Perivascular mesenchymal cells instruct ST2+ reparative macrophages to promote endovascular injury-induced neointimal hyperplasia in mice

Ping, Yan; Qin, Zhewen; Huang, Xingxiao; Sun, Qiuli; Yao, Tao; Li, Siqi; Tian, Xin; Xia, Xintong; Diao, Bowen; Zhao, Lin; Zhang, Xinyuan; Wei, Danni; Wang, Long; Shao, Mengle; Wu, Ying; Huang, Jinyu; Shan, Bo

doi:10.1038/s41467-026-68587-x

Download PDF

Article
Open access
Published: 07 March 2026

Perivascular mesenchymal cells instruct ST2+ reparative macrophages to promote endovascular injury-induced neointimal hyperplasia in mice

Yan Ping^1,2,
Zhewen Qin^1,3,
Xingxiao Huang^1,2,
Qiuli Sun^1,2,
Tao Yao^1,4,
Siqi Li^5,6,
Xin Tian^1,4,
Xintong Xia⁴,
Bowen Diao¹,
Lin Zhao¹,
Xinyuan Zhang⁷,
Danni Wei^5,6,
Long Wang²,
Mengle Shao ORCID: orcid.org/0000-0002-5488-9904^8,9,
Ying Wu ORCID: orcid.org/0009-0008-1808-3349^4,9,
Jinyu Huang² &
…
Bo Shan ORCID: orcid.org/0000-0002-1675-5809^1,10

Nature Communications volume 17, Article number: 3635 (2026) Cite this article

6420 Accesses
8 Altmetric
Metrics details

Subjects

Abstract

Tissue-resident mesenchymal stromal cells (MSC) instruct immune cell activation at injury sites, a key event in tissue repair. However, the full array of immunomodulatory mechanisms driving injury-responsive immune-stromal cell interactions is underexplored. Here, using an endovascular injury mouse model, we demonstrate that perivascular MSCs enhance arterial immune response by facilitating recruitment of ST2-expressing reparative macrophages in male mice. Time-resolved single-cell sequencing reveals an MSC-mediated, macrophage phenotypic switch, essential for vascular regeneration. Mechanistically, injury activates NFκB-dependent IL-33 production in MSCs, which acts as a paracrine signal that drives Osteopontin (OPN/SPP1) production in ST2+ macrophages and stimulates vascular smooth muscle cell (VSMC) proliferation and neointima formation. Local hydrogel-mediated delivery of siRNAs targeting Il33 or Spp1 effectively alleviates injury-induced neointimal hyperplasia, suggesting a potential therapeutic strategy to prevent restenosis and other vascular diseases. Our findings define an IL33-ST2-OPN axis that mediates functional crosstalk between perivascular MSCs and reparative macrophages, orchestrating immune-mediated reparative responses.

A discrete ‘early-responder’ stromal-cell subtype orchestrates immunocyte recruitment to injured tissue

Article 06 November 2023

HuMSC-EV induce monocyte/macrophage mobilization to orchestrate neovascularization in wound healing process following radiation injury

Article Open access 01 February 2023

Mesenchymal stem cells in treating human diseases: molecular mechanisms and clinical studies

Article Open access 22 August 2025

Introduction

Tissue repair can occur through tissue regeneration, which restores normal tissue structure, or through healing, which results in scar formation. A well-balanced tissue repair process is crucial for reinstalling tissue homeostasis and determines the pathogenesis of diseases^1,2,3,4. Immune cells play a critical role in the regulation of regenerative and inflammatory responses essential for tissue repair⁵. Following tissue injury, a large number of macrophages are mobilized and activated, contributing to tissue regeneration, inflammation and fibrosis^1,5,6. Notably, macrophages act in a sequential and coordinated cascade to remove pathogens or necrotic debris and resolve inflammation during tissue injury and repair. The prevailing paradigm suggests that macrophages initially recruited to the injured sites are pro-inflammatory, while those involved in the subsequent regeneration phase transition to an anti-inflammatory and reparative phenotype^2,5,7,8. This phenotypic transition ensures the precisely timed progression of tissue repair, preventing asynchrony, which could lead to impaired regeneration, increased fibrosis, and prolonged inflammation^1,2,5. Despite their critical functions, the precise identities and lineage origins of these reparative macrophages, as well as the mechanisms governing their acquisition of pro-repair phenotype, remain largely unclear.

Vascular repair is a tightly regulated process whose progression determines the delicate balance between proper regeneration of the endothelial monolayer and the undesired formation of vascular lesions following vessel damage. Upon endovascular injury, the vascular repair process includes endothelial cell (EC) dysfunction, vascular smooth muscle cell (VSMC) proliferation as well as active immune response within blood vessel walls^9,10,11. In response to endothelial injury, VSMCs are stimulated to proliferate and migrate from the tunica media into the tunica intima, driving a vascular remodeling process called neointima hyperplasia. This process leads to vascular wall thickening and lumen narrowing and is associated with a variety of cardiovascular diseases, such as atherosclerosis, restenosis, hypertension, and arterial aneurysm^9,10,12,13. However, the mechanisms underlying the injury-responsive immune response and their implications in VSMCs proliferation and neointima formation remain poorly understood.

A complex dialog between immune system and other microenvironmental components is crucial for tissue repair and regeneration^1,2,3. Injury triggers the recruitment, proliferation, and activation of various immune and non-immune cells, which together make up an intercellular communication network to orchestrate tissue repair^1,3,4. Recent efforts have focused on clarifying how reparative macrophages act in a paracrine and/or autocrine manner to contribute to the repair cell milieu via the production of bioactive factors, such as cytokines, growth factors and metabolites^{1,6,7,8,14,15,16}. A panel of cytokines (e.g., IL-4, IL-13, IL-10, TGFβ), growth factors (e.g., CSF-1, IGF-1) and signaling pathways (e.g., STAT6, NFκB, PPARγ, AMPK) have been reported to regulate macrophage phenotypic transitions^17,18. Notably, IL-33/ST2 signaling, through downstream effectors, such as AKT, ERK or GATA3, governs the phenotypes of macrophages and other immune cells and has been implicated in various contexts^19,20,21. However, its role in macrophage phenotypic switching during tissue injury and repair remains unclear.

Mesenchymal stromal cells (MSC), a major stromal fraction localized in perivascular regions, have recently been identified as a primary cellular source of IL-33 in peripheral tissue^{22,23,24,25,26}. MSCs are recognized for their fundamental role in providing structural support to tissues and organs^27,28. It has now become evident that they have additional functions in tissue repair, particularly by activating the immune system following injury^28,29. In the vasculature, adventitial MSCs residing in the tunica adventitia have been implicated in pathways in the contexts of vascular remodeling and disease development^27,30,31,32. Recent studies have revealed that MSCs could respond to injury-associated signals through reciprocal interactions with other immune cells to mobilize various microenvironmental components critical for tissue regeneration and wound healing^28,33,34,35. Furthermore, studies employing cell fate mapping and single-cell sequencing have underscored the tunica adventitia as a regulatory center for vascular remodeling, influencing the functions of medial and intimal cells^{31,32,36,37,38}. Notably, PDGFRβ + SCA1+ adventitial MSCs do not directly contribute to new VSMCs in the context of wire-induced artery injury³⁰. Instead, they are likely to function in a regulatory capacity, supporting the artery repair process rather than serving as a cellular source of new VSMCs³⁰.

In this study, we integrate time-resolved single-cell RNA sequencing and multi-color flow cytometry to comprehensively analyze the dynamic changes in cell-cell communication circuits within the artery wall and define how MSCs coordinate immune responses following endovascular injury. We find that mesenchymal IL-33 drives the activation of ST2+ macrophages, which promote smooth muscle cell proliferation and neointimal growth through Osteopontin (OPN/SPP1), and that disrupting this IL-33-ST2-OPN axis limits neointimal hyperplasia. These results establish a mechanistic framework for immune-stromal crosstalk in arterial repair and highlight a therapeutically targetable pathway for preventing restenosis and related vascular diseases.

Results

Arterial recruitment of CX3CR1+ macrophages is essential for endovascular injury-induced neointimal hyperplasia

Emerging evidence highlights the critical role of vascular immune cells in coordinating blood vessel repair¹¹. Using the endovascular injury mouse model where a wire is inserted into the murine femoral artery to mimic clinical coronary stent implantation³⁹, we monitored the temporal recruitment of CD45+ immune cells in the injured artery wall over a 4-week period of vascular injury and repair (Fig. 1a and Supplementary Fig. 1a). Quantification of CD45+ immune cells, together with functional activity score analysis of immune cell populations, revealed that endovascular injury induced a rapid immune response, with immune cell infiltration peaking around 2 days post-injury (Fig. 1b and Supplementary Fig. 1b–d).

