Trem1 and Isg15 were upregulated in higher‐tier cells along the developmental trajectory compared to the lower‐tier cells (Figure 4B). We also performed pseudotime trajectory analysis in the recMac subclusters (Figure 4C). This trajectory started from recMacs_3, transiting through recMacs_2 and recMacs_1 to give rise to recMacs_4, which exhibited similar expression patterns of Trem1 and Isg15 genes (Figure 4C,D). Furthermore, RNA velocity analysis further revealed discrete transitional steps from recMacs_3 to recMacs_2 and from recMacs_2 to recMacs_1 (Figure S9B). These findings reveal dynamic changes in gene expression during the developmental progression of macrophage subclusters, offering insights into the molecular mechanisms that govern their differentiation and functional specialization.

FIGURE 4.

AMs_3, AMs_7 and recMacs_4 subsets are responsible for the rapid phagocytosis of exogenous fine particles. (A) Slingshot‐based pseudotime trajectory analysis in AM subtypes. Left panel: UMAP visualization of nine distinct AM subtypes (AMs_1–AMs_9), color‐coded according to provided legend. Right panel: UMAP visualization incorporating pseudotime analysis to track progression of cellular states. (B) Expression dynamics of Trem1 and Isg15 genes across AM subtypes in pseudotime analysis. (C) Slingshot‐based pseudotime trajectory analysis in recMac subtypes. Left panel: UMAP visualization of four distinct recMac subtypes, color‐coded according to provided legend. Right panel: UMAP visualization incorporating pseudotime analysis to track progression of cellular states. (D) Expression dynamics of Trem1 and Isg15 genes across recMac subtypes in pseudotime analysis. (E) Violin plots displaying AUCell scores for GOBP_PHAGOCYTOSIS, HALLMARK_INFLAMMATORY_RESPONSE and GOBP_MACROPHAGE_CYTOKINE_PRODUCTION across distinct macrophage subsets. (F) Detection levels of Tag RNAs across different groups (left panel) and different cell clusters (right panel). (G–I) GO enrichment analysis of upregulated DEGs in (G) AMs_3, (H) AMs_7 and (I) recMacs_4. P_adjust was calculated by the Benjamini‐Hochberg method. (J and K) Volcano plot illustrating DEGs in (J) AMs_3&7 and (K) recMacs_4 between blank control and Cy3‐positive groups. Red: Upregulated DEGs (log₂ [fold change] ≥ 1 and P_adjust < 0.05), Green: Downregulated DEGs (log₂ [fold change] ≤ ‐1 and P_adjust < 0.05), Gray: Genes with no significant difference in expression (|log₂ [fold change]| < 1 or P_adjust ≥ 0.05). P_adjust was calculated by the Bonferroni method. (L) Schematic depicting interactions between particles and specific macrophage subsets exhibiting most active response to particles.

To figure out the differential responses of macrophage subclusters to fine particle administration, we measured the AUCell scores of each subcluster in selected gene sets. Among the 14 macrophage subsets identified, the recMacs_4 subcluster exhibited the highest AUCell score (0.16) in the GOBP_PHAGOCYTOSIS gene set, while recMacs_1, recMacs_2 and AMs_7 displayed robust activity with scores of 0.15, 0.15 and 0.14, respectively, underscoring their notable phagocytic ability (Figure 4E). The recMacs_4 subcluster demonstrated the highest score of 0.16 within the HALLMARK_INFLAMMATORY_RESPONSE gene set, followed by AMs_7 and AMs_3 with scores of 0.15 and 0.14, respectively (Figure 4E). In the GOBP_MACROPHAGE_CYTOKINE‐PRODUCTION gene set, both recMacs_4 and AMs_3 exhibited the highest activity with scores of 0.16, while AMs_7 followed with a score of 0.15 (Figure 4E). Ridge plots display the distribution of AUCell scores for the aforementioned gene sets across distinct macrophage subsets (Figure S10). Furthermore, we determined the detection levels of exogenous Tag RNAs to assess fine particle uptake among the various macrophage subclusters. Notably, Tag RNAs were mainly detected in the Cy3‐positive group compared to the other two groups and were even higher in AMs_3, AMs_7 and recMacs_4 among the 14 macrophage subclusters, confirming that these subclusters may act as pivotal factors in fine particle endocytosis (Figure 4F). Collectively, these results highlight the rapid response of AMs_3, AMs_7 and recMacs_4 subclusters to fine particle treatment.

