Abstract

   Mediating protein citrullination, peptidyl arginine deiminase 2 (PAD2)
   has recently been reported to influence macrophage phenotypes. However,
   the mechanisms of PAD2 on macrophage function in Pseudomonas aeruginosa
   (PA)‐induced acute lung injury syndrome (ALI) remains unclear.
   Utilizing single‐cell RNA sequencing and mass spectrometry‐based
   proteomics, a new citrullination site at arginine 171 (R171) is
   discovered within nuclear factor‐ κB (NF‐κB) p65 catalyzed by PAD2,
   which modulates PAD2‐NF‐κB p65‐importin α3 pathway and its downstream
   M1/M2 macrophage polarization. Building on these findings, a
   cell‐specific targeted therapeutic strategy using gold nanoparticles
   (AuNPs) conjugated with a novel PAD2 inhibitor, AFM41a, and an
   intercellular adhesion molecule‐1 (ICAM‐1) antibody is developed. This
   approach enables the selective delivery of the inhibitor to
   M1‐polarized macrophages in the PA‐infected alveolar niche. In vivo,
   this nanomedicine reduces excessive inflammation and promotes M1‐to‐M2
   polarization to inhibit ALI. This study highlights the role of
   PAD2‐mediated citrullination in macrophage polarization and introduces
   a promising nanoparticle‐based therapy for PA‐induced ALI.

   Keywords: acute lung injury, gold nanoparticles, ICAM‐1, macrophage
   polarization, NF‐κB p65, PAD2
     __________________________________________________________________

   Peptidyl arginine deiminase 2 (PAD2) plays a key role in regulating
   macrophage function in Pseudomonas aeruginosa‐induced acute lung injury
   (ALI). Using single‐cell RNA sequencing and proteomics, a new
   PAD2‐catalyzed citrullination site on NF‐κB p65 (171 Arginine),
   modulating macrophage polarization is identified. A nanoparticle‐based
   PAD2 inhibitor therapy is developed, reducing inflammation and
   improving ALI outcomes.

   graphic file with name ADVS-12-2413253-g004.jpg

1. Introduction

   Acute lung injury and acute respiratory distress syndrome (ALI/ARDS)
   are marked by diffuse alveolar damage and inflammation, posing
   significant global health and resource challenges.^[ [46]^1 , [47]^2 ^]
   Pneumonia caused by bacterial and viral infections remains a major risk
   factor for the development ALI.^[ [48]^3 , [49]^4 ^] Pseudomonas
   aeruginosa (PA), a Gram‐negative pathogen with increasing multidrug
   resistance (MDR), is frequently associated with hospital‐acquired
   pneumonia, particularly in immunocompromised patients, leading to ALI
   and high mortality rates.^[ [50]^5 , [51]^6 ^] Due to the high
   prevalence of MDR in PA infections and the lack of effective
   treatments,^[ [52]^7 , [53]^8 ^] it is necessary to develop novel
   non‐antibiotic therapeutics for PA‐induced ALI.

   Peptidylarginine deiminases (PAD, also known as PADI) is an enzyme
   family involved in the citrullination of arginine residues on proteins,
   a post‐translational modification that alters protein function and
   cellular processes.^[ [54]^9 ^] Among them, PAD2 is abundantly
   expressed in myeloid‐derived immune cells, and elevated levels of PAD2
   have been reported in the bronchoalveolar lavage fluid (BALF) and serum
   of ALI patients.^[ [55]^10 , [56]^11 ^] Our recent studies showed that
   the absence of PAD2 improves survival and alleviates ALI in septic
   mice.^[ [57]^11 , [58]^12 ^] In vitro experiments revealed that using a
   pan‐PAD inhibitor, BB‐Cl‐amidine, or silencing the Pad2 gene can
   attenuate the polarization of THP‐1 cells toward the M1 phenotype while
   promoting activation toward the M2 phenotype.^[ [59]^13 ^] However, the
   role of PAD2‐mediated macrophage polarization in PA‐induced ALI has not
   yet been studied in vivo.

   An in vitro study demonstrated that PAD2 can directly citrullinate
   nuclear factor‐ κB (NF‐κB) p65, leading to reduced expression of
   NF‐κB‐targeted pro‐inflammatory mediators in Pad2 deficient macrophages
   following lipopolysaccharide (LPS) stimulation.^[ [60]^14 ^] In
   contrast, another in vitro study presented conflicting evidence,
   suggesting that overexpression of PAD2 may inhibit NF‐κB activity by
   citrullinating the inhibitor of κB (IκB) kinase γ subunit, which
   facilitates NF‐κB dissociation from IκBα and subsequent nuclear
   translocation.^[ [61]^15 ^] Without in vivo studies, the exact
   mechanisms by which PAD2 modulates the NF‐κB pathway and influences
   macrophage phenotypes remain poorly understood, limiting our
   understanding of these processes in PA‐induced ALI and their potential
   for advancing therapeutic strategies.

   To address these issues, we utilized single‐cell RNA sequencing
   (scRNA‐seq) to characterize immune cell heterogeneity in the alveolar
   microenvironment of PA‐induced ALI models and explored PAD2‐dependent
   cellular and gene changes, specifically focusing on macrophage
   polarization. Additionally, we employed mass spectrometry‐based
   proteomics analysis to map specific PAD2‐mediated citrullinated
   proteins and citrullination modification sites.^[ [62]^16 , [63]^17 ^]
   Importantly, we identified a specific citrullination site within NF‐κB
   p65 catalyzed by PAD2, which modulates NF‐κB p65 activity and
   consequently affects the M1/M2 polarization transition.

   Building on these observations, we developed a dual‐function
   anti‐inflammatory nanomedicine for treating ALI. Previous studies
   showed that treatment with AFM32a, a small‐molecule PAD2 inhibitor,
   could alleviate lung injury and improve survival in murine models of
   LPS‐induced and cecal ligation puncture‐induced sepsis.^[ [64]^11 ,
   [65]^18 ^] However, in vivo administration of small molecules like
   AFM32a is often limited by rapid metabolism, poor bioavailability, and
   high off‐target toxicity.^[ [66]^19 ^] To address these challenges, we
   functionalized each AuNP to simultaneously carry a novel PAD2 inhibitor
   AFM41a^[ [67]^20 ^] and intercellular adhesion molecule‐1 (ICAM‐1)
   antibodies, enabling the AuNPs to cross the pulmonary endothelial
   barrier and target the PA‐infected alveolar niche.^[ [68]^21 , [69]^22
   ^] AuNPs are potent tools for targeted drug delivery due to their
   biocompatibility, non‐toxicity, and flexible surface functionalization
   properties.^[ [70]^23 , [71]^24 ^] AuNPs have also been shown to reduce
   mortality, renal dysfunction, and liver injury by promoting M1 to M2
   polarization.^[ [72]^25 , [73]^26 ^] Our ICAM1‐congugated AuNP
   nanomedicine successfully delivered the PAD2 inhibitor to M1‐polarized
   alveolar cells, offering a promising approach for the treatment of ALI.

2. Results

2.1. Single‐Cell Characterization of Nlrp3^+Icam1^+ Myeloid Cells in Septic
Lung

   Using the 10× Genomics scRNA‐seq platform, we analyzed 22193 BALF cells
   from PA‐induced ALI and sham models in both WT and Pad2^−/− mice
   (Figure [74]1A; Figure [75]S1A, Supporting information). Through
   nonlinear dimensionality reduction on uniform manifold approximation
   and projection (UMAP) plots, we delineated a diverse immune cell
   landscape for both WT and Pad2^−/− groups following integration
   analysis. Ten distinct clusters of BALF cells, including alveolar
   macrophages (AMs), myeloid cells, fibroblasts, Clara cells, dendritic
   cells (DCs), T cells, interstitial macrophages (IMs), NK cells, B
   cells, and red blood cells, were identified using the Seurat clustering
   algorithm (Figure 1B; Figure [76]S1B, Supporting information). Cell
   type identification within each cluster relied on the significant
   expression of well‐characterized marker genes (Figure [77]S1C,
   Supporting information). UMAP visualization (Figure [78]1C) and cell
   ratio analysis (Figure [79]1D) illustrate the predominance of AMs in
   the sham condition (84.47% in WT sham group, 84.45% in Pad2^−/− sham
   group), with myeloid cells emerging as the major population post‐PA
   infection (87.94% in WT PA group and 89.67% in Pad2^−/− PA group).

Figure 1.

   Figure 1
   [80]Open in a new tab

   scRNA‐seq analysis of bronchoalveolar lavage fluid (BALF) cells
   comparing wild‐type and Pad2^−/− mice. A. Schematic diagram showing
   scRNA‐seq pipeline of murine BALF cells WT and Pad2^−/− mice during
   PA‐pneumonia induced ALI. BALF cells from 3 mice were pooled into one
   sample for each group including WT sham, WT PA, Pad2^−/− sham, and
   Pad2^−/− PA groups. Adapted with permission from BioRender. Yu, X.
   (2025). B. Nonlinear dimensionality reduction using Uniform Manifold
   Approximation and Projection (UMAP) analysis of 22193 BALF cells from
   WT and Pad2^−/− mice reveals ten distinct clusters following
   unsupervised clustering. Each dot represents an individual cell, with
   colors indicating cluster assignment. C. Experimental Group‐based UMAP
   visualization of WT sham, WT PA, Pad2^−/− sham, and Pad2^−/− PA. This
   visualization highlights the distinct immune landscapes present before
   and after PA infection, shifting from an alveolar macrophage‐dominant
   to a myeloid cell‐dominant population. D. Distribution of cell subtype
   proportions among all cell populations in each experimental group. E.
   Pseudotime and single‐cell trajectory analysis of Chil3, Mrc1, Nlrp3
   and Icam1 genes by Monocle. F. Heatmap of M1/M2 genes, ordered based on
   their common kinetics through pseudotime using Monocle. G. The upper
   panel shows expression visualization of Chil3 and Mrc1 genes across all
   cluster populations, along with their co‐expression patterns across
   distinct cell clusters. The lower panel shows expression of Nlrp3 and
   Icam1 visualization of genes across all cluster populations, along with
   their co‐expression patterns across distinct cell clusters. Density
   plots represent the relative gene expression levels.