**Fig. 1: CX3CR1+ arterial macrophage recruitment is essential for endovascular injury-induced neointimal hyperplasia.**

The abundance of immune cells in the artery began to decline around 4–7 days post-injury (Fig. 1b and Supplementary Fig. 1b, d), coinciding with the inflammatory-to-reparative phase transition in the repair process^40,41. To characterize the temporal changes in the immune landscape during artery injury and repair at single cell resolution, we collected femoral arteries at 0, 2, and 7 days post-injury for single-cell RNA sequencing (scRNA-seq) (Fig. 1a). A total of 27,515 high-quality cells were captured and sequenced across the 3 time points (0 days [D0 or Sham], 5960 cells; 2 days [D2], 11,016 cells; 7 days [D7], 10,539 cells). After combining the data from all 3 time points, clustering and visualization using Uniform Manifold Approximation and Projection (UMAP) revealed 4 major vascular cell populations, each expressing specific markers: VSMC (Myh11, Acta2, Tagln), MSC (Pdgfrb, Pdgfra, Dcn), EC (Pecam1, Cdh5), and Immune cell (Ptprc) (Supplementary Fig. 1e, f and Supplementary Data 1).

Endovascular injury induced a pronounced recruitment of various arterial immune cell types, including B cells, macrophages, mast cells, neutrophils, NK/T cells, dendritic cells (DC), monocytes, and group 2 innate lymphoid cells (ILC2) (Fig. 1c, Supplementary Fig. 1c, d and Supplementary Data 1). Consistent with histological observations, these immune cell populations exhibited distinct temporal dynamics throughout the vascular repair progress (Fig. 1d, e and Supplementary Fig. 1b–d, S1i–k). Notably, a substantial increase in macrophages residing within the arterial wall (i.e., arterial macrophages) was detected at 7 days and 14 days post femoral artery injury (FAI) (Fig. 1d, e and Supplementary Fig. 1c, d), coinciding with the period of active VSMC proliferation³⁸. Their abundance eventually declined to baseline levels as vascular regeneration progressed by 28 days post-injury (Supplementary Fig. 1b and S1i–k). In addition to the single-cell transcriptomics, we also employed multi-color flow cytometry to comprehensively analyze vascular immune cells at the time points of 0, 2, and 7 days post-injury (Fig. 1a). Consistent with the histological and scRNA-seq data, the abundance of CD45+ immune cells increased ~16-fold in the injured artery immediately post-injury and remained 8-fold higher at 7 days post-injury compared to uninjured controls (Fig. 1f). t-SNE analysis of the multi-color flow cytometry data identified 11 distinct immune cell populations, including macrophages, monocytes, neutrophils, ILC2s, mast cells, plasmacytoid DCs (pDCs), conventional DCs (cDCs), NK cells, B cells, CD8+ T cells, and CD4+ T cells (Fig. 1g; Supplementary Fig. 1l and Supplementary Data 1). The distinct immune landscapes at each time point were reflected by the altered proportions and abundance of these immune subsets (Fig. 1h–j).

As the most abundant immune cell population in the injured arteries, macrophages showed robust changes after injury, with ~60-fold increase at 2 days and over a 20-fold increase at 7 days post-FAI (Fig. 1h, j and Supplementary Fig. 2a). Both flow-cytometric analyses using Cx3cr1^GFP reporter mice and anti-CX3CR1 antibodies, as well as in combination with scRNA-seq analysis, confirmed that the increased macrophages in the injured arterial wall were predominantly CX3CR1+ cells (Fig. 1k, l and Supplementary Fig. 2b–f). To evaluate the role of macrophages in neointimal formation, we generated an inducible macrophage depletion mouse model (Cx3cr1^CreER, Rosa26-iDTR, denoted as Mac△ mice), in which tamoxifen administration sensitized CX3CR1+ cells to diphtheria toxin (DT) by activating diphtheria toxin receptor (DTR) expression (Fig. 1m). DT-induced depletion effectively eliminated arterial macrophages without affecting the abundance of other immune cell populations within the arterial wall (Supplementary Fig. 2g). Consequently, neointimal formation was markedly reduced in Mac△ mice compared with control animals at 4 weeks post-injury (Fig. 1n). Collectively, these results demonstrate that arterial macrophages are essential for the neointimal hyperplasia following endovascular injury.

Identification of a subset of reparative macrophages that facilitates VSMC proliferation upon endovascular injury

Following endovascular injury, artery repair begins with an initial inflammatory phase that predominates during the first 3 days post-injury, before transitioning into a reparative phase to complete the repair process^40,41 (Fig. 2a). Next, we examined whether arterial macrophages displayed distinct activation states at the 2 days (D2, inflammatory phase) and 7 days (D7, reparative phase) post-FAI. Compared to the baseline state (D0), genes upregulated in D2 macrophages were mainly associated with inflammation-related pathways, such as Innate immune response, Cell chemotaxis, and Phagocytosis (Fig. 2b, left). In contrast, D7 arterial macrophages showed elevated expression of genes associated with vascular regeneration, including those related to VSMC proliferation, Regeneration, and Collagen metabolism (Fig. 2b, right). These findings suggested distinct functional states of arterial macrophages across different phases of the repair process. Further unsupervised clustering identified 4 macrophage subsets (Cluster 1-4) within the arterial wall during vascular repair, each showing substantial temporal changes in abundance and transcriptional profiles (Fig. 2c, d, and Supplementary Fig. 3a). Clusters 1, 2, and 3 predominantly emerged in the arterial wall after injury, whereas Cluster 4 represented homeostatic resident macrophages as being observed only in arteries under normal condition (Fig. 2d). High Cx3cr1 expression in Clusters 1, 2, and 3, together with the absence of Adgre1 in Cluster 3 (Supplementary Fig. 3b), identified Cluster 3 as monocytes and suggested that Clusters 1 and 2, which were sequentially recruited during the reparative phase, originate from these infiltrated monocytes (Cluster 3). Pseudo-temporal ordering and differentiation analyses supported this, showing that the Clusters 1 and 2 derived from Cluster 3 monocytes (Fig. 2e, f, and Supplementary Fig. 3c–e). To validate their cellular origin, GFP-labeled monocytes were i.v. injected into wild-type mice at D4 (see Methods for experimental details). By D7, ~50% of arterial macrophages were GFP+ cells by flow cytometry analysis, indicating that newly infiltrated monocytes are the primary source of Cluster 2 macrophages during the reparative phase of vascular repair (Fig. 2g).

**Fig. 2: Characterization of arterial macrophage heterogeneity uncovers a subset of reparative macrophages that facilitate injury-induced VSMC proliferation.**

In addition to their distinct cellular origins, Cluster 1 macrophages expressed higher levels of inflammatory genes, while Cluster 2 macrophages showed relative upregulation of genes associated with vascular repair and regeneration (Supplementary Fig. 3f,g). To develop a flow cytometry protocol capable of distinguishing inflammatory from reparative macrophages, we examined top-ranked differentially expressed genes (DEGs) in each subcluster to identify suitable cell-surface markers. Ly6c2 (encoding LY6C protein) was dominantly expressed in Cluster 1, while Cd63 and Cd72 (encoding CD63 and CD72 proteins, respectively) were expressed relatively higher among surface markers in Cluster 2 (Supplementary Fig. 3h, and Supplementary Table 2). Notably, LY6C was more abundant in D2 macrophages, whereas CD63 and CD72 were predominant in D7 macrophages at protein levels (Supplementary Fig. 3i). In agreement, over 90% of D2 macrophages were LY6C positive, but this proportion decreased to ~35% at D7 (Fig. 2h, i). In contrast, a LY6C-CD63 + CD72+ subset of arterial macrophages, which was rarely detected at the onset of injury, was abundantly observed at D7 (Fig. 2h, j). This inverse occurrence of these 2 macrophage clusters prompted us to investigate whether LY6C+ macrophages and the LY6C-CD63 + CD72+ counterparts function differently during the artery repair process. First, we used the CSF1R inhibitor Ki20227 to temporarily deplete vascular monocytes/macrophages at different time points after injury⁴². Notably, Ki20227 treatment administered immediately after injury (D0-D3) resulted in markedly reduced immune cell infiltration, particularly macrophages and downregulated inflammation genes (e.g., Ccl2, Cxcl1, Cxcl10, Il6, Nos2, Il1b), without affecting cell proliferation-related genes (e.g., Ccna2, Ccnb1, Ccnd1, Ccne1, Mki67, Pcna) (Fig. 2k, l, and Supplementary Fig. 4a, b). In contrast, macrophage depletion between D4-D7 resulted in decreased expression of genes related to cell proliferation, but not those associated with inflammation (Fig. 2m, n, and Supplementary Fig. 4c). In addition, 5-ethynyl-2’-deoxyuridine (EdU) staining showed there were approximately 50% fewer proliferating VSMCs (Fig. 2o) as well as nearly 80% reduction in neointimal formation (Fig. 2p) in animals treated with Ki20227 between D4-D7. These results collectively demonstrate that LY6C+ macrophages represent inflammatory macrophages (herein referred to as InfMAC) that mediate vascular inflammatory response following injury, while the LY6C-CD63 + CD72+ subset functions as reparative macrophages (referred to as RepMAC) that promote VSMC proliferation and neointimal hyperplasia.