Furthermore, we delineated the transcriptional alterations and enriched gene ontology (GO) pathways underlying the reaction of AMs_3, AMs_7 and recMacs_4 subsets to fine particle treatment. As illustrated in Figure 4G, the upregulated DEGs in the AMs_3 subset were enriched in epidermal growth factor receptor binding and carbohydrate binding, implying this subset primarily responded to particles through signal transduction and cell recognition [36, 37, 38]. The enriched GO molecular functions in the AMs_7 subset included immune receptor activity and pattern recognition receptor (PRR) activity, suggesting this subset likely employed PRRs to mediate its reaction (Figure 4H). Similarly, the enriched GO molecular functions in the recMacs_4 subset included PRR activity, toll‐like receptor binding and cell adhesion molecule binding, indicating that this subset relied on early surface recognition and cell adhesion in response to fine particles (Figure 4I). To gain mechanistic insights into the heterogeneity of fine particle response, we sorted out the effects of fine particle treatment on AMs_3, AMs_7, and recMacs_4. Compared with the blank control group, genes ascribed to acute‐phase responses (e.g., Saa3 and Stat3) [39, 40], phagocytosis (e.g., Plscr1, Thbs1 and Vav1) [41, 42, 43], interleukin‐1 (IL‐1) signaling (e.g., Il1r2, Il1rap and Il1rn) [44, 45, 46] and inflammatory chemokines (e.g., Ccl4, Ccr1 and Cxcl3) [47, 48, 49] were significantly upregulated in the AMs_3&7 subsets of the Cy3‐positive group (Figure 4J, log₂ [fold change] ≥ 1 and P_adjust < 0.05). Furthermore, GO functional enrichment analysis demonstrated that the upregulated DEGs in the AMs_3&7 subsets were mainly enriched in response to IL‐1, phagocytosis, acute‐phase response, as well as cell chemotaxis (Figure S11A, P_adjust < 0.05). In addition, the analysis of DEGs in the recMacs_4 subset of the Cy3‐positive group recognized significant upregulation of genes associated with phagocytosis (e.g., Fcgr3 and Tlr4) [50, 51] and inflammatory chemokines (e.g., Ccl4, Ccr1 and Cxcl3) [47, 48, 49] (Figure 4K, log₂ [fold change] ≥ 1 and P_adjust < 0.05). Furthermore, GO functional enrichment analysis demonstrated that the upregulated DEGs in the recMacs_4 subset were enriched in phagocytosis, PRR signaling pathway, cellular response to biotic stimulus and cell chemotaxis (Figure S11B, P_adjust < 0.05). Together, these findings reveal the specialized roles of each macrophage subset in fine particle clearance and immune responses, offering insights into the diverse functions of pulmonary macrophages in response to particles (Figure 4L).

In the analysis of macrophage subpopulations, AMs_3 and AMs_7 subsets exhibited higher expression levels of Cd274 (Figure S12A), the gene encoding programmed death‐ligand 1 (PD‐L1) [52]. Notably, within the Cy3‐positive group, the expression of Cd274 gene was found to be the highest (Figure S12A). To define the transcriptional profile of the PD‐L1^+/high AM subpopulation, we employed FACS to isolate both PD‐L1^+/high and PD‐L1^−/low AM subpopulations for RNA sequencing (Figure S12B). The results revealed that the PD‐L1^+/high AM subset exhibited significant upregulation of DEGs associated with phagocytosis (e.g., Tlr4, Cd36 and Mertk) [50, 53, 54] compared to the PD‐L1^−/low AM subset (Figure S12C, log₂ [fold change] ≥ 1 and P_adjust < 0.05), which aligns with our scRNA‐seq results. GO enrichment analysis of the upregulated DEGs in the PD‐L1^+/high AM subpopulation demonstrated significant enrichment in pathways related to phagocytosis, cellular response to environmental stimulus and chemotaxis (Figure S12D, P_adjust < 0.05). Additionally, immunofluorescence micrographs of lung tissue cryosections demonstrate the co‐localization of a PD‐L1⁺ AM with engulfed fine particles in the alveoli (Figure S12E). These findings suggest that PD‐L1 signaling may play an important role in the enhanced phagocytic activity observed in AMs_3&7 subsets following exposure to particles.