   To explore the potential relationship between AMs, IMs, and myeloid
   cells in the sham and PA groups, we analyzed the dynamic immune states
   and cell transitions within the macrophage/myeloid cell compartment
   using the Monocle algorithm.^[ [81]^27 ^] Pseudotime analysis revealed
   that AMs from the sham group were predominantly localized at the
   starting branch of the trajectory, while myeloid cells from the PA
   group were situated at the terminal end (Figure [82]S1D, Supporting
   information). IMs, present in both sham and PA conditions, were mainly
   positioned at the third branch (Figure [83]S1D, Supporting
   information). Gene expression analysis of macrophages and myeloid cells
   revealed distinct ALI‐associated phenotypic and functional changes. AMs
   from the sham group exhibited upregulation of M2‐associated genes,
   Chil3 and Mrc1 (Figure [84]1E,F). Following PA‐induced ALI, myeloid
   cells showed increased expression of M1‐related genes, including Nlrp3,
   Icam1, Il1b, Cxcl2, Nos2, Il12a and Mmp9,^[ [85]^28 ^] along with
   decreased expression of Chil3 and Mrc1(Figure [86]1E,F). To further
   investigate shifts in macrophage polarization between the sham and PA
   conditions, we evaluated the expression of Chil3, Mrc1, Nlrp3 and Icam1
   across all cell populations. Consistent with the pseudotime analysis,
   Chil3^+Mrc1^+Nlrp3 ^− Icam^− cells were primarily AMs from the sham
   group, whereas Chil3^−Mrc1^− Nlrp3^+Icam1^+ cells were mainly myeloid
   cells from the PA group (Figure [87]1G).

2.2. Pad2 Knockout Alters Myeloid Cell Composition and Function in ALI

   Given that the Nlrp3^+Icam1^+ myeloid cell compartment represents the
   largest emergent population of alveolar immune cells after PA‐induced
   ALI, we performed an unsupervised cluster analysis to investigate the
   heterogeneity among all myeloid cells. Five transcriptionally distinct
   subcluster of myeloid cells (Clusters 0–4) were identified (Figure
   [88]2A). A heterogeneous cell distribution was observed among WT PA and
   Pad2^−/− PA groups across these subclusters (Figure [89]2B,C). Cluster
   0 myeloid cells emerged as a prominent population post‐PA infection,
   characterized by the expression of Il12a, Il1a, Ccl4, and Csf3 genes,
   indicative of a pro‐inflammatory M1‐like phenotype. Cluster 1 myeloid
   cells displayed elevated levels of Mrc1, Anxa5, Ear2, and S100a10,
   aligning with an M2‐like anti‐inflammatory phenotype. Cluster 2 myeloid
   cells were characterized by the expression of Ngp, Mmp9, Camp, Cd177,
   and Ly6c2 genes, suggesting a neutrophil‐like phenotype involved in
   inflammation and antimicrobial defense. Cluster 3 myeloid cells
   expressed Ddx3y, Eif2s3y, and Eya3 genes associated with RNA metabolism
   and protein synthesis in cell development. Finally, Cluster 4 myeloid
   cells were distinguished by high expression of S100a1, Ctsk, Bcar3, and
   Lima1, which are linked to cell adhesion and migration (Figure [90]2D).

Figure 2.

   Figure 2
   [91]Open in a new tab

   Subcluster analysis of myeloid cells from WT and Pad2^−/− mice after
   PA‐induced ALI. A. UMAP analysis of five distinct sub‐clusters
   (Clusters 0, 1, 2, 3, and 4) of myeloid cells from WT and Pad2^−/− PA
   mice, with each cluster color‐coded for identification. B. Experimental
   group‐based UMAP visualization of myeloid cells from WT PA and Pad2^−/−
   PA groups. C. Distribution of cell‐subtype proportions among myeloid
   cell populations in WT PA and Pad2^−/− PA groups. D. Heatmap showing
   gene expression within the myeloid cell compartment from WT and
   Pad2^−/− PA mice in the PA‐induced ALI condition, annotated by cluster
   types. E. Expression profiles of pro‐inflammatory genes (Il6, Il1b,
   Nos2, Cxcl1, Il12a, S100a8, S100a9, Ngp, Ltf, Stfa2l1, Cxcr2, Mmp9,
   Ifitm1, Trem3, Csf3r, Retnlg, Slc2a3, and Ly6g5b) and anti‐inflammatory
   genes (Arg1, Mrc1, Chil3, Retnla, and Il10) across myeloid cell
   clusters, illustrated by dot plots. F. Enrichment analysis of
   representative GO biological pathways for Cluster 0 myeloid cells,
   illustrated by dot plots. G. Expression profiles of NF‐κB pathway genes
   across myeloid cells from WT PA and Pad2^−/− PA groups, illustrated by
   dot plots.

   Next, we compared the differences in the relative proportions of the
   subclusters among myeloid cells. Notably, the deletion of Pad2 resulted
   in a marked reduction of M1‐like Cluster0 myeloid cells (44.7% in WT PA
   vs 22.9% in Pad2^−/− PA) and an increase in M2‐like Cluster1 myeloid
   cells (25.5% in WT PA vs 35.8% in Pad2^−/− PA) (Figure [92]2C).
   Figure [93]2E showed the expression of M1/M2‐related genes among all
   myeloid cell clusters, with Cluster 0 myeloid cells displaying a
   distinct M1‐polarization signature, characterized by the highest
   M1‐related genes and the lowest M2‐polarization genes. Gene Ontology
   (GO) enrichment analysis identified activated pathways in Cluster 0
   myeloid cells, suggesting positive regulation of inflammatory responses
   (Figure [94]2F). Notably, NF‐κB signaling (NIK/NF‐kappaB signaling and
   I‐kappaB kinase/NF‐kappaB signaling) was among these activated
   pathways, confirmed as an important mechanism for promoting M1
   polarization state.^[ [95]^29 ^] Additionally, GO enrichment analysis
   for Clusters 1, 2, 3, and 4 is shown in Figure [96]S2A, Supporting
   information. The heatmap further demonstrates that NF‐κB signaling
   genes are expressed at higher levels in Cluster 0 myeloid cells
   compared to other myeloid cell clusters (Figure [97]S2B, Supporting
   information). Furthermore, the expression of genes critical for NF‐κB
   signaling was decreased in Pad2^−/− PA compared to WT PA group
   (Figure [98]2G).

2.3. Effect of PAD2 Inhibition and Knockout on Macrophage Polarization in
PA‐Induced ALI

   Our scRNA‐seq analysis revealed that myeloid cells exhibited notable
   plasticity and polarization characteristics in the immune response.^[
   [99]^30 , [100]^31 ^] To assess the impact of Pad2 deletion on M1/M2
   polarization function, we conducted a comparative analysis of M1 and
   M2‐related gene expression across all BALF cells between WT and
   Pad2^−/− mice. Following PA‐induced ALI, our scRNA‐seq analysis
   revealed an upregulation of M2‐related genes (Chil3, Mrc1, Il10, Arg1,
   Msr1, Retnla) in Pad2^−/− mice compared to their WT counterparts.
   Conversely, Pad2^−/− mice showed a downregulation of M1‐related genes
   (Il1a, Il1b, Tnf, Cxcl2, Ccl4, Il12a, Cxcl1, Il6) compared to WT mice
   after PA infection (Figure [101]3A).

Figure 3.

   Figure 3
   [102]Open in a new tab

   PAD2 modulates M1/M2 polarization in PA‐induced ALI. A. Gene expression
   levels of M2 markers (Chil3, Mrc1, Il10, Arg1, Msr1, Retnla) and M1
   markers (Il1a, Il1b, Tnf, Cxcl1, Cxcl2, Ccl4, Il12a, Il6) across all
   BALF cells, illustrated in a dot plot for WT PA and Pad2^−/− PA group
   comparison. B. Western blot analysis for M2 proteins (Ym1 and CD206)
   and M1 proteins (NLRP3 and pro‐IL‐1β) in BALF cell lysates from WT and
   Pad2^−/− mice 24 h post‐PA inoculation (n = 3). Relative protein
   expression levels are shown in the right panel. C. ELISA results
   showing levels of IL‐1β, TNF‐α, Ym1, and IL‐10 in BALF from WT and
   Pad2^−/− mice, across sham conditions and 24 h post‐PA inoculation (n =
   3–6). D. Western blot analysis comparing the expression of NLRP3, iNOS,
   and pro‐IL‐1β proteins in BMDMs from WT and Pad2^−/− mice, treated with
   PA‐derived LPS (250 ng mL^−1, 24 h) (n = 3). Relative protein
   expression levels are shown on the right panel. E. ELISA result showing
   IL‐1β levels in BMDMs culture supernatant from WT and Pad2^−/− mice,
   treated with PA‐derived LPS (250 ng mL^−1, 24 h). Cell culture
   supernatants were analyzed to determine the concentrations of released
   mature IL‐1β cytokine (n = 3). F. Western blot analysis of NLRP3, iNOS,
   and pro‐IL‐1β proteins in RAW 264.7 cells. Cells were stimulated with
   250 ng mL^−1 PA‐derived LPS or without PA as control, and
   simultaneously treated with or without 1 µm AFM41a for 24 h (n = 3).
   Relative protein expression levels are shown on the right panel. G.
   ELISA result showing IL‐1β levels in RAW264.7 cell culture supernatant.
   Cell culture supernatants were analyzed to determine the concentrations
   of released mature IL‐1β cytokine (n = 3). Results in B–G are
   representative of at least 3 independent experiments. Data for bar
   charts in B were analyzed using unpaired Student's t‐tests. Data for
   bar charts in C–E were analyzed using two‐way analysis of variance
   (ANOVA). Data for bar charts in F and G were analyzed using one‐way
   ANOVA. Data are presented as means ± SEM. Asterisks (*) denote
   statistical significance, with p‐values indicated as follows: * p <
   0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