Perivascular PDPN^hi MSCs exhibit heightened immunoregulatory activity in artery wall upon endovascular injury

In addition to the remarkable changes in immune cell abundance and frequency, our scRNA-seq dataset uncovered extensive interactions among MSCs, VSMCs, ECs, and immune cells, which were significantly intensified following endovascular injury (Fig. 3a, b, and Supplementary Fig. 5a). Notably, the intensity of these cell-cell interactions was found to be both cell type-dependent and highly dynamic throughout the repair process (Fig. 3c, Supplementary Fig. S5a,b). Among all detected cell types, MSCs exhibited the most pronounced interactions with immune cell populations, particularly macrophages. These interactions became progressively strengthened during tissue repair (Fig. 3c, and Supplementary Fig. 5b). Further analysis identified 2 distinct subpopulations of MSCs localized in arteries, designated as MSC1 and MSC2 (Fig. 3d). In contrast to MSC1 which represented the major MSC population in the normal state, MSC2 was predominantly detected at 7 days post-injury and exhibited elevated expression of genes involved in extracellular matrix deposition and immune cell regulation (Fig. 3f, g), indicating their specific recruitment associated with endovascular injury. As a recognized marker of MSCs, Podoplanin (PDPN) expression is upregulated in response to various contexts, including infection, tissue injury, chronic inflammation and cancer, and is implicated in the migration and activation of MSCs^24,43. Notably, the injury-associated MSC2 cells expressed higher levels of PDPN (PDPN^hi) compared to the MSC1 counterparts (Fig. 3h–k), which was consistent with the increased presence of PDPN+ MSCs in injured arteries, as quantified by flow cytometry analysis (Fig. 3l). Bulk RNA sequencing experiments using isolated PDGFRβ + PDPN + MSCs from femoral arteries after wire injury or sham operations (Fig. 3m) revealed that endovascular injury caused substantial alterations in the transcriptomes of arterial MSCs (Fig. 3n, and Supplementary Fig. 5c). Notably, the 1796 upregulated genes in MSCs from the injury group were predominantly enriched in immune response-related pathways (Fig. 3o), confirming the heightened injury-induced immunoactivation of these MSCs. These new findings suggest that arterial PDPN^hi MSCs with immunomodulatory features are involved in the immune remodeling and tissue repair processes following endovascular injury.

**Fig. 3: Endovascular injury-activated MSCs exhibit heightened immunoregulatory features.**

IL-33 expression in artery adventitial MSCs is activated by NFκB signaling pathway upon endovascular injury

Growing evidence highlights the critical role of MSCs in modulating immune cell functions and remodeling microenvironment via the release of secreted factors^43,44. To identify MSC-derived factors regulating artery repair, we analyzed genes that encoded secreted proteins whose expression was altered by FAI in bulk RNA-seq data (MSC_FAI vs. Sham). Among the 50 DEGs, Il33 is the top1 upregulated gene and predominantly expressed in MSCs (Supplementary Fig. 5d, e). Notably, Il33, encoding the alarmin cytokine IL-33 that orchestrates reparative type 2 immune responses in multiple settings^19,45,46, was among the most strongly upregulated genes in the FAI group compared to Sham counterparts (Fig. 4a). Akin to findings in white adipose tissue, pancreas and lung^{22,23,24,25,26}, murine artery IL-33 expression was predominantly detected in PDGFRβ + or PDGFRα + MSC subsets located in the adventitial layer, as evidenced by scRNA-seq analysis of FAI model shown in Fig. 1, immunofluorescence staining of injured artery sections with IL-33 antibodies, and flow cytometry using the Il33^GFP reporter mouse model (Fig. 4b, c, and Supplementary Fig. 5e, S6a–c). Moreover, the upregulation of Il33 gene expression, rather than the changes in the abundance of Il33-expressing cells, appeared to contribute to the higher IL-33 levels observed in injured arteries (Fig. 4d–f, and Supplementary Fig. 6a–c). Notably, plasma IL-33 levels were significantly higher in patients with in-stent restenosis (ISR) compared with those without ISR (Fig. 4g). EC denudation is a key initiating event in multiple vascular pathologies, including stent-induced artery injury^9,39. To model this process in vitro, we co-cultured Il33-expressing PDGFRβ + MSCs isolated from inguinal adipose tissue²² with conditioned media (CM) derived from mechanically injured cultured human and mouse EC lines, respectively (Fig. 4h, and Supplementary Fig. 6d). As expected, exposure to the CM from both injured HUVECs and bEND.3 cells (CM_Inj) induced a 2-3-fold increase in Il33 expression compared to CM from normal ECs (CM_Nor) (Fig. 4i,j, and Supplementary Fig. 6e). To identify the secreted factor(s) mediating IL-33 expression, primary MSCs were treated with the CM which was subjected to 95 °C boiling or filtration through a 10-kDa-cutoff filter (Frac^<10K). The results pointed to the presence of heat-sensitive macromolecules derived from injured ECs contributed to the regulation of IL-33, as the effect of CM_Inj on Il33 induction was abolished only when boiled CM was applied, but not with the filtered CM (Fig. 4k, l). Hence, we further analyzed the bulk RNA-seq dataset of MSCs shown in Fig. 3 using Gene Set Enrichment Analysis (GSEA) and found that the NFκB signaling pathway, along with related pathways, such as the TNF signaling pathway and IL-17 signaling pathway, were highly enriched in the FAI group (Fig. 4m). Indeed, exposure to CM_Inj or the fractions containing materials with molecular weight of 10 kDa or above (Frac^>10K) resulted in robust activation of the IKK-NFκB signaling pathway in MSCs, as evidenced by the elevated levels of phosphorylated IKKα/β and p65 (Fig. 4n). In agreement, the effects of CM_Inj and Frac^>10K on Il33 expression were largely blocked by inhibition of the IKK-NFκB signaling pathway using p65 inhibitor JSH23 or through Rela knockdown via siRNA (Fig. 4o–q). Moreover, several secreted proteins were identified to be upregulated in the conditioned media derived from injured ECs, including B2M, MIF, CXCL5 and ITGB1, which are known to activate NFκB signaling pathway (Supplementary Fig. 6d, f, and Supplementary Data 2). In addition, sequence analysis of the promoter region of murine Il33 gene identified 3 putative p65 binding sites (B.S.-1 ~ 3) (Fig. 4r). Chromatin Immunoprecipitation (ChIP) assays demonstrated that the addition of CM_Inj significantly enhanced the direct binding of p65 proteins to chromatin regions of B.S.-2 and -3 (Fig. 4r), supporting the direct regulation of NFκB signaling pathway on Il33 transcription in MSCs.

**Fig. 4: IL-33 expression is enhanced by endovascular injury-activated NFκB signaling pathway in perivascular MSCs.**

IL-33 ablation in MSCs impedes the recruitment of reparative macrophages and limits neointimal hyperplasia following endovascular injury

To determine the role of IL-33 in MSC-mediated immune remodeling in the injured arterial wall, we developed a mouse model in which Il33 was selectively inactivated in PDGFRβ + cells using a tamoxifen-inducible Cre system (Pdgfrb^CreER, Il33^loxp/loxp; denoted as Pb-Il33^KO mice) (Fig. 5a). In this model, tamoxifen administration resulted in a > 80% reduction in Il33 expression within the arterial wall of Pb-Il33^KO mice (Fig. 5b), confirming that PDGFRβ + cells serve as the primary source of IL-33 in the artery. While there were no significant effects on the total number of CD45+ immune cells (Fig. 5c), the absence of IL-33 in MSCs led to notable changes in the composition of vascular immune cells, as revealed by immune cell profiling using multi-color flow cytometry (Fig. 5d, e). Interestingly, similar frequencies of type 2 immune cells, including ILC2s, mast cells, and CD4 + T cells, along with comparable mRNA levels of type 2 cytokines (Il4, Il5 and Il13) and inflammatory genes (Ccl2, Cxcl1, Cxcl2, Il1b, Tnf) were detected in the arteries of Pb-Il33^KO mice and their control counterparts (Fig. 5d–f, and Supplementary Fig. 7a, b). These findings suggested that elevated IL-33 did not activate conventional type 2 immune responses in the injured arteries. Nevertheless, a ~ 36% reduction of total macrophages was observed in the arteries of Pb-Il33^KO mice (Fig. 5f), accompanied by a trend towards higher neutrophil accumulation. Neutrophils primarily act during the acute phase of vascular injury, when the immune response peaks⁴⁷. Their contribution to late-stage neointimal hyperplasia is likely less significant compared with macrophages, which at this stage have largely acquired a reparative phenotype. Moreover, immunofluorescence staining of artery sections showed that IL33-expressing MSCs were localized in close proximity to macrophages in the adventitial layer of the injured artery (Fig. 5g). Meanwhile, the number of LY6C-CD63 + CD72+ macrophages was significantly decreased in the injured arteries of Pb-Il33^KO mice compared to their control counterparts (Fig. 5h, i), indicating that mesenchymal IL-33 was driving the recruitment of reparative macrophages.