2.5. Exogenous Particles Induce Active Communication Between Specific Macrophage Subsets and Other Cell Subpopulations

We analyzed cellular crosstalk patterns that contribute to fine particle–induced pulmonary inflammation by integrating ligand and receptor information from scRNA‐seq data to develop a putative interaction network among lung cell populations. The number and strength of intercellular interactions increased among lung cell subpopulations following fine particle treatment (Figure 5A; Figure S13). Compared with the blank control group, the Cy3‐positive group exhibited largely enhanced interactions associated with inflammatory signaling pathways (e.g., IL6 and TNF) [55, 56], chemokines (e.g., CXCL and CCL) [57] and cell migration and adhesion (e.g., MIF and SPP1) [58, 59] (Figure 5B). Furthermore, the Cy3‐positive group showed more pronounced interactions associated with inflammatory signaling pathways (e.g., IL6 and TNF) [55, 56], chemokines (e.g., CXCL and CCL) [57] and cell proliferation and migration (e.g., APRIL and MIF) [58, 60] compared with the Cy3‐negative group (Figure 5B). These results substantiate that particle intrusion triggers extensive intercellular communication networks within pulmonary cell subpopulations, characterized by enhanced ligand–receptor interactions involving proinflammatory mediators and tissue remodeling factors.

FIGURE 5.

AMs_3, AMs_7 and recMacs_4 subsets are key signaling nodes in fine particle–disturbed cellular communication network. (A) Histogram showing the strength and number of cell interactions calculated by CellChat in blank control, Cy3‐negative and Cy3‐positive groups. (B) Prominent cell communication signaling pathways ordered based on differences in relative information flow between blank control and Cy3‐positive groups (left panel), blank control and Cy3‐negative groups (middle panel), and Cy3‐negative and Cy3‐positive groups (right panel). (C) Chord diagram illustrating inferred intercellular communication network across pulmonary cell populations for C‐X‐C motif chemokine ligand (CXCL) and macrophage migration inhibitory factor (MIF) signaling pathways in blank control, Cy3‐negative and Cy3‐positive groups. (D) Heatmaps depicting roles of CXCL (upper panel) and MIF (lower panel) signaling pathways in pulmonary cell subpopulations in blank control, Cy3‐negative and Cy3‐positive groups.

We further analyzed the CXCL and MIF signaling pathways between macrophage subsets and other lung cell populations to validate the role of specific macrophage subpopulations in fine particle–related responses. The results showed that AMs_7 and recMacs_4 exhibited increased incoming interactions within the CXCL signaling pathway in the Cy3‐positive group (Figure 5C). Consistently, the CXCL signaling pathway was intensified both from AMs_7 to AMs_6, AMs_7, recMacs_4 and neutrophils, and from recMacs_4 to AMs_6, AMs_7, DCs and neutrophils in the Cy3‐positive group (Figure 5C). Additionally, the AMs_3, AMs_7 and recMacs_4 subsets harbored increased outgoing interactions in the MIF signaling pathway in the Cy3‐positive group (Figure 5C). In the Cy3‐positive group, the relative importance of AMs_7 and recMacs_4 as senders of the CXCL signaling pathway, as well as that of AMs_3, AMs_7 and recMacs_4 as mediators and influencers of the MIF signaling pathway, was distinctly enhanced (Figure 5D). These results imply that the ability of specific macrophage subsets to orchestrate responses in the CXCL and MIF signaling pathways improved following fine particle treatment. The cellular communication network analysis further revealed a complex and redundant signaling landscape, with the AMs_3, AMs_7 and recMacs_4 subsets appearing as central nodes in the predicted cellular communication network.