   Subsequent validation through Western blot analysis of BALF cell
   lysates from WT and Pad2^−/− mice suggested a regulatory role of PAD2
   in M1/M2 polarization. Specifically, the Pad2^−/− PA group showed
   elevated levels of Ym1 and CD206 proteins and lower levels of NLRP3 and
   IL‐1β proteins (Figure [103]3B). ELISA assays were conducted to analyze
   immune biomarkers associated with M1 and M2 phenotypes in BALF from WT
   and Pad2^−/− mice in both sham and PA groups. Our results revealed the
   activation of both pro‐inflammatory and anti‐inflammatory markers after
   PA‐induced ALI (Figure [104]3C). The levels of M1‐related IL‐1β and
   TNF‐α were attenuated in Pad2^−/− PA mice, whereas those of M2‐related
   Ym1 and IL‐10 were significantly higher in this group (Figure [105]3C).
   To investigate the influence of Pad2 deletion in vitro, we treated bone
   marrow‐derived macrophages (BMDMs) from WT and Pad2^−/− mice with
   PA‐derived LPS for 24 h. Significant upregulation of M1‐related NLRP3,
   IL‐1β, and iNOS proteins was observed following LPS treatment, with
   Pad2 KO significantly reducing their levels (Figure [106]3D,E).
   Furthermore, we used a novel PAD2 inhibitor, AFM41a, to explore whether
   inhibiting PAD2 activity could affect macrophage polarization.^[
   [107]^32 ^] We stimulated mouse RAW 264.7 cells with LPS, in the
   presence or absence of AFM41a, and found that inhibiting PAD2 activity
   with AFM41a also decreased M1‐polarization protein expression,
   including NLRP3, IL‐1β and iNOS (Figure [108]3F,G).

2.4. Nuclear Translocation Modulated by PAD2‐Dependent Citrullination of
NF‐κB p65

   Our investigation demonstrates that PAD2 plays a crucial role in
   modulating the M1/M2 polarization transition in ALI. NF‐κB is the major
   transcription factor responsible for activating M1 macrophage
   polarization.^[ [109]^29 ^] Given PAD2's ability to citrullinate
   proteins and influence their activity, we hypothesized that PAD2 might
   interact with NF‐κB via citrullination, thereby affecting NF‐κB pathway
   activity. To test this, we prepared BALF cell lysates from WT and
   Pad2^−/− mice and subjected them to in‐gel digestion and liquid
   chromatography‐tandem mass spectrometry (LC‐MS/MS) analysis (Figure
   [110]4A). Proteomics and citrullinomics analysis of NF‐κB pathway
   proteins revealed citrullinated NF‐κB p65, specifically identifying the
   citrullinated peptide, DPAGRPLLLTPVLSHPIFDNR, which contains R171
   (Figure [111]4B). This citrullinated peptide was detected in the WT
   group but not in the Pad2^−/− group. The citrullination of R171 was
   confirmed by observing a ≈1 Da heavier citrullinated peptides in WT
   mice (2328.28 Da) compared to the uncitrullinated peptides in Pad2^−/−
   mice (2327.28 Da) in the MS spectrum (Figure [112]S3, Supporting
   information). To validate the interaction between PAD2 and p65, BMDMs
   from WT and Pad2^−/− mice were stimulated with PA‐derived LPS for 1 h.
   Our results showed that anti‐ NF‐κB p65 antibody was
   co‐immuno‐precipitated PAD2 from the WT group, whereas this
   co‐immunoprecipitation was inhibited in the Pad2^−/− group
   (Figure [113]4C).

Figure 4.

   Figure 4
   [114]Open in a new tab

   Identification of PAD2‐dependent citrullination site within NF‐κB p65.
   A. Schematic diagram showing the citrullinome pipeline of murine BALF
   cells from the WT and Pad2^−/− group. Adapted with permission from
   BioRender. Yu, X. (2025). B. Mass spectrometry identification of
   citrullinated peptides at arginine 171 (R171) in NF‐κB p65 from BALF
   cells of WT mice. The citrullination site is highlighted in green (*).
   Notably, no citrullinated form of the peptides was identified detected
   from BALF cells of Pad2^−/− mice, indicating the PAD2‐dependent
   citrullination at the R171 residue. BALF cells from 3 mice were pooled
   into one sample. C. BMDMs from WT and Pad2^−/− mice were collected and
   stimulated with PA‐derived LPS (250 ng mL^−1, 1 h), and cells were
   lysed for protein extraction, followed by immunoprecipitation with
   anti–p65 antibody or control IgG. The immunoprecipitants were probed
   with anti‐PAD2 by Western blotting. D. BMDMs from WT and Pad2^−/− mice
   were stimulated with PA‐derived LPS (250 ng mL^−1, 1 h) and control.
   Nuclear and cytosol extracts were prepared separately and examined by
   Western blotting using a p65 antibody, normalized against β‐actin
   (cytosol extracts) and Lamin A/C (nuclear extracts) (n = 4–5). Relative
   protein expression levels are shown in the right panel. E.
   Immunofluorescence analysis of NF‐κB p65 expression in BMDMs from WT
   and Pad2^−/− mice treated with PA‐derived LPS (250 ng mL^−1, 1 h)
   compared to control (n = 5), with cells stained for NF‐κB p65 (in red)
   and nuclei counterstained with DAPI (in blue) under a confocal
   microscope. Relative protein expression levels are shown in the right
   panel. Scale bars represent 50 µm. Results in C–E are representative of
   at least 3 independent experiments. Data for all bar charts were
   analyzed using two‐way ANOVA. Data are presented as means ± SEM.
   Asterisks (*) denote statistical significance, with p‐values indicated
   as follows: * p < 0.05; ** p < 0.01; *** p < 0.001.

   Upon external stimuli, cytosolic NF‐κB p65 becomes activated and
   translocated into the nucleus to regulate the expression of
   pro‐inflammatory mediators.^[ [115]^33 ^] To study this process, we
   utilized BMDMs from both WT and Pad2^−/‐ mice and isolated cytosolic
   and nuclear extracts separately to assess the levels of NF‐κB p65.
   Following 1‐h PA‐derived LPS stimulation, we observed decreased
   expression of cytosolic NF‐κB p65, accompanied by its translocation
   into the nucleus, resulting in a notable increase in nuclear p65.
   Conversely, Pad2 deletion attenuated the accumulation of nuclear NF‐κB
   p65 while maintaining cytosolic levels after LPS stimulation
   (Figure [116]4D). Immunofluorescence analysis of NF‐κB p65 expression
   in BMDMs from WT and Pad2^−/− mice further confirmed that Pad2 deletion
   impairs the nuclear localization of NF‐κB p65 following LPS stimulation
   (Figure [117]4E).

2.5. Citrullination of NF‐κB p65 at Arginine 171 Enhances its Interaction
with Importin α3

   NF‐κB p65 remains inactive in the cytoplasm until activated, upon which
   it is rapidly imported into the nucleus by importin family proteins.^[
   [118]^34 , [119]^35 ^] To determine which importin proteins mainly
   facilitate NF‐κB p65 translocation into the nucleus during PA‐induced
   ALI, we examined the expression of importin family genes in our
   scRNA‐seq data. Among BALF cells from WT and Pad2^−/− mice after
   PA‐induced ALI, only the Kpna4 gene (encoding importin α3) exhibited
   significantly elevated expression levels (Figure [120]5A).
   Additionally, within our myeloid‐derived cell cluster, including
   myeloid cells, IMs, AMs, and DCs, Kpna4 gene displayed notably higher
   expression compared to other importin isoforms (Figure [121]5B).

Figure 5.

   Figure 5
   [122]Open in a new tab

   Citrullination of arginine 171 site in NF‐κB p65 modulates its
   interaction with importin α3. A. Gene expression levels of Kpna1,
   Kpna2, Kpna3, Kpna4, Kpna6, Kpna7, Kpnb1, Ipo4, Ipo5, Ipo7, Ipo8, Ipo9,
   Ipo11 and Ipo13 across all BALF cells, depicted in a violin plot for WT
   sham and WT PA group comparison. B. Gene expression level of Kpna1,
   Kpna2, Kpna3, Kpna4, Kpna6, Kpna7, Kpnb1, Ipo4, Ipo5, Ipo7, Ipo8, Ipo9,
   Ipo11 and Ipo13 across myeloid cells, interstitial macrophages,
   alveolar macrophages, and dendritic cells, illustrated by violin plots.
   C. Schematic illustration of the establishment of R171A mutant RAW
   264.7 cells using lentivirus transfection. The process involves using a
   three‐plasmid packaging system to transfect 293T cells, producing and
   collecting the transgenic lentivirus, and then transducing RAW 264.7
   cells. Adapted with permission from BioRender. Yu, X. (2025). D. R171A
   mutant RAW 264.7 macrophage cells and the corresponding WT cells were
   stimulated with PA‐derived LPS (250 ng mL^−1, 1 h), and were lysed for
   protein extraction, followed by immunoprecipitation with anti–importin
   α3 or control IgG. The immunoprecipitants were probed with anti‐NF‐κB
   p65 by Western blotting. E. R171A mutant RAW 264.7 macrophage cells and
   the WT cells were stimulated with PA‐derived LPS (250 ng mL^−1, 1 h)
   and control. Nuclear and cytosol extracts were prepared separately and
   examined by Western blotting using p65 antibody, normalized against
   β‐actin and Lamin A/C (n = 4–5). Relative protein expression levels are
   shown in the right panel. F. Western blot analysis of NLRP3, iNOS, and
   pro‐IL‐1β proteins in R171A mutant RAW 264.7 cells and WT cells,
   stimulated with PA‐derived LPS (250 ng mL^−1, 24 h) and control (n =
   3). Relative protein expression levels are shown in the right panel.
   Results in D‐F are representative of at least 3 independent
   experiments. Data for all bar charts were analyzed using two‐way ANOVA.
   Data are presented as means ± SEM. Asterisks (*) denote statistical
   significance, with p‐values indicated as follows: * p < 0.05; ** p
   < 0.01.