**Fig. 5: MSC IL-33 inactivation affects endovascular injury-induced arterial immune remodeling and dampens neointimal hyperplasia.**

Next, we asked whether inactivation of IL-33 in MSCs affected injury-induced neointimal formation. In Pb-Il33^KO mice subjected to endovascular injury, the pathological thickening of the arterial SMC layer was reduced by ~70% (Fig. 5j). Compared to the control group, there was a remarkable decrease in VSMC proliferation in injured arteries with Il33 deletion, as evidenced by an ~51% reduction of Ki67+ VSMCs (Fig. 5k, l) and an ~72% decrease in EdU+ VSMCs (Fig. 5m). To confirm that these effects were due to the loss of local IL-33 signal, we delivered a siIl33-containing hydrogel to selectively silence Il33 expression at the femoral artery injury sites (Fig. 5n). As a result, the locally applied siIl33-containing hydrogel effectively suppressed both IL-33 mRNA and protein levels in injured arteries (Fig. 5o, p), which was associated with the reduced endovascular injury-induced thickening of SMC layers (Fig. 5q). These results confirm the local effect of IL-33 in promoting neointimal hyperplasia.

IL-33-ST2 signaling-activated reparative macrophages drive the neointimal hyperplasia in response to endovascular injury

To determine whether IL-33 directly influences VSMC proliferation, we supplemented VSMC cell line MOVAS cultures with recombinant IL-33 proteins and assessed the impact on cell proliferation. VSMCs treated with IL-33 showed comparable growth rates and proliferative capacities to those treated with vehicle (Supplementary Fig. 7c–e), suggesting the involvement of indirect IL-33-dependent mechanism(s) by which IL-33 stimulates VSMC proliferation. Therefore, we hypothesized that IL-33’s effect on facilitating VSMC proliferation requires the presence of other cell types that express ST2, the obligate IL-33 receptor¹⁹. Indeed, the expression levels of Il1rl1 gene (encoding ST2 protein) in the artery were remarkably increased immediately following endovascular injury, peaking at 1-2 weeks post-injury when extensive VSMC proliferation occurred (Fig. 6a). Importantly, these observations were in line with the results showing that IL1RL1 gene expression was also notably upregulated in human arteries following stent surgery (Fig. 6b). Moreover, the endovascular injury-induced ST2 expression coincided with the recruitment of ST2-expressing immune cells (Fig. 6c–e). Supporting our hypothesis that MSC-derived IL-33 was required for the recruitment of ST2+ immune cells to the injured sites, the number of ST2+ immune cells was decreased by ~45% in the injury arteries of Pb-Il33^KO mice (Fig. 6f–h). To further investigate the immune cells regulated by local IL-33-ST2 signaling within injured arteries, multi-color flow cytometry analysis identified 7 distinct types of ST2+ immune cells, including macrophages, monocytes, neutrophils, ILC2s, mast cells, B and T cells (Fig. 6i). Notably, ~58% of ST2+ immune cells in the injured arteries were macrophages, and the abundance of these ST2+ macrophages was reduced by ~55% in injured arteries when IL-33 was inactivated in MSCs (Fig. 6j).

**Fig. 6: IL-33-activated arterial macrophages promote endovascular injury-induced VSMC proliferation and neointima formation.**

Transcriptomic datasets of IL33-treated BMDM (IL-33^UP) and arterial macrophages after injury (D7_Mac^UP) revealed that 451 common genes were upregulated (Fig. 6k). These upregulated genes were significantly enriched in pathways related to wound healing and VSMC proliferation (Fig. 6k), which implied that IL-33-activated ST2+ macrophages acquired the reparative phenotype. In agreement, ST2+ macrophages were predominantly found within the RepMAC population in the arteries at D7 (Fig. 6l), with nearly 60% of RepMACs identified as ST2+ cells (Fig. 6m). Thus, we generated a mouse model in which Il1rl1 was specifically knocked out in myeloid cells (Lyz2 ^Cre, Il1rl1^loxP/loxP; denoted as Mac-Il1rl1^KO mice) to determine the role of macrophage ST2 in artery injury and repair. Macrophage ST2 deficiency substantially restrained the neointima formation in response to endovascular injury (Fig. 6n–p). Mac-Il1rl1^KO mice exhibited significantly reduced VSMC proliferation, evidenced by ~60% reduction in Ki67+ VSMCs and ~78% decrease in the number of EdU-incorporated VSMCs in injured arteries (Fig. 6q–s). Taken together, these new results demonstrate that IL-33-activated ST2+ reparative macrophages are crucial for facilitating VSMC proliferation and promoting neointimal hyperplasia.

ST2+ macrophage-derived Osteopontin stimulates VSMC proliferation upon artery injury

To elucidate how ST2+ macrophages promote VSMC proliferation, cultured VSMCs were subjected to the CM derived from IL33-treated macrophages (Fig. 7a). The CM exerted enhanced capacity in promoting VSMC growth, an effect which was blocked in the absence of ST2 (Fig. 7b, and Supplementary Fig. 8a, b). These observations led to the hypothesis that IL-33-induced secreted factor(s) from macrophages contribute to the stimulation of VSMC proliferation. Therefore, we profiled the putative ligand-receptor interactions between macrophages and VSMCs, and identified several macrophage-derived ligands—Spp1, Tgfb1, Osm, Tnf, Lgals9, and Pdgfb—whose interactions with their receptors on VSMCs were enhanced following endovascular injury (Fig. 7c). Among these ligands, Spp1 (which encodes Osteopontin, OPN) appeared to be a promising candidate for the following reasons. First, Spp1 gene expression and OPN proteins were predominantly detected in macrophages in the context of artery injury (Fig. 7e–g). Secondly, Spp1 expression was markedly upregulated in macrophages within injured arteries (Fig. 7d–g). In addition, the enhanced macrophage-VSMC interactions mediated by Spp1 were evident at D14 (Supplementary Fig. 8c, d). Consistently, arterial Spp1 mRNA levels were markedly elevated after endovascular injury (Fig. 7h), following a pattern similar to Il33 (Fig. 4f). Moreover, plasma samples from IRS patients also showed increased OPN protein levels (Fig.7i).

**Fig. 7: ST2+ macrophages-derived osteopontin drives neointimal hyperplasia upon endovascular injury.**

Vascular Spp1 expression levels in injured arteries were reduced by ~50% in Pb-Il33^KO mice in comparison with control littermates (Fig. 7j). In addition, the absence of ST2 also resulted in nearly 50% reduction of Spp1 expression in the injured arteries (Fig. 7k). Therefore, we set to determine whether the IL-33 receptor signaling pathway promotes OPN production in macrophages. As expected, IL-33 treatment led to over 4-fold upregulation of Spp1 gene expression in the macrophages, while this induction was abolished in ST2-deficient cells (Fig. 7l). In agreement, siRNA-mediated silencing Spp1 in macrophages using siRNA significantly reduced the expression levels of proliferation-related genes in VSMCs exposed to the aforementioned IL-33-treated macrophage-derived CM (Fig. 7m, n, and Supplementary Fig. 8e). In vivo, hydrogel-embedded siSpp1 effectively inhibited the progression of injury-induced neointima formation (Fig. 7o). Together, these results demonstrate that macrophage-derived OPN, driven by paracrine IL-33 signaling, promotes VSMC proliferation and neointimal hyperplasia following endovascular injury.