3. Discussion

Currently, various methods have been developed to recognize refined cell populations, including flow cytometry, scRNA‐seq, spatial transcriptomics, multiplex immunohistochemistry/immunofluorescence and combined applications of these methods [61, 62, 63, 64, 65]. Compared with other methods, scRNA‐seq provides an unbiased and comprehensive view of the transcriptome, and its single‐cell resolution enables in‐depth analysis of cellular heterogeneity in broad complex systems, such as immune cells and stem cells. [17] However, scRNA‐seq cannot directly identify the particle–phagocytized cells from sequencing data owing to the lack of unique identifiers for inhaled fine particles. This challenge makes monitoring and analysis of particle engulfment at single‐cell precision extremely difficult. With the advancement of technology, the integration of scRNA‐seq with other experimental techniques, such as mass cytometry (cytometry by time ‐ of ‐ flight, CyTOF) and FACS, attempts to infer particle engulfment and elucidate relevant transcriptional information [66, 67]. CyTOF achieves single‐cell analysis of the association between cellular responses and the dose of metal‐based particles [66]. However, the limited types of available metal labels restrict the exploration of this relationship following detailed cell subset stratification [68]. Furthermore, fluorescent labeling allows the sorting of particle–engulfed cells via the integration of FACS and scRNA‐seq, while the single‐cell fluorescence information cannot be reconstructed, thereby preventing its joint analysis with transcription profiling data. Herein, we established Cy3‐labeled LbL‐AuNPs loaded with poly(A) tail‐modified Tag RNAs to identify the prominent fine particle–responsive cell subpopulations through FACS and scRNA‐seq. Following LbL‐AuNP treatment, FACS was employed to distinguish cells that phagocytosed fine particles (Cy3‐positive cells) from those treated with fine particles but which did not phagocytose them (Cy3‐negative cells). Subsequently, scRNA‐seq analysis generated a cellular atlas of mouse lung tissues and successfully evaluated particle internalization based on Tag RNA detection levels.

In the cell‐type‐specific transcriptome atlas, 13 cell populations were identified, including immune cells (AMs, IMs, recMacs, neutrophils, DCs, NK cells, T cells, plasma cells and B cells) and nonimmune cells (AT1, AT2, endothelial cells and fibroblasts), consistent with the results of previous annotation studies of lung cells [18, 27]. Herein, AMs, IMs and recMacs achieved high scores in gene sets related to inflammatory responses and phagocytosis compared with other cell types, reflecting that they were the most active cell populations in response to fine particle inhalation. A similar study generated a mumps virus (MuV) strain tagged with green fluorescent protein (GFP) and Cre recombinase to delineate the differential susceptibility of immune cells [69]. Infected animal models demonstrated that AMs were the early cellular targets of MuV in vivo, which showed high susceptibility and responsiveness to respiratory pathogens [69]. Another study revealed the roles of different macrophage lineages in Mycobacterium tuberculosis (Mtb) infection [70, 71, 72]. AMs were found to form an anti‐inflammatory M2‐type population conducive to Mtb replication and spread [70, 71, 72], while IMs were associated with a more bactericidal immunological setting [71, 72]. Despite initial research on the functional heterogeneity within macrophage lineages [67], the varying responses of different pulmonary macrophage populations to foreign particles remain underexplored. Our results suggest that the main fine particle–responsive cell populations (including AMs, IMs and recMacs) not only exhibited robust phagocytic abilities but also elevated inflammatory responses, reflecting a high degree of responsiveness.

We further analyzed gene set scores and Tag RNA characteristics among the 14 macrophage subsets, and identified three main fine particle‐responsive macrophage subsets: AMs_3, AMs_7 and recMacs_4. This study expands on the potential functions of these macrophage subsets, with the recMacs_4 subset demonstrating the most pronounced phagocytosis, inflammatory response and cytokine production, highlighting its importance in particle clearance and the initiation of immune responses. In addition, the AMs_3 and AMs_7 subclusters showed strong inflammatory responses and macrophage factor–producing capabilities, further confirming their effector functions against external stimuli. A recent study revealed the spatiotemporal dynamics of pulmonary macrophage subpopulations at single‐cell resolution upon SARS‐CoV‐2 infection [26]. Among the different macrophage subsets, Slamf9⁺ macrophages efficiently recruited neutrophils to collaboratively phagocytose viruses and further into the Trem2⁺ and Fbp1⁺ subclusters to alleviate inflammatory responses after viral clearance [26]. In our study, the functional heterogeneity of specific macrophage subsets similarly served as a pivotal mechanism to counter particle treatment. At the transcriptional level, the AMs_3&7 and recMacs_4 subsets in the Cy3‐positive group showed an upregulation of the genes associated with cell chemotaxis (e.g., Ccr1, Ccl4 and Cxcl3) and phagocytosis (e.g., Plscr1, Fcgr3 and Tlr4). Likewise, these specific macrophage subclusters exhibited enhanced interactions in the CXCL and MIF signaling pathways. This observation underscores the potential role of these subclusters in mediating cellular communication. Overall, the findings of this study enhance our understanding of the cell‐type‐specific interaction patterns in macrophage responses to fine particles and offer novel perspectives for studying macrophage functions in environmental stimuli.