   Given our findings that PAD2 citrullinates NF‐κB p65 at the arginine
   171 site and regulates its nuclear translocation, we hypothesized that
   citrullination by PAD2 facilitates its interaction with importin α3,
   thereby facilitating nuclear translocation efficiency. To test this
   hypothesis, we generated a stable p65‐R171A mutant RAW 264.7 cell line
   using plasmids injection and lentivirus transfection (Figure [123]5C).
   The R171A mutation replaced the arginine (R) at position 171 with
   alanine (A), making it incapable of being citrullinated by PAD2 (Figure
   [124]S4A, Supporting information). Co‐immunoprecipitation experiments
   confirmed that anti‐importin α3 could co‐immuno‐precipitate NF‐κB p65
   in WT cells, whereas this interaction was inhibited in R171A mutant
   cells (Figure [125]5D). These findings demonstrate that PAD2‐mediated
   citrullination of NF‐κB p65 at R171 is crucial for its interaction with
   importin α3, thereby enhancing nuclear translocation.

   To further investigate, we separately isolated cytosolic and nuclear
   extracts from WT and R171A mutant cells post‐LPS stimulation. Our
   results demonstrated that the R171A mutant inhibited the translocation
   of p65 from the cytosol to the nucleus, reducing nuclear translocation
   efficiency (Figure [126]5E). Given that NF‐κB p65 activated M1
   polarization upon inflammatory stimulation, we treated WT and R171A
   mutant cells with PA‐derived LPS for 24 h. We observed a significant
   increase in the expression levels of NF‐κB‐targeted M1 polarization
   proteins, including NLRP3, IL‐1β, and iNOS, upon LPS stimulation in WT
   cells. In contrast, the R171A mutant decreased the expression of these
   M1 polarization proteins (Figure [127]5F). Additionally, we collected
   the supernatant from these macrophage cultures and assessed the release
   of pro‐inflammatory cytokines, including IL‐1β, IL‐6, and TNF‐α, which
   are also regulated by NF‐κB p65. ELISA results revealed that the R171A
   mutant reduced the expression of these pro‐inflammatory cytokines after
   LPS stimulation and inhibited their release from the cells (Figure
   [128]S4B, Supporting information).

2.6. In Vitro Attenuation of Inflammatory Response via Targeted
Gold‐Nanoparticle‐Mediated PAD2 Inhibition

   To investigate the therapeutic effect of PAD2 on macrophage
   polarization, we developed a nanodrug delivery strategy involving
   cell‐specific PAD2 inhibition targeting ICAM‐1, a surface marker highly
   expressed on M1 macrophages as identified in our scRNA‐seq results
   (Figure [129]1E–G). We selected 30 nm gold nanoparticles (AuNPs)
   (Figure [130]S5, Supporting information) as the drug delivery carrier
   due to their biocompatibility and non‐toxicity, properties previously
   demonstrated to mitigate inflammation in mouse models.^[ [131]^23 ,
   [132]^24 ^] A self‐assembled polyethylene glycol (PEG) layer was added
   to the nanoparticle to passivate the surface and prevent aggregation.
   These nanoparticles were co‐conjugated with ICAM‐1 antibody or IgG
   control via a covalent bond, and with a novel PAD2 inhibitor AFM41a
   through electrostatic interactions (Figure [133]6A).

Figure 6.

   Figure 6
   [134]Open in a new tab

   ICAM‐1 antibody‐loaded AuNPs enhance uptake by macrophages and deliver
   drugs in vitro. A. Schematic illustration of molecular structures of
   anti‐ICAM1‐AFM41a‐loaded AuNPs and anti‐IgG‐AFM41a‐loaded AuNPs.
   Adapted with permission from BioRender. Yu, X. (2025). B. Schematic
   illustration of macrophage polarization state after PA‐derived LPS
   stimulus and cellular uptake of anti‐ICAM1‐AuNPs. The
   anti‐ICAM1‐AFM41a‐loaded AuNPs can be taken up by macrophages and
   directly promote the transition from M1 to M2 polarization. Adapted
   with permission from BioRender. Yu, X. (2025). C. Violin plots
   depicting the expression levels of Icam1 gene in BALF cells from WT
   sham and WT PA mice. D. Western blot analysis of ICAM‐1 protein in RAW
   264.7 cells stimulated with PA‐derived LPS (250 ng mL^−1, 24 h) and
   control. Relative ICAM‐1 protein expression level is shown in the right
   panel (n = 3). E. RAW 264.7 cells were stimulated with PA‐derived LPS
   (250 ng mL^−1, 6 h) or left untreated as a control. Cells were
   subjected to immunofluorescence staining with PE anti‐mouse ICAM‐1
   antibody to evaluate the expression of ICAM‐1. BF: bright field. Scale
   bars represent 50 µm. F. RAW 264.7 cells (left panel) and BMDMs (right
   panel) were pre‐treated with PA‐derived LPS (250 ng mL^−1, 6 h) or left
   untreated as control, then treated with PE‐anti‐ICAM1‐AFM41a‐loaded
   AuNPs or PE‐anti‐IgG‐AFM41a‐loaded AuNPs (20 µg mL^−1, 3 h). Cells were
   analyzed by immunofluorescence to assess the uptake of AuNPs. A
   zoomed‐in view reveals the distribution of AuNPs within the cells.
   Scale bars represent 100 µm. G. RAW 264.7 cells were pre‐treated with
   or without PA‐derived LPS (250 ng mL^−1, 6 h), then treated with
   PE‐anti‐ICAM‐1‐AFM41a‐loaded AuNPs (20 µg mL^−1, 18 h). ELISA analyzed
   cell culture supernatants to show the concentrations of IL‐1β and IL‐6
   cytokines (n = 5). Results in D‐G were representative of at least three
   independent experiments. Data for bar charts in D were analyzed using
   unpaired Student's t‐tests. Data for bar charts in G were analyzed
   using one‐way ANOVA. Data are presented as means ± SEM. Asterisks (*)
   denote statistical significance, with p‐values indicated as follows: *
   p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

   To evaluate the loading performance of anti‐ICAM‐1 or control IgG
   antibodies and the AFM41a inhibitor on the nanoparticles, we used
   dynamic light scattering (DLS) and zeta potential measurements to
   determine nanoparticle size and surface charge before and after
   conjugation. A significant increase in particle size following antibody
   conjugation confirmed the successful attachment of the larger ICAM‐1
   antibody or IgG control via covalent bonding to the AuNP surface
   (Figure [135]S6A–C). Additionally, the zeta potential shift from mildly
   negative to mildly positive indicates the successful electrostatic
   binding of the positively charged small molecule inhibitor AFM41a
   (Figure [136]S6D). The polydispersity index (PDI) remained ≈0.2,
   suggesting a narrow size distribution and minimal aggregation. To
   further quantify the level of antibody loading, we performed a
   bicinchoninic acid (BCA) assay (Figure [137]S7, Supporting
   Information), which indicated that ≈93.1 antibody molecules were
   conjugated per AuNP, confirming efficient surface functionalization.
   The AFM41a drug loading efficiency was evaluated using high‐performance
   liquid chromatography (HPLC). A standard curve was established based on
   the AFM41a peak area, allowing the quantification of the drug loading
   efficiency by comparing the AFM41a concentrations in the supernatant
   before and after AuNP conjugation. The HPLC‐based analysis yielded an
   average AFM41a loading of 7005 molecules per AuNP (Drug/NPs; Figure
   [138]S8, Supporting Information).

   Leveraging ICAM‐1 overexpression in inflamed endothelial cells and
   macrophages during ALI, the anti‐ICAM‐1 antibody directs AuNP uptake
   into macrophages via receptor‐mediated endocytosis. Inside the
   lysosome, high ionic strength and acidic pH (≈4.5–5.0) protonate
   PEG‐carboxyl groups, weakening electrostatic interactions and
   facilitating AFM41a release. AuNP accumulation induces osmotic swelling
   and physical stress (e.g., PEG degradation, aggregation), promoting
   drug escape into the cytosol, where PAD2 is primarily located. AFM41a
   inhibits PAD2 activity and the citrullination process. This targeted
   delivery blocks the PAD2‐NF‐κB p65‐importin α3 pathway, thereby
   mitigating inflammation and promoting the shift from M1 to M2
   polarization (Figure [139]6B). Following PA‐induced ALI, scRNA‐seq
   analysis showed a notable upregulation of Icam1 expression in BALF
   cells compared to sham mice (Figure [140]6C). To further validate ICAM1
   role in macrophages, RAW 264.7 macrophages were stimulated with
   PA‐derived LPS, which resulted in a marked increase in ICAM‐1
   expression. This finding was confirmed by both Western blot and
   immunofluorescence results (Figure [141]6D,E).