Discussion

Coronary artery disease (CAD), the leading cause of death worldwide, arises from progressing atherosclerosis, a multiplex disease characterized by chronic vascular wall inflammation, progressive narrowing of the vessel lumen and eventual plaque formation^12,48,49. Percutaneous angioplasty is a commonly used clinical treatment for CADs, but is ineffective in about 16–20% of all patients as the surgery-caused endovascular injury frequently leads to restenosis, which is a re-narrowing of the vessel caused primarily by excessive VSMC proliferation and extracellular matrix deposition^10,48,50. Although modulated VSMCs are the primary cellular component of the neointimal lesions, the roles of various adventitial cells, including MSCs and macrophages, as well as their interactions are essential for maintaining vascular homeostasis and facilitating remodeling in response to injury^42,51,52,53. In this study, we identified a subset of ST2-expressing arterial macrophages that respond to paracrine signals derived from adventitial MSCs and adopt a pro-repair phenotype. Among a panel of potential secreted factors mediating MSC-macrophage functional crosstalks (Figs. 3 and 4), IL-33 was found to stimulate OPN production in ST2-expressing macrophages, thereby promoting VSMC proliferation. Locally applied hydrogel-mediated delivery of siRNAs targeting Il33 or Spp1 at injury sites effectively restrains VSMC proliferation and prevents blood vessel narrowing in an endovascular injury model, highlighting the role of IL33-ST2-OPN regulatory axis in the context of injury-induced neointimal hyperplasia (Fig. 8). These findings reveal an intercellular regulatory network linking injury-responsive adventitial resident cells and reparative immune cells that drive neointima formation, representing a promising therapeutic target for preventing restenosis and other vascular diseases.

**Fig. 8: Proposed model: PDGFRβ+ perivascular mesenchymal cells direct ST2+ macrophage response during endovascular injury.**

Macrophages, the most abundant immune cells within large vessel walls, originate from both tissue-resident sources and circulating monocytes^42,54, and maintain arterial homeostasis through regulation of VSMC activity and collagen production^42,54,55. We and other groups have revealed that the monocyte-macrophage compartment represents the largest and most significantly altered population among arterial immune cells after arterial injury (Fig. 1)^51,53. Damage-associated molecular patterns (DAMPs) activate signaling pathways that recruit monocytes, which differentiate into macrophages at injury sites^1,6. Previous studies have revealed that the injury-induced macrophage accumulation within adventitial layer includes both tissue-resident macrophages and monocyte-derived macrophages (MDMs), with both types belonging to the CX3CR1+ macrophage population⁵⁴. Consistently, our study confirmed that the majority of macrophages in injured arteries are derived from newly-recruited circulating monocytes, primarily of the CX3CR1 lineage (Fig. 1). The genetic ablation of CX3CR1+ macrophages effectively reduced injury-induced restenosis, in agreement with prior findings that global Cx3cr1 deficiency protects against neointimal hyperplasia⁵⁶. It is well documented that macrophages undergo phenotypic transitions from an inflammatory to a reparative state during tissue regeneration and wound healing^5,57. However, the precise cellular origins of inflammatory versus reparative macrophages remain poorly understood. Recent studies in skin wound healing indicate that CX3CR1^hi reparative macrophages are long-term tissue-resident, while inflammatory CX3CR1^low/int macrophages are derived from circulation⁵⁸. Our data indicated that, in the injured artery, both inflammatory and reparative macrophages are CX3CR1+ cells (Fig. 1). Specifically, lineage tracing experiments revealed that the reparative macrophages recruited to the sites of vascular injury are derived from circulating monocytes (Fig. 2g), which supports a hypothetical model suggesting that newly recruited monocytes give rise to reparative macrophages following injury⁵.

Regarding cellular identity, single-cell resolution characterization of arterial macrophage heterogeneity using scRNA-seq and multi-color flow cytometry revealed a LY6C+ macrophage subset enriched with inflammatory markers (denoted as InfMAC), while the another LY6C-CD63 + CD72+ subset exhibits reparative features (denoted as RepMAC). Pseudotime trajectory analysis inferred their common monocytic origin and implied possible transitions between these two subpopulations, warranting future investigation using bona fide lineage-tracing experiments (Fig. 2e). The regulation of the inflammatory-to-reparative transition of macrophages remains an active area of investigation, particularly in the contexts of atherosclerosis and restenosis. Notably, pseudotime trajectory analysis revealed the coexistence of inflammatory and reparative macrophage populations in injured arteries suggests that macrophage phenotypic transitions during tissue repair do not occur as a simple binary switch but instead reflect a dynamic equilibrium in which both macrophage subsets coexist and remain functionally active (Fig. 2e and Supplementary Fig. 3). The tissue microenvironment, especially cell-cell interactions mediated by local signaling, is also pivotal in driving macrophage phenotype transitions during tissue repair^1,6. MSCs, residing in the adventitial layer of large blood vessels, are capable of sensing and responding to a wide range of stimuli through reciprocal interactions with immune cells^33,34. Previous studies have shown that PDGFRβ + SCA1+ MSCs do not contribute to de novo VSMCs in the model of wire-induced artery injury, as shown through genetic lineage tracing experiments. Instead, the pre-existing VSMCs contribute significantly to the formation of new VSMCs during injury-induced vascular regeneration³⁰. The cross-talk between MSCs and immune cells, like macrophages, is considered to be critical for restoring tissue homeostasis after injury²⁸. Our study found that in response to endovascular injury, a PDPN^hi subset of adventitial MSCs emerges with an enhanced immune response, characterized by heightened IL-33 production. Thereby, these PDPN^hi MSCs activate ST2-expressing arterial macrophages to acquire pro-repair phenotype and subsequently facilitate VSMC proliferation. Hence, we identified the IL-33 obligate receptor ST2 as a functional marker for reparative macrophages based on 2 key findings: 1) its expression pattern strongly overlaps with CD63 and CD72 in arterial macrophages at 7 days following arterial injury, and 2) macrophage-specific deficiency of ST2 effectively disrupted the reparative features of arterial macrophages. Further studies are required to elucidate whether the pro-repair activity of IL-33-activated ST2+ macrophages is specific to injured arteries or is also applicable to other contexts of tissue injury and repair.

Considerable efforts have been dedicated to uncovering how the injury-recruited macrophages influence vascular diseases⁵¹. Reparative macrophages facilitate tissue repair through the production of various effectors, including growth factors (e.g., VEGF, PDGFA), matrix metalloproteinases (e.g., MMP-8/10/28), and bioactive metabolites^6,7,8,14,15. Our results highlight that Spp1, a selective marker enriched in TREM2+ macrophages^59,60, plays a functional role in MSC-dependent immunoregulation crucial for vascular repair. Notably, both TREM2+ and TREM1+ macrophages⁶¹ exhibited temporal increases during vascular injury and repair (Supplementary Fig. 3j). In the injured artery, OPN is predominantly produced by ST2+ reparative macrophages activated by paracrine IL-33 signaling. Moreover, the hydrogel-based siRNA delivery approach targeting local Il33 and Spp1 gene expression has proven effective in blocking the progression of neointimal hyperplasia (Figs. 5 and 7). Given the thrombogenic risk associated with coated stents, bare stents combined with hydrogel therapy could offer a safer alternative by promoting rapid endothelialization while being less prone to restenosis^10,48,50. In line with the clinical application of hydrogel in treating cardiac and artery disorders^62,63, our studies suggest that hydrogel-based delivery system could serve as a promising strategy for treating cardiovascular diseases, such as restenosis following the surgery of percutaneous angioplasty. A key limitation of our current study is lack of in vivo validation in humans to confirm the translational relevance of our findings from mice. Future investigations involving human subjects will be essential to elucidate the role of MSC-instructed reparative macrophages in vascular repair after injury and to evaluate their potential as a therapeutic target for preventing neointimal hyperplasia. In addition, the EC lines used in our in vitro experiments may not fully recapitulate the phenotypic features of arterial EC. Finally, the mechanisms governing macrophage phenotypic transition in this context remain incompletely defined and warrant future investigation.

In conclusion, we reveal a previously unrecognized MSC–Macrophage signaling axis that shapes vascular repair after injury. Injury-activated adventitial MSCs release IL-33 to recruit and activate reparative ST2+ macrophages, which drive VSMC proliferation via OPN secretion. Targeting this IL-33–ST2–OPN pathway offers a promising strategy to limit neointimal hyperplasia and prevent restenosis. These findings significantly advance our understanding of how immune–stromal interactions regulate tissue repair, particularly in the vascular context, and uncover MSC–Macrophage crosstalk as a potential therapeutic target for preventing restenosis and related vascular diseases.