The role of PD‐L1, encoded by the CD274 gene [73], in regulating biological functions of macrophages remains controversial [74, 75, 76]. Numerous studies have focused on the tumor‐associated macrophages (TAMs), which are the predominant PD‐L1‐expressing cell populations in tumor microenvironment and are generally thought to exhibit immunosuppressive function [74, 77, 78]. Nevertheless, a recent scRNA‐seq study has identified PD‐L1^+/high TAMs as more mature, activated, and immunostimulatory, compared with the PD‐L1^−/low subgroup [79]. PD‐L1 signaling is reported to be rapidly upregulated during the monocyte‐to‐macrophage maturation process [79]. In the macrophages differentiated from human pluripotent stem cells, PD‐L1 knockout could hinder macrophage development and inhibit inflammatory program initiation [80]. Besides, PD‐L1 expression is also associated with the phagocytic function of macrophages [74, 81, 82]. The process of macrophages recognizing and phagocytosing yeast is dependent on PD‐L1 as a fungal‐binding receptor [81]. Similarly, the absence of PD‐L1 signaling could suppress M1 macrophage polarization and undermine their antimicrobial capability [82]. Despite these progresses, the functional significance of PD‐L1 expression in macrophages needs to be further explored, especially in response to inhaled particles. Building on these studies, our scRNA‐seq transcriptomic analysis has revealed that AMs_3&7 subpopulations marked by high PD‐L1 expression possess a transcriptional signature indicative of enhanced phagocytic and environmental sensing capabilities, suggesting a broader role for PD‐L1 in macrophage‐mediated particle clearance. To definitively validate this role, future experiments should investigate the development of these subsets and, importantly, the precise role of PD‐L1 itself in orchestrating their phagocytic function.

In fact, there are certain limitations in this study. First, although our study identified the primary particle‐responsive macrophage subpopulations, a key limitation was the lack of functional validation via direct depletion or inhibition experiments due to the absence of subgroup‐specific antibodies. We are eager to address this limitation to substantially enhance our understanding of the phenomena under investigation in future studies. Second, it is worth noting that our analysis using CellChat predicted the probability of ligand‐receptor interactions but did not demonstrate functional validation of these signaling events. This approach offers initial insights into potential cell communication networks, yet the true functional significance of these interactions still remains to be empirically verified. Thirdly, the joint analysis strategy utilizing Tag RNA aimed to identify the detection of particle uptake, while a potential for artificial transcriptomic effects could not be avoided in cells internalizing particles. Finally, our analysis was conducted within a 6 h time window, which effectively captured the acute phagocytic and early inflammatory responses but did not encompass the longitudinal dynamics or the inflammatory resolution phase. Despite these limitations, we believe that our approach, which integrates gene expression profiling, pathway enrichment analysis, pseudotime analysis, functional scoring and cell communication, provides a robust foundation for our conclusions. The current work establishes an essential foundation and provides a compelling rationale for the critical future experiments that would build upon our initial findings.