   To visualize the macrophage uptake of the nanodrug in vitro, we
   leveraged the localized surface plasmon resonance (LSPR) of the AuNPs
   to enhance the fluorescence signal of nearby fluorophores. In our
   assay, individual AuNPs (LSPR peak = 526 nm, Figure [142]S9A,
   Supporting information) aggregate within the macrophage cytoplasm after
   being engulfed. We selected phycoerythrin (PE) fluorophores with an
   emission wavelength matching the LSPR peak wavelength of the aggregated
   AuNPs to facilitate plasmon‐enhanced fluorescence (PEF).^[ [143]^36 ^]
   Both the ICAM1‐AFM41a‐AuNPs (nanodrug conjugates) and their
   non‐targeted IgG‐AFM41a‐AuNPs controls were tagged with a secondary
   PE‐anti‐Rat‐IgG antibody complex, and we confirmed their strongly
   enhanced fluorescence intensity upon aggregation prior to application
   (Figure [144]S9B, Supporting information).

   After LPS stimulation, BMDMs were incubated with ICAM1‐AFM41a‐AuNPs at
   concentrations of 1, 5, 10, 20, and 40 µg mL^−1 (nanodrug‐treated
   group), or without ICAM1‐AFM41a‐AuNPs (untreated group). Both groups
   were imaged after 3 h (see Experimental Section). We observed a
   remarkable granulate‐like fluorescence signal (indicating AuNP
   aggregates) in the cytoplasm of the nanodrug‐treated BMDMs, with the
   signal increasing proportionally with the dosing concentration of the
   ICAM1‐AFM41a‐AuNPs nanodrug (Figure [145]S9C, Supporting information).
   AuNP uptake was also evaluated between the ICAM‐1 target group
   (ICAM1‐AFM41a‐AuNPs) and the IgG control group (IgG‐AFM41a‐AuNPs) with
   LPS stimulation, to assess M1‐targeted delivery in macrophages. The
   anti‐ICAM‐1 target group showed a notably stronger uptake compared to
   the IgG control group, in both RAW264.7 cells and BMDMs
   (Figure [146]6F). To assess the anti‐inflammatory effect of the
   ICAM1‐AFM41a‐AuNPs, we collected supernatant from each group 24 h after
   LPS stimulation. Treatment with AuNPs led to a marked reduction in the
   levels of pro‐inflammatory cytokines IL‐1β and IL‐6 in both the
   anti‐ICAM1 target and IgG control groups compared to the untreated
   group. Importantly, the anti‐ICAM1 target group exhibited a
   significantly greater reduction in IL‐1β and IL‐6 production compared
   to the IgG control group (Figure [147]6G).

2.7. Therapeutic Efficacy of ICAM1‐AFM41a‐AuNPs in PA‐Induced ALI

   To evaluate the targeted delivery and therapeutic efficacy of
   ICAM1‐AFM41a‐AuNPs nanodrug in a PA‐induced ALI model, we conducted an
   in vivo experiment as outlined in Figure [148]7A. Immunohistochemical
   (IHC) staining was performed to assess ICAM‐1 expression in lung
   tissues from both sham and PA‐treated mice. ICAM‐1 was primarily
   expressed in pulmonary vascular endothelial cells under sham
   conditions, whereas its expression was upregulated in activated
   alveolar myeloid cells following PA‐induced ALI (Figure [149]7B).
   Additionally, IHC staining for ICAM‐1 was performed on the heart,
   liver, spleen, and kidney in PA‐induced ALI mice. Notably, ICAM‐1
   expression was higher in the lung compared to other organs, supporting
   the targeted delivery of anti‐ICAM1‐Ab‐conjugated‐AuNPs to lung tissue
   (Figure [150]S10, Supporting information). To further investigate the
   effect of AuNPs in vivo, mice were inoculated with PA to induce ALI,
   followed by intravenous administration of ICAM1‐AFM41a‐AuNPs,
   IgG‐AFM41a‐AuNPs, and free AFM41a 4 h after PA infection. Lungs were
   harvested for microscopic imaging 8 h after PA infection. We visualized
   nanoparticle distribution within the whole lung tissue and within
   frozen sections of the lung through plasmon‐enhanced fluorescence. The
   intensity of the enhanced fluorescence was sufficient to mask the
   autofluorescence background of the lung tissue. The frozen sections
   show that the ICAM1‐AFM41a‐AuNPs nanodrug could transport from the
   bloodstream to the alveolar airspace, where most macrophages and PA
   bacteria are located (Figure [151]7C). Multiple sites within the lung
   tissue showed significantly higher accumulation and stronger
   fluorescence emission in ICAM1‐AFM41a‐AuNPs‐treated mice compared to
   IgG‐AFM41a‐AuNPs‐treated controls, demonstrating that the
   ICAM1‐AFM41a‐AuNPs nanodrug specifically targeted the injured lung
   areas.

Figure 7.

   Figure 7
   [152]Open in a new tab

   Therapeutic performance of the ICAM‐1 antibody‐loaded AuNPs in
   PA‐induced ALI. A. Experimental procedures for evaluating the targeting
   and therapeutic efficacy of AuNPs in PA‐induced ALI. 4 h after PA
   inoculation, free AFM41a, PE‐anti‐IgG‐AFM41a‐loaded AuNPs, and
   PE‐anti‐ICAM1‐AFM41a‐loaded AuNPs were administered intravenously.
   Lungs were collected for imaging 8 h post‐infection. Blood, BALF, and
   lung tissue were collected 24 h post‐infection. Adapted with permission
   from BioRender. Yu, X. (2025). B. Immunohistochemical (IHC) staining
   for ICAM‐1 in lung tissues of sham and PA‐induced ALI mice. A zoomed‐in
   view shows ICAM1^+ alveolar myeloid cell from PA group. Scale bars
   represent 50 µm. C. The immunofluorescence of PE‐labeled anti‐IgG AuNPs
   and anti‐ICAM1‐AFM41a‐loaded AuNPs were measured using whole lung
   tissue (left panel) and lung frozen sections (right panel) to analyze
   lung‐targeted efficacy of AuNPs. The red arrows indicate the locations
   of AuNPs. Scale bars represent 50 µm. The statistical analysis results
   of fluorescence intensity are shown in the lower right panel. D.
   Histopathological examination of lung injury using H&E‐stained lung
   tissue sections from mice with free AFM41a, PE‐anti‐IgG‐AFM41a‐loaded
   AuNPs, PE‐anti‐ICAM1‐AFM41a‐loaded AuNPs, or without treatment 24 h
   after PA infection. Scale bars represent 50 µm. E. ELISA results
   showing concentrations of IL‐1β, IL‐6, TNF‐α, CXCL‐10, and CCL‐3 in the
   BALF of mice treated with free AFM41a, PE‐anti‐IgG‐AFM41a‐loaded AuNPs,
   PE‐anti‐ICAM1‐AFM41a‐loaded AuNPs, or without treatment mice 24 h
   post‐PA infection (n = 3–6). F. Western blot analysis for NLRP3, CD206,
   iNOS, Ym1 and NF κB p65 proteins in BALF cell lysates from mice treated
   with free AFM41a, PE‐anti‐IgG‐AFM41a‐loaded AuNPs,
   PE‐anti‐ICAM1‐AFM41a‐loaded AuNPs, or without treatment mice 24 h
   post‐PA infection (n = 5). Relative protein expression levels are shown
   in the right panel. G. IHC staining for Ym1 in lung tissues of mice
   treated with PE‐anti‐IgG‐AFM41a‐loaded AuNPs or
   PE‐anti‐ICAM1‐AFM41a‐loaded AuNPs within the PA‐induced ALI group (n =
   3). Scale bars represent 50 µm. Results in B‐G are representative of at
   least three independent experiments. Data for bar charts in C were
   analyzed using unpaired Student's t‐tests. Data for bar charts in E and
   F were analyzed using one‐way ANOVA. Data are presented as means ± SEM.
   Asterisks (*) denote statistical significance, with p‐values indicated
   as follows: * p < 0.05; ** p < 0.01; *** p < 0.001, **** p < 0.0001.

   Further analysis included histological and cytokine assessments of the
   lung tissues and BALF 24 h after infection. Histological examination
   revealed a significant reduction in inflammatory cell infiltration,
   bacterial spread, and cellular debris in the ICAM1‐AFM41a‐AuNPs‐treated
   mice compared to both IgG‐AFM41a‐AuNPs‐treated and free AFM41a‐treated
   mice, indicating enhanced therapeutic efficacy of ICAM1‐AFM41a‐AuNPs
   (Figure [153]7D). Cytokine analysis demonstrated decreased levels of
   pro‐inflammatory markers: IL‐1β, IL‐6, TNF‐α, CXCL‐10, and CCL‐3 in the
   BALF from ICAM1‐AFM41a‐AuNPs‐treated mice, affirming the
   anti‐inflammatory effect of the nanodrug (Figure [154]7E). BALF cells
   were further analyzed for the expression of M1 markers (NLRP3, iNOS)
   and M2 markers (CD206, Ym1) by Western blot. As shown in
   Figure [155]7F, treatment with free AFM41a and IgG‐AFM41a‐AuNPs reduced
   NLRP3 and iNOS expression but did not significantly affect CD206 and
   Ym1 levels after PA infection. In contrast, treatment with
   ICAM1‐AFM41a‐AuNPs significantly decreased NLRP3 and iNOS expression
   while increasing CD206 and Ym1 expression. Additionally, total NF‐κB
   p65 protein levels remain unchanged across all groups (Figure [156]7F).
   IHC staining of Ym1 confirmed that ICAM1‐AFM41a‐AuNPs treatment
   increased Ym1 expression in alveolar cells, indicating enhanced M2
   polarization (Figure [157]7G).