Methods

Animals

All animal experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee at Zhejiang University. All mice were from C57BL/6 background in this study. Male C57BL/6 mice (7-8 weeks old) were obtained from Hangzhou Ziyuan Laboratory Animal Technology Co., Ltd. Cx3cr1^GFP mice (Strain#: 005582), Cx3cr1^CreER mice (Strain#: 020940), Rosa26-iDTR mice (Strain#: 007900) were obtained from The Jackson Laboratory and kindly provided by Dr. Peng Shi (Zhejiang University). Pdgfrb^CreER mice (Strain#: 029684), Il33^{flox/flox-eGFP} mice (Strain#: 030619) were obtained from The Jackson Laboratory. Lyz2^Cre mice (Strain#: T003822), Il1rl1^flox/flox mice (Strain#: T005741) were generated by GemPharmatech. Macrophage specific Il1rl1 gene knockout mice (Mac-Il1rl1^KO) were generated by intercrossing Il1rl1^flox/flox mice with Lyz2^Cre mice. Inducible macrophage ablation mice (Mac^Δ) were generated by crossing Cx3cr1^CreER mice with Rosa26-iDTR mice. The mice were co-housed, breeding and maintained at 23 ± 3 °C with a humidity of 35 ± 5% under a 12-hour dark-light cycle, with free access to water and food in the SPF facilities of the Animal Facility of Zhejiang University. For the animal study, the number of animals in each experiment has been indicated in the Figure legends or Supplementary Data 3. In this study, male mice were used to avoid estrogen-related variability that could confound the mechanistic interpretation of vascular injury responses.

Mouse studies

For the establishment of the wire-induced femoral artery injury (FAI) model³⁸, 8-week-old male mice were anesthetized via intraperitoneal injection of avertin (#75-80-9, Sigma) and the femoral artery was isolated by carefully removing the surrounding vein and connective tissues using microsurgical forceps. Then, a 0.25 mm flexible 21-gauge wire (tips of a CROSS-IT 200XT guide wire, #1003313H, Abbott Laboratories) was inserted into the femoral artery and advanced 5–10 mm towards the iliac artery, where it was left in place for 3 minutes. Blood flow was restored following ligation of the profunda femoris branch. Mice were placed on a warming pad until fully awake post-surgery and were euthanized via intraperitoneal injection of avertin (#75-80-9, Sigma) at the indicated time points for further analysis.

For pharmacological depletion of macrophages, 8-week-old male mice were administered Ki20227 (#7688, Selleck) orally at a dosage of 40 mg/kg body weight, or vehicle (0.5% methylcellulose) at the indicated time points post-FAI. To inhibit inflammatory macrophages, administration was conducted once a day for 2 consecutive days right after surgery. To inhibit reparative macrophages, daily administration commenced at 4 days post-FAI and continued for 3 consecutive days. Mice were euthanized and then sacrificed at 4 days, 7 days, or 28 days post-FAI, and femoral artery samples were collected for further analysis.

For tamoxifen-induced gene knockout in mice, tamoxifen (#10540-29-1, Sigma) was dissolved in corn oil (#8001-30-7, Aladdin) at a concentration of 20 mg/mL. Mice carrying Pdgfrb^CreER allele or control animals, received intraperitoneal (i.p.) injections of tamoxifen at a dosage of 50 mg/kg body weight for 5 consecutive days to induce Cre-mediated recombination. For the mice carrying Cx3cr1^CreER allele or control animals, tamoxifen injection at a dosage of 50 mg/kg body weight were performed for 2 consecutive days prior to FAI surgery and at 1, 7 and 14 days post-FAI.

To selectively deplete monocyte-derived macrophages by diphtherial toxin injection, 8-week-old male MacΔ mice (Cx3cr1^CreER; Rosa26-iDTR) and control animals were treated with tamoxifen as described above, followed by intraperitoneal (i.p.) injections of diphtheria toxin (DT, #D381867, Aladdin) at a dosage of 8 μg/kg body weight at 1, 7, and 14 days post-FAI. Mice were euthanized and then sacrificed at indicated time points post-FAI.

For in vivo siRNA delivery via hydrogel, a 20% (w/v) Pluronic F127 (#P2443, Sigma) solution was prepared by dissolving the polymer in sterile PBS and stirring overnight at 4 °C to ensure complete dissolution. siRNA (siIl33, siSpp1, or scramble control siRNA) was incorporated at a final concentration of 30 µg per 50 µL of gel, and all procedures were performed on ice to maintain the liquid state prior to application. 50 μL of 20% Pluronic gel F127 containing 30 μg siIl33, siSpp1, or scramble siRNA (siNC) was applied to the adventitial surfaces of injured arteries right after surgery, respectively. Femoral artery samples were harvested at indicated time for further analysis. Small interfering RNAs were purchased from GenePharma: siNC (sense 5’-UUCUCCGAACGUGUCACGUTT-3’ and antisense 5’-ACGUGACACGUUCGGAGAATT-3’), siIl33 (sense 5’-AGACUCCGUUCUGGCCUCACCAUAA-3’ and antisense 5’-UUAUGGUGAGGCCAGAACGGAGUCU-3’), siSpp1 (sense 5’- GCCAUGACCACAUGGACGATT-3’ and antisense 5’- UCGUCCAUGUGGUCAUGGCTT-3’).

For in vivo evaluation of VSMC proliferation using EdU incorporation, mice were i.p. administered with EdU (#ST067, Beyotime) at a dosage of 50 mg/kg body weight at indicated time points. EdU incorporation evaluation of the cells in the femoral artery samples collected at the indicated time was performed with immunofluorescence staining using the BeyoClick™ EdU Cell Proliferation Kit (#C0075, Beyotime). For in vitro proliferative cell detection, VSMC cell line MOVAS were cultured and incubated with 10 μM EdU for 2 hours. Cells were then washed with PBS for three times and fixed with 4% PFA. Incorporated EdU was detected and evaluated using the BeyoClick™ EdU Cell Proliferation Kit (#C0075, Beyotime). MOVAS cell line was obtained from ATCC and cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin.

For in vivo monocyte infiltration assay⁴⁴, leukocyte pools from the blood of 8-week-old Cx3cr1^GFP mice were subjected to erythrocyte lysis prior to the incubation of indicated flow antibodies. Then, CD45 + CD11b + CD115+ circulating monocytes were isolated from the leukocytes by FACS. For each mouse, 1×10⁶ viable monocytes were resuspended in 100 μL PBS and injected into the mouse via tail vein at 4 days post-FAI. Three days post injection, GFP-expressing cells in the femoral artery were evaluated using flow cytometry.

Arterial cells preparation from murine femoral artery and flow-cytometric assays

For single arterial cells preparation, fresh femoral arteries were harvested and placed in a petri dish containing pre-cooled PBS. Perivascular adipose tissues and connective tissues were carefully removed, and the arteries were cut into fine pieces before incubation in a digestion solution containing 1.5 mg/mL Collagenase I (#17018029, Gibco) and 0.25 mg/mL bovine serum albumin (BSA, #A8020, Solarbio) in Hank’s balanced salt solution (#BL559A, Biosharp). The digestion was performed at 37 °C with shaking at 120 rpm. Detached cells in the supernatant were collected every 10 min and suspended in pre-cooled FACS buffer (Ca²⁺/Mg²⁺ free PBS supplemented with 2% FBS). Then, fresh digestion solution was added into the remaining undigested tissue for 3–4 rounds more digestion processes. All collected cells from each artery sample were pooled, filtered through a 70 μm cell strainer, and centrifuged at 600 g for 5 min. The cell pellets were resuspended in 100 μL RBC Lysis Buffer (#420301, Biolegend) and incubated on ice for 1 min for erythrocyte lysis. Then, the cells were harvested and resuspended in 1 mL pre-cooled FACS buffer for further analysis.