4. Conclusions

Identifying pulmonary cell subpopulations that remove exogenous particles and modulate inflammatory responses is critical for understanding pathogenesis and developing therapeutic methods for lung injury. Here, we established Cy3‐labeled LbL‐AuNPs loaded with poly(A) tail‐modified Tag RNAs to capture fine particle‐responsive cell populations and characterize their functional heterogeneity via FACS and scRNA‐seq. In the single‐cell transcriptome atlas, specific macrophage subsets (AMs_3, AMs_7 and recMacs_4) demonstrated the most direct engulfment and activation against exogenous particles. Driver events at the transcriptional level, including genes ascribed to the cell chemotaxis (e.g., Ccr1, Ccl4 and Cxcl3) and phagocytosis (e.g., Plscr1, Fcgr3 and Tlr4), were upregulated in these three subsets. Moreover, AMs_3&7 subsets with high Cd274 expression showed a significant enrichment of acute‐phase signaling, including phagocytosis, cellular response to environmental stimulus and chemotaxis. Cell communication analysis revealed that these three cell subsets established enhanced interactions with different cell populations in response to fine particles, implying their potential roles in inflammatory modulation.

In summary, our developed LbL‐AuNP model colabeled with fluorescence and Tag RNA partially addresses the limitation of scRNA‐seq in directly evaluating single‐cell particle phagocytosis and offers a new approach to study cellular responses to fine particles. Herein, we systematically investigated the responses of pulmonary cell subpopulations to exogenous particles, encompassing cellular immune response, functional heterogeneity of subclusters and remodeling of intercellular interactions. These findings enhance our understanding of the interactions between particles and biological systems, and provide theoretical guidance for the design, development and optimization of nanoadjuvants. Additionally, by precisely identifying and analyzing changes in crucial cell subpopulations and signaling pathways following particle inhalation, this study offers a scientific basis and intervention strategies for addressing potential pulmonary diseases caused by atmospheric fine particles.

5. Experimental Section

5.1. Preparation and Modification of Tag RNAs for Particle Loading

To prepare the in vitro transcription template, the GFP fragment, serving as the template DNA, was cloned into the pGM‐T vector (TIANGEN, China). The GFP fragment does not align to the murine genome. Thus, appropriate plasmids were linearized using the SalI‐HF enzyme (NEB, USA) and were used as DNA templates for the in vitro transcription reaction. The TranscriptAid T7 High‐Yield Transcription Kit (Thermo Fisher Scientific, USA) was utilized to generate RNA fragments containing the target sequence through in vitro transcription, following the provided experimental guidelines. Subsequently, DNase I was employed to treat the in vitro transcription products to digest template DNA. In accordance with the manufacturer's instructions, poly(A) tails were added to RNA fragments using a polyadenylation kit (Thermo Fisher Scientific, USA). Following purification, RNA fragments with poly(A) modifications were obtained and designated as “Tag RNAs”, which were used for the subsequent loading of particles.

5.2. Synthesis of LbL‐AuNPs

AuNPs were synthesized via the reduction of tetrachloroauric acid with sodium citrate, following a previously reported procedure [83]. Briefly, a 1.0 mL solution of 1% gold chloride trihydrate (AuCl₄·3H₂O, ≥ 99%, Acmec Biochemical, China) was diluted to 100 mL and heated until boiling under reflux. Subsequently, 2.5 mL of 1% trisodium citrate (C₆H₅Na₃O₇, Acmec Biochemical, China) was added under vigorous stirring, and the mixture was boiled for an additional 10 min. The synthesized AuNP dispersion was alkalinized to pH 11.0 with 1 N NaOH, after which MUA (Sigma‐Aldrich, USA) was introduced at a concentration of 0.1 mg/mL. The AuNPs were then purified via centrifugation and washed with DNase/RNase‐free water thrice. Subsequently, LbL‐AuNPs were prepared according to a previously reported method [19]. Furthermore, purified AuNPs were sequentially incubated in polyelectrolyte solutions of PEI (average Mw ∼25,000 Da; Sigma‐Aldrich, USA) to form PEI‐AuNPs. Then, PEI‐AuNPs were incubated with Cy3 NHS ester (Duofluor, China) for 6 h to produce Cy3‐PEI‐AuNPs, which were then purified using DNase/RNase‐free water thrice. Next, Cy3‐PEI‐AuNPs were sequentially incubated with Tag RNAs and PEI, ultimately forming PEI/RNA/Cy3‐PEI‐AuNPs. Each coating step with PEI or Tag RNAs was conducted for 30 min. Following each coating step, the particles were purified thrice to remove any unbound polyelectrolytes.