3. Discussion

   The composition and function of alveolar immune cells provide insights
   into the pathogenesis of ALI.^[ [158]^6 , [159]^37 ^] Our analysis of
   single‐cell transcriptional profiles of alveolar immune cells reveals
   significant transitions in macrophage polarization phenotypes in
   PA‐induced ALI. Specifically, we observed co‐activation of
   Nlrp3^+Icam1^+ M1‐like cells during PA‐induced ALI, replacing the
   M2‐like cells present in unstimulated alveoli (Figure [160]1E–G). This
   finding is supported by recent studies suggesting that NLRP3 is a key
   regulator of macrophage‐induced inflammation and M1 polarization.^[
   [161]^38 , [162]^39 ^] ICAM‐1 is well‐known for its role in regulating
   leukocyte adhesion and recruitment from circulation to sites of
   inflammation in vascular endothelial cells. However, ICAM‐1 expression
   is also induced in macrophages upon inflammatory stimulation, and its
   function and mechanisms in macrophages are less well understood.^[
   [163]^40 , [164]^41 ^] Our data demonstrate that ICAM‐1 is notably
   upregulated on macrophages both in vivo and in vitro upon
   proinflammatory stimulus (Figures [165]1, [166]6). Several recent
   studies have implicated ICAM‐1 in the regulation of macrophage
   polarization.^[ [167]^28 , [168]^42 , [169]^43 ^] In line with these
   studies, our findings suggest that the activation of Nlrp3^+Icam1^+
   myeloid cells contributes to M1 polarization and exacerbated
   inflammation in the pathobiology of PA‐induced ALI.

   Subcluster analysis of Nlrp3^+Icam1^+ myeloid cells revealed that
   C0_myleoid cells exhibited an M1‐polarization state and activated NF‐κB
   signaling (Figure [170]2E,F). The NF‐κB signaling pathway,
   predominantly composed of the p50/p65 heterodimer, mediates the
   transcriptional expression of NLRP3, ICAM‐1, and proinflammatory
   cytokines, promoting M1 polarization.^[ [171]^44 , [172]^45 , [173]^46
   ^] Additionally, we observed a shift from M1 to M2 polarization
   phenotype following Pad2 deficiency (Figure [174]3A). Our study
   indicates that both Pad2 gene knockout and PAD2 inhibition can mitigate
   excessive M1 polarization‐associated inflammation in vivo and in vitro
   (Figure [175]3B–G). Pad2 deletion attenuated the activity of the NF‐κB
   pathway in ALI (Figure [176]2G). Previous studies have demonstrated
   that decreased NF‐κB activity can contribute to an M2 polarization
   shift and reduce lung injury.^[ [177]^47 , [178]^48 , [179]^49 ^] Based
   on these findings, we hypothesized that PAD2 might directly
   citrullinated NF‐κB proteins, thereby promoting M1 polarization and
   inflammation. Through MS‐based proteomics analysis of alveolar cells,
   we identified PAD2‐dependent citrullination of NF‐κB p65 at arginine
   171 (Figure [180]4B; Figure [181]S3, Supporting information). The
   reduced translocation of cytosolic p65 into the nucleus due to Pad2
   deficiency after PA further validated our hypothesis
   (Figure [182]4A–C).

   To explore PAD2's role in citrullinating the R171 site, we generated
   R171A mutant RAW 264.7 cells alongside WT controls. Notably, mutations
   at the R171 site reduced p65's interaction with importin α3
   (Figure [183]5D), impairing its nuclear translocation efficiency in
   LPS‐stimulated macrophages (Figure [184]5E). These findings demonstrate
   that PAD2‐mediated citrullination at R171 is crucial for the downstream
   interaction between p65 and importin α3, thereby facilitating p65's
   nuclear localization. Previous studies have shown that PAD4 can
   directly citrullinate p65 at residues R35, R50, R73, and R149 in
   neutrophils, activating NF‐κB and prolonging inflammation in vitro.^[
   [185]^44 ^] In contrast, our study reveals the critical role of PAD2 in
   citrullinating p65 at R171 in macrophages, influencing their
   polarization. Notably, PAD2 has been shown to significantly improve
   survival rates in various sepsis‐induced ALI models compared to PAD4.^[
   [186]^11 , [187]^12 ^] Our study contributes to understanding the
   underlying mechanisms of PAD2's role in ALI immunology, particularly
   through in vivo studies.

   Our study supports the idea that shifting M1‐polarized cells toward
   anti‐inflammatory M2 polarization in alveoli through targeted delivery
   of PAD2 inhibitor‐AFM41a can alleviate PA‐induced ALI. Conventional
   therapeutics for intravenous small‐molecule AFM41a inhibitors face
   major drawbacks, including systemic distribution that leads to
   insufficient drug concentrations at the site of lung injury and a lack
   of cell‐specific targeting ability.^[ [188]^19 , [189]^50 ^] To address
   these issues, we demonstrated that ICAM‐1 antibody‐loaded AuNPs can
   specifically deliver the AFM41a inhibitor to Icam1^+ M1‐polarized cells
   in the injured alveoli. During PA‐induced ALI, both activated alveolar
   cells and lung vasculature highly express ICAM‐1 (Figure [190]7B;
   Figure [191]S10, Supporting information),^[ [192]^21 , [193]^22 ^]
   making it an effective mediator for transporting AuNPs from the
   bloodstream to the lung airspace. Once in the alveoli, the ICAM‐1
   antibody targets Icam1^+ M1‐polarized cells, allowing AFM41a to inhibit
   PAD2's citrullination activity and the associated PAD2‐NF‐κB
   p65‐importin α3 pathway. Our result showed that ICAM‐1‐coated AuNPs
   significantly increased enhanced the delivery of AFM41a into the
   injured alveoli compared to the IgG control (Figure [194]7C).

   Moreover, previous studies have shown that AuNPs can attenuate
   inflammation and promote M2 polarization,^[ [195]^25 , [196]^26 ^]
   making them a dual‐function M2‐polarizing nanomedicine when combined
   with the novel AFM41a inhibitor. Our study demonstrated that
   ICAM1‐AFM41a‐AuNPs efficiently target M1‐polarized alveolar cells and
   induce a notable M1‐to‐M2 polarization shift in PA‐induced ALI,
   compared to non‐targeted AuNPs and free AFM41a treatment
   (Figure [197]7E–G). Therefore, this specific M1‐like
   macrophage‐targeted AuNP therapy holds promise for mitigating the
   hyperinflammatory response in ALI. However, its efficacy in human ALI
   remains uncertain, and the safety and effectiveness of
   ICAM1‐AFM41a‐AuNPs must be validated in clinical trials. Additionally,
   due to the low abundance of citrullination events in whole‐cell
   extracts,^[ [198]^51 , [199]^52 , [200]^53 ^] we identified only one
   citrullination site at R171 by comparing data from WT and Pad2^−/−
   mice. It remains unclear whether other arginine residues are also
   citrullinated by PAD2 and whether R171 is the most critical site.
   Further investigation into all potential citrullination sites on p65 is
   necessary and could reveal additional therapeutic targets.

   In summary, our scRNA‐seq analysis reveals that PA‐induced ALI
   activates the NF‐κB pathway in alveolar cells, creating a persistent
   M1‐like pro‐inflammatory microenvironment. PAD2‐mediated citrullination
   of NF‐κB p65 at arginine 171 enhances its interaction with importin α3,
   facilitating nuclear localization and promoting NF‐κB‐mediated M1
   polarization. Importantly, our findings demonstrate that targeting
   M1‐polarized macrophages with ICAM‐1 antibody‐loaded AuNPs delivering
   the novel AFM41a inhibitor, effectively shifts M1 macrophages toward an
   M2 phenotype. This approach offers a promising therapeutic strategy for
   ALI.

4. Experimental Section

Animals

   Wild‐type male C57BL/6J and Pad2^−/− mice (8–12 weeks old) were used in
   this study. Wild‐type mice were purchased from the Jackson Laboratory
   (Bar Harbor, ME, USA). Pad2^−/− mice were kindly provided by Dr Scott
   Coonrod (Cornell University, Ithaca, New York, USA). The mice were
   housed in a specific pathogen‐free environment with a controlled
   temperature of 23 ± 1.5 °C and relative humidity of 70 ± 20%. All mice
   used for in vivo and in vitro experiments were age‐ and sex‐matched.

   All animal experiments were performed with approval from the University
   of Michigan (PRO00011567).

Acute Lung Injury Model

   Mice aged 8–12 weeks were exposed to intranasal administration of PA
   (19660; ATCC) solution to induce PA‐induced ALI. Briefly, a PA solution
   was prepared at a concentration of 8.25 × 10^7 colony‐forming units
   (CFU) mL^−1 in PBS (Gibco). The mice were anesthetized with ketamine
   (Dechra Veterinary Products, 80 mg kg^−1) and xylazine (Akorn, 20
   mg kg^−1), and then held vertically. Subsequently, PA solution (15 µL)
   was instilled into each nostril, totaling 30 µL to achieve a final
   bacterial load of 2.5 × 10^6 CFU. Mice inoculated with sterile PBS (30
   µL) served as sham controls. All mice were euthanized by CO[2] 24 h
   after inoculation.