For flow-cytometric analysis, isolated cells were incubated with fluorescence-conjugated flow antibodies at 4 °C for 30 min. The following antibodies and working concentrations were used: CD45-BV510 (#103138, Biolegend, 1:400), CD3-AF488 (#100210, Biolegend, 1:200), CD4-eF450 (#48-0041-82, Invitrogen, 1:400), CD8a-AF700 (#557959, BD Biosciences, 1:100), B220-BV605 (#103244, Biolegend, 1:400), NK1.1-APC (#108710, Biolegend, 1:400), CD11b-PerCP/Cyanine5.5 (#101228, Biolegend, 1:400), F4/80-APC/Cy7 (#123118, Biolegend, 1:400), CD11c-Dazzle594 (#117348, Biolegend, 1:800), I-A/I-E-BV711 (#107643, Biolegend, 1:800), LY6C-PE (#128008, Biolegend, 1:2000), LY6G-PerCP (#127653, Biolegend, 1:200), ST2-PE-Cyanine7 (#25-9335-80, Invitrogen, 1:400), CD117-BV785 (#105841, Biolegend, 1:400), CD127-BV650 (#135043, Biolegend, 1:400), CD206-BV421 (#141717, Biolegend, 1:400), CD11b-BV421 (#101236, Biolegend, 1:400), F4/80-APC (#123116, Biolegend, 1:400), CX3CR1-PerCP/Cyanine5.5 (#19010, Biolegend, 1:400), CD45-PerCP/Cyanine5.5 (#103132, Biolegend, 1:400), CD31-PerCP/ Cyanine5.5 (#102420, Biolegend, 1:400), PDGFRβ-PE (#13600, Biolegend, 1:100), PDGFRβ-PE-Cyanine7 (#25-1402-8, Invitrogen, 1:100), SCA1-PE-Cyanine7 (#108113, Biolegend, 1:2000), OPN-PE (#IC808P, R&D, 1:100), and Ki67-FITC (#652410, Biolegend, 1:400) (Supplementary table 5). For intracellular Ki67 staining, cells post the staining of cell membrane markers were further fixed, permeabilized and stained with Ki67-FITC antibodies using the Foxp3 Staining Set (#00-5523-00, eBioscience) according to the manufacturer’s instructions. For the evaluation of intracellular lipid contents, the cells were further incubated with 2 μM BODIPY™ 493/503 (#D3922, Thermo Fisher) for an additional 20 min following cell membrane protein staining. For intracellular OPN staining, cell suspensions were incubated with Brefeldin A (3 ug/mL; #HY-16592, MCE) for 6 h prior to being fixed, permeabilized and stained with OPN-PE antibodies using the Foxp3 Staining Set (#00-5523-00, eBioscience).

Immune cell populations were identified by flow cytometry based on the following surface marker expression profiles: Macrophage (Mac) as CD45⁺CD11b⁺CD3^-B220^-NK1.1^-LY6G^-F4/80⁺; Neutrophils (Neu) as CD45⁺CD11b⁺F4/80⁻LY6G⁺; Monocyte (Mon) as CD45⁺CD11b⁺F4/80⁻LY6G⁻LY6C⁺; B cells as CD45⁺CD11b^-CD3^-B220⁺; Natural Killer cell (NK) as CD45⁺CD11b⁻B220⁻CD3⁻NK1.1⁺; CD4⁺ T cell as CD45⁺CD11b⁻B220⁻CD3⁺CD4⁺; CD8⁺ T cell as CD45⁺CD11b⁻B220⁻CD3⁺CD8⁺; Classic Dendritic cell (cDC) as CD45⁺CD11b⁺CD11c⁺MHCII⁺; plasma DC(pDC) as CD45⁺ CD11b^-LY6G^-F4/80^-CD3^-B220^-NK1.1^-CD11c⁺; ILC2 as CD45⁺CD11b^-CD3^-B220^-NK1.1^-CD11c^- CD117⁺; Mast cells as CD45⁺CD11b^-CD3^-B220^-NK1.1^-CD11c^-CD117^hiMHCII^low (Supplementary Data 1). Flow cytometric analyses were performed using Aurora CS (Cytek) or CytoFLEX LX (Beckman). Flow data were analyzed with FlowJo V10 (version 10.8.1). All gating strategies are provided in Supplementary Fig. 9.

RNA extraction and qRT-PCR

Total RNA was isolated using TRIzol reagent (#15596018CN, Thermo Fisher) from indicated cells or tissues. cDNA was generated by PrimeScript RT reagent Kit with gDNA Eraser (Perfect Real Time; #RR047A, Takara). Real-time quantitative PCR was performed using TB Green PCR reagents (#RR420A, Takara) and Roche LightCycler 480 II. Actb or Rps18 were used as internal controls for normalization. Oligonucleotide primer pairs used in this study are listed in Supplementary Table 1.

Immunoblot analysis and antibodies

Protein extracts from cells were prepared by homogenization in RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris-HCl, pH 7.4) containing complete protease-inhibitor cocktail (#P8340, Sigma). Protein extracts were separated by SDS-PAGE and then transferred onto a PVDF membrane (#IPVH00010, Millipore). After incubation with the indicated antibodies, blots were developed with SuperSignal West Pico Chemiluminescent substrate (#34580, Thermo Fisher) or Immobilon Western Chemiluminescent HRP substrate (#WBKLS0100, Millipore). The primary antibodies against P65 (#A22676, 1:1000), Phos-P65 (#AP1294, 1:1000) were purchased from Aclonal, and β-Tubulin (#2128S, 1:1000), IKKα/β (#96794, 1:1000), Phos-IKKα/β (#2697, 1:1000) were purchased from Cell Signaling Technology (Supplementary Table 5). Validation information is available on the manufacturers’ websites. Unprocessed western blots were presented in the Source Data File.

Histology and immunofluorescence staining

For histological analysis, femoral artery samples were fixed in 10% formalin for 2 days and then embedded in paraffin. The paraffin-embedded samples were sectioned into 5 μm thick slices and mounted on adhesion microscope slides for hematoxylin-eosin (H&E) staining. All images were acquired using an EVOS AMF7000 microscope.

For immunofluorescence staining, antigen retrieval of the rehydrated sections was performed using Antigen Retriever Tris-EDTA (#G1206, Servicebio). Sections were then blocked with PBS supplemented with 3% goat serum (#16210064, Thermo Fisher) and 0.1% Triton X-100 (#9002-93-1, Solarbio) at room temperature for 1 hour, followed by incubation with primary antibodies at 4 °C overnight. The sections were washed three times with PBS, and incubated with Alexa Fluor-conjugated secondary antibodies at room temperature for 1 hour, and then mounted with DAPI-containing antifade mounting medium (#S2110, Solarbio). The primary antibodies and their working concentrations were as follows: αSMA (#ab7817, Abcam, 1:400), CD45 (#ab208022, Abcam, 1:400), IL-33 (#ab187060, Abcam, 1:400), and MAC2 (#CL8942AP, Cedarlane, 1:800) (Supplementary Table 5). EdU click staining was performed using the BeyoClick™ EdU Cell Proliferation Kit (#C0075, Beyotime) according to the manufacturer’s instruction. Images were acquired using a Zeiss LSM800 confocal microscope and analyzed with ImageJ (NIH, https://imagej.nih.gov/ij/) or Photoshop software.

Cellular assays

For bone marrow-derived macrophages (BMDMs) preparation⁶⁴, femurs and tibias from 8 ~ 12-week-old male mice were isolated to extract bone marrow. Macrophage precursors were isolated using a Ficoll-Paque PLUS Media (#17144003, Cytiva) gradient and cultured in BMDM growth medium, which consisted of DMEM containing 10% FBS, 10 ng/ml M-CSF (#315-02, PeproTech), and 1% penicillin/streptomycin, on bacteriological petri dishes (#351029, Corning) for 7 days. The culture medium was refreshed every two days. Mature BMDMs were then maintained in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin for subsequent experiments.

For primary mesenchymal stromal cells (MSCs) preparation, MSCs were retrieved from stromal vascular fractions of murine inguinal white adipose tissue (WAT) using FACS²². In brief, depots of inguinal WAT were minced thoroughly before being incubated for 80 min in 10 mL digestion buffer (1XHBSS, 1.5%BSA and 1 mg/ml Collagenase D (#11088882001, Roche) at 37 °C in a shaking water bath. The digested mixture was placed on the ice immediately after digestion and then filtered through a 100 μm cell strainer into 20 mL pre-chilled 2% FBS/PBS. After centrifugation (600 g for 5 min at 4 °C), the stromal pellet was resuspended and subjected to erythrocyte lysis. Cells were then filtered again through a 40 μm cell strainer before centrifugation at 600 g for 5 min at 4 °C. After centrifugation, cells were resuspended in pre-cooled FACS buffer and incubation on ice before flow antibodies incubation. For sorting PDGFRβ + MSCs, primary antibodies were added to the cells and incubated at 4 °C for 15 min. The cells then were resuspended in 2% FBS/PBS for sorting using a BD Bioscience FACSAria cytometer at Zhejiang University School of Medicine. The primary antibodies and the dilution were as follows: CD45-PerCP/Cyanine5.5 (#103132, Biolegend, 1:400); CD31-PerCP/Cyanine5.5 (#102420, Biolegend, 1:400); PDGFRβ-PE (#136006, Biolegend, 1:75).