5.3. Animal Experimentation

All animal experiments were performed under protocols approved by the Animal Ethical Committee of Shandong First Medical University (approval number: W202505090728). Seven‐week‐old female BALB/c mice were obtained from Vital River Laboratories (Beijing, China) for all animal experiments. To investigate how pulmonary immune cells respond to fine particles, mice were treated via intratracheal instillation with LbL‐AuNPs at 2 mg/kg body weight for 2, 6 and 12 h. The fine particle solution (1 mg/mL) was formulated in PBS. For intratracheal instillation, the fine particle solution was instilled into the trachea of anesthetized mice using a 20‐gauge catheter. Control animals were given the same volume of vehicle solution (PBS only).

5.4. Detection of Cell Populations in Lung Tissue via Flow Cytometry

Following particle treatment, lung tissues were collected from mice for analyzing cell populations using an Attune NxT flow cytometer (Thermo Fisher Scientific, USA). Lung single cell suspensions were generated via enzymatic digestion using Liberase TM (Roche, Switzerland) and DNase I (Solarbio, China) at 37°C for 30 min. Red blood cells (RBCs) were lysed using RBC lysis buffer (Solarbio, China) for 3 min. Cell suspensions were filtered twice using a 70‐µm cell strainer to achieve single‐cell suspensions, and then 1.0 × 10⁶ cells were isolated from each tissue sample for staining. Fluorescent dye‐conjugated antibodies were utilized to label cell surface markers during a 30‐min incubation at 4°C, protected from light. Nonspecific antibody binding was blocked with anti‐mouse CD16/CD32 (Biolegend, China) before staining. Table S1 lists the antibodies utilized for flow cytometry.

5.5. Single‐Cell Suspension Preparation for Sequencing

All single, 4',6‐diamidino‐2‐phenylindole (DAPI)‐negative living lung cells were sorted using FACS into cells with or without Cy3 positivity. These cells were derived from single‐cell suspensions of the lung tissue of control mice treated with PBS and mice treated with LbL‐AuNPs for 6 h. Additionally, sorted cells were stained with Trypan blue and microscopically evaluated to quantify cell numbers, assess viability, and evaluate aggregate formation. All samples demonstrated cell viability above 95%, and no aggregates were detected.

5.6. Library Preparation and Sequencing

The prepared single‐cell suspensions of lung tissue were subjected to capture using the 10× Genomics Chromium system (10× Genomics). Following the manufacturer's protocol, cells were loaded into the Chromium Controller to generate gel bead‐in‐emulsions. Subsequently, single‐cell 3ʹ library and gel bead reagent kits (10× Genomics, USA) were employed to perform cell barcode labeling and cDNA library construction. Finally, indexed sequencing libraries were prepared utilizing the Chromium Single Cell 3ʹ Library Kit (10× Genomics, USA). Sequencing was performed on the Illumina NovaSeq X Plus platform at Shanghai Novogene Corporation (Novogene, China). The raw data were subjected to quality control to remove low‐quality reads.

5.7. Construction of a Reference Genome

The official mouse reference genome mm10‐2020‐A (https://www.10xgenomics.com/support/software/cell‐ranger/latest/release‐notes/cr‐reference‐release‐notes#2020‐a) provided by Cell Ranger was used as the baseline genomic sequence. The Tag RNA sequence in FASTA format was incorporated into this genomic sequence and designated as “Targetgene”. Additionally, the corresponding gene transfer format (GTF) annotation file was updated to include “Targetgene” as an independent gene entry. Furthermore, the Cell Ranger mkref function was employed to specify the newly constructed genomic sequence and the updated GTF annotation file, thereby completing the indexing of the Cell Ranger reference genome.

5.8. Transcriptome Data Processing and Quality Control

The Cell Ranger Single‐Cell Software Suite (version 7.1, 10× Genomics) was utilized to align sequencing data to the customized reference genome, followed by transcript‐level quantification. To ensure data quality, low‐quality data were filtered according to the following thresholds: (i) genes detected in fewer than 3 cells were excluded, (ii) cells harboring < 200 or >7,500 detected genes were removed, and (iii) cells with a mitochondrial gene count proportion exceeding 15% were discarded. The DoubletFinder tool (https://github.com/chris‐mcginnis‐ucsf/DoubletFinder) was implemented to detect and remove potential doublets.