Sample Preparation and Single‐Cell RNA Sequencing

   BALF samples were collected from both WT and Pad2^−/− mice after PA
   infection and sham control. BALF samples from three mice per group were
   pooled to generate samples (n = 3 mice per sample) for the isolation of
   BALF cells. The BALF samples were centrifuged for 5 min at 400 rcf at 4
   °C. The cell pellet was then resuspended in red blood cell lysis buffer
   (500 µL, 00‐4333‐57; eBioscience) and incubated for 5 min at room
   temperature. Following this, cold PBS (500 µL, Gibco) was added to
   dilute the RBC lysis buffer, and the mixture was centrifuged for 5 min
   at 400 rcf at 4 °C. The resulting BALF cells were resuspended in PBS
   containing 1% weight/volume FBS (Gibco), and cell viability was
   determined using automated cell counters (Invitrogen). The single‐cell
   suspension was thoroughly mixed and loaded onto a 10× Chromium system
   to capture no more than 10 000 single cells using the Chromium Next GEM
   Single Cell 3′ GEM, Library & Gel Bead Kit (10× Genomics). The cells
   were partitioned into Gel Beads in the Chromium instrument. DNA
   amplification and library construction were performed with cell lysis
   and barcoded reverse transcription of RNA. The resulting libraries were
   sequenced using an Illumina HiSeq 4000 next‐generation sequencing
   platform. Data quality analysis and mapping to ensemble gene symbols
   were conducted using CellRanger (10x Genomics).

scRNA‐Seq Data Processing

   The CellRanger output data were imported into the Seurat R package
   (version 5.0.1) for unsupervised clustering analysis. Prior to
   clustering, filtering procedures were implemented to eliminate
   multiplets and damaged cells, while sources of variation deemed
   uninformative were regressed out. Identification of variable genes was
   achieved through iterative selection based on the dispersion versus
   average expression profile of each gene. Normalization and scaling of
   gene expression values within individual cells were conducted using the
   SCTransform algorithm. Dimensionality reduction and visualization of
   the data were performed using Principal Components Analysis^[ [201]^54
   ^] and Uniform Manifold Approximation and Projection (UMAP),
   incorporating the top 30 principal components. Parameters for UMAP were
   set to min. dist. = 0.3 and n. neighbor = 30. Subsequently, cells were
   clustered using an unsupervised clustering approach with default
   parameters for the Seurat package (resolution = 0.6). Cluster‐specific
   marker genes were identified utilizing the FindAllMarkers function in
   Seurat, with criteria set at a p‐value < 0.01 and log (fold change)
   > 0.25 within the target cluster. Visualization of gene expression
   patterns across cell clusters was accomplished using UMAP plots and
   heatmaps generated with functions available in the Seurat package.
   Differentially expressed genes (DEGs) were determined using the
   FindAllMarkers function with default parameters, specifying log (fold
   change) > 0.26, p‐value < 0.01, and min.pct > 0.1. For Gene Ontology
   (GO) enrichment analysis, DEGs across different cell types were
   utilized to conduct cluster‐specific pathway enrichment analysis.
   Myeloid cell re‐clustering, UMAP visualization, DEG analysis, and GO
   enrichment analysis were performed with dims = 1:15, resolution = 0.2,
   and the same default parameters as previously described.

   For single‐cell trajectory analysis, Monocle v.2.30.0 was utilized to
   explore developmental trajectories among alveolar macrophages,
   interstitial macrophages, and myeloid cells.^[ [202]^27 , [203]^55 ^]
   The scRNA‐seq data were scaled, normalized, and clustered in Seurat,
   then downsampled to 1500 cells and imported into a Monocle object. DEGs
   for each cluster were identified using the FindVariableFeatures
   function, and the top 2000 genes with the lowest q‐values were selected
   for pseudotime analysis. Cell trajectories were visualized using the
   plot_cell_trajectory function, with the trajectory's starting point
   (root_state) defined by cells from the sham condition.
   Plot_genes_in_pseudotime function and plot_pseudotime_heatmap function
   were used to show the dynamic trend of the selected significant gene
   expression levels.

Cell Culture

   Bone marrow‐derived macrophages (BMDMs) were isolated from tibiae and
   femurs of both WT and Pad2^−/− mice for in vitro experiments. Bone
   marrow cells were harvested and cultured in 75 mm^2 Petri dishes
   containing IMDM (Gibco) supplemented with 20% FBS, 1%
   penicillin/streptomycin (Lonza Inc), and 30% L929 cell (CCL‐1; ATCC)
   supernatant at 37 °C/5% CO[2] incubator. The L929 cell supernatant was
   generated by incubating L929 cell fibroblasts in IMDM with 10% FBS for
   6 days to produce macrophage colony‐stimulating factor (M‐CSF). After 7
   days, BMDMs were collected and resuspended in Opti‐MEM (Gibco) to the
   desired concentrations. Mouse RAW 264.7 cells were cultured in DMEM
   (Gibco) supplemented with 10% FBS and 1% penicillin/streptomycin at 37
   °C/5% CO[2] incubator. For experiments involving LPS stimulation, cells
   were challenged by PA‐deprived LPS (250 ng/mL, L9143; Sigma–Aldrich) in
   Opti‐MEM, while control cells were treated with Opti‐MEM alone for the
   same duration. For experiments assessing AFM41a inhibitor treatment,
   cells were exposed to PA‐deprived LPS (250 ng mL^−1) with or without
   AFM41a (1 µm) inhibitor followed by a 24‐h incubation.

Preparation of AuNPs

   Custom‐synthesized 30 nm spherical gold nanoparticles functionalized
   with NHS‐PEG‐SH were ordered from Nanopartz (DC11‐30). The surface
   charge was adjusted to a desired range of −25 to −30 mV to enable
   electrostatic binding with the positively charged AFM41a inhibitor. A
   2.6 mg quantity of NHS‐PEG‐terminated AuNPs was first dissolved in PBS
   (0.5 mL) and sonicated for 1 min to form a uniform colloidal solution.
   Next, sterilized low endotoxin, azide‐free (LEAF) mouse anti‐ICAM‐1
   antibody (200 µg, 116132; BioLegend) or rat IgG2b control (200 µg,
   408210; BioLegend), along with AFM41a inhibitor (10 µg), were added to
   create a total solution volume of 1 mL. The mixture was vortexed and
   incubated for 30 min at 30 °C, followed by centrifugation (12 000 rcf,
   5 min). The supernatant was removed, and the pellet was resuspended in
   0.01× PBS with 0.1% Tween 20. This washing step was repeated three
   times. After the final centrifugation, the pellet was resuspended in 1x
   PBS and stored at 4 °C. For imaging study, the ICAM‐1 or IgG conjugated
   AuNPs were then stained with PE‐anti‐rat IgG2b antibody (4 µg mL^−1,
   408213; BioLegend) for 30 min at room temperature (25 °C) and subjected
   to the same washing protocol described above.

Binding of AuNPs to Macrophages In Vitro

   BMDMs and RAW 264.7 cells were seeded in culture slides (354108;
   Corning) and incubated overnight. The cells were then treated with or
   without PA‐derived LPS (250 ng mL^−1) for 6 h, followed by three washes
   with PBS. Cells were subsequently incubated with PE‐fluorescent labeled
   ICAM‐1Abs‐coated AFM41a‐loaded AuNPs or PE‐fluorescent labeled IgG2b
   Abs‐coated AFM41a‐loaded AuNPs sample at 37 °C/5% CO[2] incubator. For
   imaging, after 3 h of incubation with AuNPs, cells were washed three
   times with 1x PBS, and fixed with 4% paraformaldehyde for 30 min at
   room temperature. Prolong Gold anti‐fade reagent with DAPI
   ([204]P36931; Invitrogen) was added to the cells, and a coverslip was
   applied. The cells were imaged by a KEYENCE BZ‐X800 microscope. For
   cytokines analysis, after 18 h of incubation with AuNPs, cells were
   washed three times with PBS, and the culture supernatants were
   collected by centrifugation (300 rcf, 5 min) for subsequent analysis.

In Vivo Delivery of AuNPs

   An ALI mouse model was generated as previously described. 4 h post‐PA
   inoculation, the mice were grouped randomly and intravenously (i.v.)
   with either PE‐fluorescent labeled ICAM‐1 Abs‐coated AFM41a‐loaded
   AuNPs (1 mg per kg of body weight), PE‐fluorescent labeled IgG2b
   Abs‐coated AFM41a‐loaded AuNPs (1 mg kg^−1) or free AFM41a (20
   mg kg^−1), or PBS as a control. At predetermined time intervals, lungs
   were collected at 4 h post i.v. and both lung tissue and frozen lung
   sections (10 µm) were prepared for imaging using KEYENCE BZ‐X800
   microscope. 24 h post‐PA inoculation, the mice were euthanized, and
   BALF and lungs were collected for further analysis.

Measurement of Cytokines and Chemokines

   The levels of IL‐1β, TNF‐α, IL‐10, Ym1, IL‐6, CXCL‐10, and CCL‐3 in
   BALF and supernatant were measured by the core of UMICH Immune
   Monitoring Shared Resource using the core‐developed sandwich ELISA.

Western Blotting

   For Western blotting, cells were washed two times with ice‐cold PBS and
   lysed with RIPA buffer (89900; Thermo Fisher Scientific) plus halt
   protease inhibitor cocktail (87787; Thermo Fisher Scientific) for
   30 min on ice. After addition of 4× Laemmli sample buffer (Bio‐Rad),
   samples were separated by 10% SDS‐PAGE electrophoresis, transferred to
   nitrocellulose membrane (Bio‐Rad), and proteins of interest were
   incubated overnight with diluted primary antibodies (CD206, 1:1000,
   24595, Cell signaling Technology; NLRP3, 1:1000, 15101, Cell signaling
   Technology; iNOS, 1:1000, PA5‐17106, Invitrogen; Ym1, 1:1000, AF2446,
   R&D system; ICAM‐1, 1:1000, ab222736, Abcam; IL‐1β, 1:1000, 12703, Cell
   signaling Technology; NF‐κB p65, 1:1000, 8242, Cell signaling
   Technology; Lamin A/C, 1:1000, 2032, Cell signaling Technology and
   β‐actin, 1:2500, 4970, Cell signaling Technology) and then incubated
   for 1 h with a secondary antibody (Invitrogen). The membranes were
   visualized using enhanced chemiluminescence (1705061; Bio‐Rad) within
   the luminescent image analyzer (Thermo Fisher Scientific).