The samples were merged into a single Seurat object using the merge function. The data were normalized using the NormalizeData function with the LogNormalize method. The top 3,000 highly variable genes were identified using the FindVariableFeatures function with the “vst” method. Subsequently, the data were scaled and subjected to principal component analysis using the ScaleData and RunPCA functions, respectively. To correct for batch effects, integration across samples was performed using the IntegrateLayers function with the “HarmonyIntegration” method.

5.9. Unsupervised Cell Clustering and Annotations

Following integration, the top 30 principal components were used to further generate UMAP dimensionality reductions of the RunUMAP function. Unsupervised clustering was performed using the FindClusters function with multiple resolutions and the result at resolution 0.5 was ultimately selected for subsequent annotation. Clustering results were visualized using UMAP [84]. Annotation of the cell clusters was based on the known markers obtained from previous studies [25, 26, 27] and the CellMarker database (http://biocc.hrbmu.edu.cn/CellMarker/).

5.10. Pseudotime Trajectory Inference

Pseudotime‐based cell trajectory analysis was conducted using the Slingshot package (version 2.1.0) [35]. The Seurat object was first converted to a SingleCellExperiment object to match the input format required by Slingshot. The Slingshot function was employed to construct the global lineage structure based on the minimum spanning tree algorithm. This function also utilized UMAP dimensionality reduction and clustering labels from the Seurat object to identify cell trajectories. The calculation of each trajectory ensured a distinct starting point and endpoint. For enhanced clarity and conciseness, multiple trajectories were plotted on the same chart. Trajectories of gene expression changes over pseudotime were visualized using ggplot2.

5.11. Cell–Cell Communication Analysis

To explore the interactions between different cell types, we used the CellChat software package (version 2.1.2) [85] to determine active ligand–receptor interactions. CellChat objects were created from single‐cell transcriptomic data of three groups. After preprocessing with the CellChatDB.mouse ligand–receptor database, intercellular communication probabilities were modeled. The data analysis and visualization were performed following the guidelines provided on the CellChat GitHub page (https://github.com/sqjin/CellChat).

5.12. Gene Set Enrichment Analysis

In this study, gene sets related to inflammatory responses, phagocytosis and cytokine production were collected from relevant literature [2, 86] and MSigDB (https://www.gsea‐msigdb.org/gsea/msigdb/mouse/genesets.jsp). The irGSEA package (version 3.1.2) was employed to perform gene set enrichment scoring on the scRNA‐seq data. This method evaluated the activity of various cell populations within specific gene sets, revealing heterogeneity in cellular functions and states.

5.13. Differential Gene Expression and GO Enrichment Analysis

Differential gene expression analysis was performed using the Seurat “FindMarkers” function at default parameters. Genes with P_adjust < 0.05 (Bonferroni correction) and |log₂ (fold change)| ≥ 1 were selected as significantly DEGs. GO enrichment analysis was performed using ClusterProfiler (version 4.14.6). Representative GO terms that satisfy P_adjust < 0.05 (Benjamini‐Hochberg correction) were selected for illustration. P values were adjusted for multiple testing using Bonferroni correction for DEGs and Benjamini‐Hochberg correction for GO enrichment analysis. The volcano plots (Figure 4J,K), bar plots (Figure 4G–I), and bubble plots (Figure 2I) were plotted by http://www.bioinformatics.com.cn.

5.14. Statistical Analysis

Statistical analyses were performed using GraphPad Prism 8 (GraphPad software) and R software (version 4.3.1). Statistical analysis performed for each experiment are indicated either in the respective methods or in the figure legends. P < 0.05 was considered statistically significant (* p < 0.05; ** p < 0.001).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll72130‐sup‐0001‐SuppMat.docx

Shan Q., Dong Z., Li N., et al. “Deciphering the Heterogeneity of Pulmonary Macrophages in Response to Fine Particles.” Small 22, no. 10 (2026): e07293. 10.1002/smll.202507293

Data Availability Statement

The scRNA‐seq data generated in this study have been deposited in the Genome Sequence Archive in National Genomics Data Center (CRA031812). The custom code used for analysis is available from the corresponding author upon reasonable request. All other data supporting the findings of this study are available within the article and its supplementary information files.