Immunohistochemical Analysis

   Formalin‐fixed, paraffin‐embedded lung tissue sections were stained
   with the Anti‐Ym1 antibody (ab230610; Abcam), and paraffin‐embedded
   heart, liver, liver, spleen, lung, and kidney tissue sections were
   stained with the Anti‐ICAM‐1 antibody (ab222736; Abcam) by the In‐Vivo
   Animal Core at the University of Michigan. Digital images of the
   stained tissue sections were acquired using a KEYENCE BZ‐X800
   microscope.

Immunocytochemistry Analysis

   For NF‐κB p65 nuclear/cytosolic localization analysis, BMDMs were
   seeded on culture slides (354105, Corning) and incubated overnight.
   Subsequently, BMDMs were blocked with a solution containing 5% BSA
   (Thermo Fisher Scientific) and 0.05% Triton X‐100 (Thermo Fisher
   Scientific) in PBS for 1 h. Immunostaining was conducted using an
   anti‐NF‐κB p65 antibody (1:100, 8242; Cell signaling Technology)
   overnight at 4 °C, followed by incubation with a second antibody
   (ab150064; Abcam). Nuclear staining was achieved using DAPI (62248;
   Thermo Fisher Scientific). Microscopic imaging was performed using a
   Nikon A1si confocal microscope. For ICAM‐1 expression analysis, RAW
   264.7 cells were subjected to blocking and then incubated with PE
   anti‐mouse CD54 Antibody (1:100, 116107; Biolegend) overnight at 4 °C.
   The sections were rinsed twice to remove any unbound antibodies before
   imaging with a KEYENCE BZ‐X800 microscope.

Immunoprecipitation

   Immunoprecipitation was conducted in accordance with the manufacturer's
   instructions provided in the classic magnetic IP/Co‐IP kit manual
   (88804, Thermo Fisher Scientific). Cells were washed once with ice‐cold
   PBS and were lysed in IP lysis buffer with protease inhibitor cocktail,
   incubated on ice for 10 min, and centrifuged at 13000 rcf at 4 °C for
   10 min. The supernatant was incubated with indicated antibodies (2 µg
   per sample) overnight at 4 °C and then incubated with Protein A/G
   Agarose for 1 h at room temperature. The bound proteins were eluted by
   incubating with elution buffer for 10 min and then boiling in 4×
   loading buffer and subjected to Western blotting.

Nuclear‐Cytosolic Extraction

   Nuclear and cytosolic extracts were prepared as per the manufacturer's
   protocol using the NE‐PER kit (78833, ThermoFisher Scientific).
   Briefly, cells were lysed in Cytoplasmic Extraction Reagent I (CER I,
   200 µL) buffer supplemented with a protease inhibitor cocktail to
   extract cytosolic proteins. Following the addition of Cytoplasmic
   Extraction Reagent II (CER II, 11 µL) buffer, the cytosolic fraction
   was obtained by centrifugation at 16 000 rcf for 10 min at 4 °C. The
   resulting pellet containing nuclei was then resuspended in Nuclear
   Extraction Reagent (NER, 100 µL) buffer supplemented with a protease
   inhibitor cocktail. Subsequent centrifugation at 16 000 rcf for 10 min
   at 4 °C yielded the nuclear protein extract in the supernatant.

Identification of Citrullination Sites by LC‐MS/MS

   BALF samples from three mice per group were pooled to obtain samples (n
   = 3 mice per sample) for BALF cell isolation. The samples were
   centrifuged at 400 rcf at 4 °C for 5 min to pellet the cells, which
   were then treated with RBC lysis buffer (eBioscience) to remove red
   blood cells. The isolated BALF cells were suspended in RIPA buffer
   containing a protease inhibitor cocktail and heated at 95 °C for 15
   min. Protein concentration was determined using Qubit fluorometry
   (Invitrogen). Subsequently, 10 µg of each sample was subjected to
   SDS‐PAGE using 10% Bis‐Tris NuPage Mini‐gel (Invitrogen) with the MES
   buffer system. Gel electrophoresis was performed for 2 cm, and the gel
   was then cut into ten equally sized bands. In‐gel digestion of each
   band was carried out with trypsin using a robot (DigestPro, CEM),
   involving washing with 25 mm ammonium bicarbonate followed by
   acetonitrile, reduction with 10 mm dithiothreitol at 60 °C, alkylation
   with 50 mm iodoacetamide at room temperature, and digestion with
   sequencing grade trypsin (Promega) at 37 °C for 4 h. The digestion
   process was quenched with formic acid, and the supernatant was directly
   analyzed. Half of each digested sample was subjected to nano LC‐MS/MS
   using a Waters M‐Class LC system coupled to a ThermoFisher Exploris 480
   mass spectrometer. Peptides were loaded onto a trapping column and
   eluted over a 75 µm analytical column packed with XSelect CSH C18 resin
   (Waters). The column temperature was maintained at 55 °C using a column
   heater (Sonication), and the mass spectrometer operated in
   data‐dependent mode with the Orbitrap set at 60000 FWHM and 15000 FWHM
   for MS and MS/MS, respectively, with a 3‐s cycle for both MS and MS/MS.
   Each sample was analyzed for 5 h. Mascot with Trypsin/P enzyme
   parameters was used for data search against the SwissProt Mouse
   database appended with Pseudomonas aeruginosa BL04. A differential
   modification of 0.984 on arginine was specified to account for
   citrullination by PAD2. Data were filtered and analyzed using Scaffold
   5 proteome software (version: 5.3.3).

Establishment of Mouse R171A RAW 264.7 Stable Cell Lines

   Subcloning, lentivirus production, transduction, and cell selection
   were performed by the University of Michigan Vector Core. A geneblock
   encoding mouse Rela, with arginine 171 mutated to alanine (R171A) and
   an HA tag fused to the C‐terminus through a 4‐amino acid linker, was
   synthesized by Twist Biosciences. This Rela R171A‐HA geneblock was then
   cloned into the NheI/XhoI‐digested pLentiLox CAG‐mcs‐mPGK‐meGFP T2a
   Puromycin lentiviral vector. DNA preparation for lentiviral production
   was carried out using the Qiagen Plasmid Plus Midiprep DNA kit (Qiagen)
   following the manufacturer's instructions. Lentivirus packaging vectors
   psPAX2 and pC1‐VSVG were co‐transfected with either pLentilox CAG‐Rela
   R171A‐HA (for R171A mutant) or mPGK‐GFP/puro (for WT) plasmid using
   standard PEI precipitation methods. PEI precipitation involved
   incubating the plasmids with PEI (molecular weight 2500, Polysciences,
   Inc) in OptiMEM (Life Technologies) at room temperature for 20 min
   before adding to DMEM with 10% FBS media. This DNA/PEI‐containing media
   was added to transfect 293T cells (ATCC). Culture supernatant was
   collected after 72 h, supplemented with 8 µg mL^−1 polybrene (Sigma),
   and applied to RAW 264.7 cells. Cells were then incubated at 37 °C, 5%
   CO[2] for 48 h before undergoing puromycin selection for 5 days.

Quantification of AFM41a Using High‐Performance Liquid Chromatography (HPLC)

   HPLC was performed using a Thermofisher Vanquish Core LC system
   equipped with a 3 µm C18 reversed‐phase column (4.6 × 150 mm) and a
   UV–Vis detector set at λ = 220 nm for AFM41a detection. AFM41a was
   dissolved in water (1 mg mL^−1), and HPLC samples were prepared by
   mixing 800 µL of the aqueous sample with 200 µL of acetonitrile. The
   mobile phase was an isocratic mixture of 20% acetonitrile and 80% water
   (ACN–H₂O) with 0.1% trifluoracetic acid, ensuring optimal separation of
   AFM41a. The flow rate was maintained at 0.5 mL min^−1, and the column
   temperature was held at 30 °C. Chromatograms were analyzed to assess
   peak area, and a standard curve was generated by titrating different
   AFM41a concentrations to establish a linear correlation between peak
   area and drug concentration. This standard curve was then used to
   quantify drug loading efficiency by measuring AFM41a concentrations in
   the supernatant before and after conjugation to AuNPs. The difference
   between pre‐ and post‐conjugation supernatant concentrations was used
   to calculate the amount of AFM41a successfully loaded onto the AuNP
   surface.

Statistical Analysis

   The statistical analysis was performed using GraphPad Prism 8 software.
   For comparisons involving three or more groups, the one‐way or two‐way
   ANOVA with Turkey's multiple comparisons test was applied to compare
   individual groups. For comparisons between the two groups, an unpaired
   t‐test was utilized. A significance level of p‐value < 0.05 was set to
   determine statistical significance, indicating meaningful differences
   between the groups.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   X.Y. and Y.S. contributed equally to this work. X.Y., Y.S., K.K., and
   Y.L. conceptualized the study. X.Y., Y.S., T.D., O.W., L.S., L.B.,
   P.R., and Y.L. developed the methodology. X.Y., Y.S., and Y.L.
   performed the formal analysis. X.Y., Y.S., T.D., O.W., L.S., J.M.,
   M.Z., and C.Q carried out the investigation. X.Y. and Y.S. wrote the
   original draft. X.Y., Y.S., Y.L. K.K., and J.M. wrote, reviewed, and
   edited the final draft. X.Y., Y.S., K.K., and Y.L. visualized the
   study. K.K. and Y.L. supervised the study. Y.L. and K.K. administered
   the project. Y.L. acquired funds.

Supporting information

   Supporting Information
   [205]ADVS-12-2413253-s001.pdf^ (4MB, pdf)

Acknowledgements