For siRNA transfection, primary MSCs and BMDMs were transfected with siRNAs using BEX CUY21EDIT II Electroporator with the pulse conditions of 240 V, 4 ms, 50Ω. All siRNAs in this study were synthesized at GenePharma Co. Ltd and their sequences are listed above or siRela (sense 5’-GCGACAAGGUGCAGAAAGATT-3’ and antisense 5’- UCUUUCUGCACCUUGUCGCTT-3’),

For the preparation of conditioned media, the cell lines of human umbilical vein EC (HUVECs) and mouse brain EC (bEND.3) were obtained from the National Collection of Authenticated Cell Cultures and kindly provided by Dr. Yong Liu (Wuhan University), and cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin. For the preparation of conditioned media (CM), HUVECs were subjected to mechanical injury through being scraped for several times. The remained cells were washed with PBS, and the fresh media were added for an additional incubation for 24 hours. The culture media were then collected and diluted with an equal volume of fresh media to generate the conditioned media (CM). For coculture assays, primary MSCs were incubated in the indicated CMs for 24 hours before further analyses. The CM were also subjected to be boiled at 95 °C for 5 minutes, or filtered using Vivaspin protein concentrator spin columns (#11668019, Cytiva) for the fractions of Frac^>10K and Frac^<10K. For secretome study, conditioned medium (CM) derived from the bEND.3 cells were collected, concentrated, and processed for protein profiling via high-resolution liquid chromatography-tandem mass spectrometry (LC − MS/MS). The quantified data of detected proteins were listed in Supplementary Data 2.

For Chromatin Immunoprecipitation (ChIP) assays⁴⁴, primary MSCs were cultured in 10 cm dishes until reaching 90% confluency. The cells were cross-linked by incubating in 1% formaldehyde at room temperature for 10 min, followed by the addition of 125 mM glycine and incubation for 2 minutes at 4 °C to quench the reaction. Cells were resuspended in 1 mL of Farnham lysis buffer (5 mM PIPES pH 8.0, 85 mM KCl, 0.5% NP-40, 1 mM DTT, and Protease inhibitor cocktail (#P8340, Sigma) for 10 minutes to isolate nuclear material. Crude nuclear pellets were re-suspended with lysis buffer (50 mM Tris-HCl pH 7.9, 1% SDS, 10 mM EDTA, 1 mM DTT, and Protease inhibitor cocktail). Lysates were sonicated into 150–300 bp fragments using the setting of 6 cycles of 30 s on/60 s off with a Bioruptor (Diagenode) at maximum power. Lysates were centrifuged at 13,000 x g for 10 min at 4 °C to remove cell debris, then the supernatant were collected and diluted 1:10 with dilution buffer (20 mM Tris-HCl pH 7.9, 0.5% Triton X-100, 2 mM EDTA, 150 mM NaCl, 1 mM DTT, and Protease inhibitor cocktail). Chromatin was pre-cleared by adding 20 μL per mL Protein G Sepharose™ 4 Fast Flow (#17-0618-01, GE Healthcare Bio-sciences) for 1 h at 4 °C and then incubated with anti-p65 antibody overnight (#8242, Cell Signaling technology) at 4 °C. Following overnight incubation, 80 μL washed Pierce protein G beads were added and incubated for 2 h at RT. Beads were captured by centrifugation and then washed with low salt wash buffer (20 mM Tris-HCl pH 7.9, 2 mM EDTA, 125 mM NaCl, 0.05% SDS, 1% Triton X-100, and Protease inhibitor cocktail), high salt wash buffer (20 mM Tris-HCl pH 7.9, 2 mM EDTA, 500 mM NaCl, 0.05% SDS, 1% Triton X-100, and Protease inhibitor cocktail), LiCl wash buffer (10 mM Tris-HCl pH 7.9, 1 mM EDTA, 250 mM LiCl, 1% NP-40, 1% sodium deoxycholate, and Protease inhibitor cocktail), and 1x Tris-EDTA (TE). Chromatin was eluted by elution buffer (100 mM sodium bicarbonate, 1% SDS) for 15 min at RT. Crosslinks were reversed through addition of 5 M NaCl and incubation at 65 °C overnight. RNA and protein were removed by incubation with RNase (#11119915001, Roche) and proteinase K (#EO0491, ThermoFisher Scientific). DNA was purified through Phenol:Chloroform:IsoAmyl (25:24:1) extraction and Isopropanol precipitation. Purified DNA was resuspended in 20 μL ddH2O and subjected to qPCR (ChIP-qPCR). The amount of each DNA sequence in ChIP samples was calculated relative to input using the ΔCt method. The sequences of primers for ChIP-qPCR are B.S._1 (Forward 5’-GGCTTGGAGCTCTCATTATCT-3’ and Reverse 5’-GACTAGTAAGAGCATCATGCAC-3’), B.S._2 (Forward 5’-TTCAACCTAGAGTCAGGCTG-3’ and Reverse 5’-TGCCTGCAACTTTATCGC-3’), B.S._3 (Forward 5’-GCCTCACCTGATATTTTATGTGG-3’ and Reverse 5’-AAAGCTGAAAATGGGGCTC-3’).

Bulk RNA sequencing and analysis

Total RNA was extracted using TRIzol reagent (#15596018CN, Thermo Fisher) from indicated cells isolated from the femoral arteries using FACS, and was eluted in 10 µL of sterile water. Bulk RNA sequencing was performed by Novogene Co., Ltd. (Beijing, China). Briefly, SMARTer Ultra Low RNA Kit (#634936, Takara) was used to synthesize cDNA. After library amplification, cDNA was tagmented and bar coded using TruePrep DNA Library Prep Kit V2 (#TD503-02, Vazyme). The quantity and quality of final libraries were evaluated using the Qubit 2.0 Fluorometer and Agilent 2100 bioanalyzer. Then cDNA libraries were pooled and sequenced on the Illumina Novaseq 6000 platform using a 150 bp paired-end modality. The sequencing process utilized the Sequencing by Synthesis (SBS) method, employing four fluorescently labeled dNTPs, DNA polymerase, and sequencing primers.

The process of raw data was performed on Galaxy platform (https://usegalaxy.org/). Quality Control was performed with FastQC (Version 0.12.1). After trimming low quality sequences using Cutadapt (Version 4.9), RNA STAR (Version 2.7.11a) was used to map reads to the mouse reference genome (GRCm38/mm10). Gene expression quantification was performed using featureCounts (Version 2.0.3) against mouse Gencode version 20 annotations. Differential expression analysis was conducted with the DESeq2 package (version 1.42.0) in R, and p-values were adjusted using the Benjamini-Hochberg method to manage the false discovery rate, ensuring statistically rigorous detection of differentially expressed genes. The differentially expressed genes (DEGs) identified between different groups were analyzed for pathway enrichment. Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses, and Gene Set Enrichment Analysis (GSEA) were performed using the ClusterProfiler package (version 4.10.1) in R. All bulk RNA-seq data have been deposited to GEO (GSE281009).

Single-cell RNA sequencing and analyses

For single-cell RNA sequencing (scRNA-seq), arterial cell isolation from murine femoral arteries at 2 days (D2) and 7 days (D7) post-surgery or Sham operations was performed as described above. The resuspended cells were stained with 10 μg/mL DAPI (#BL105A, Biosharp, 1:1000), filtered through a 40-μm cell strainer, and subjected to fluorescence-activated cell sorting (FACS) using a Cytek Aurora CS to isolate singlets. For each experimental condition, alive cells from 6 mice of the FAI or Sham group were pooled to create one sample per group. Single-cell RNA sequencing (scRNA-seq) was conducted by Berry Genomics (Beijing, China) using the Chromium Single Cell 3’ Reagent Kits v3 (10X Genomics) in accordance with the manufacturer’s protocol. Briefly, cells with a viability of over 85% were washed with 0.04% BSA in DPBS and resuspended to a concentration of 700–1200 cells/μL. The cells were then encapsulated in droplets, each containing a uniquely barcoded bead, enabling the reverse transcription of mRNA into cDNA. The cDNA libraries were subsequently amplified and sequenced on the Illumina NovaSeq platform. Raw sequencing data were processed using Illumina’s bcl2fastq converter to generate FASTQ files for further analysis. All scRNA-seq data have been deposited to GEO (GSE281008).

For scRNA-seq data analysis and visualization, the sequencing data were processed into gene expression matrix with the CellRanger (version 7.2.0, 10X Genomics). Downstream analysis of the data was performed with the R package Seurat (Version 5.0.1)⁶⁵. Genes expressed in fewer than three cells were removed. Cells with >300 expressed genes, <6,000 expressed genes, mitochondrial gene content <10%, and unique molecular identifier (UMI) counts <40,000 were retained for further analysis, yielding to 27,515 cells. Samples of the Sham, FAI_D2, and FAI_D7 conditions were integrated using R package harmony (Version 1.2.0)⁶⁶. Gene expression counts were normalized using Log-normalization with a scale factor of 10,000, and the top 2,000 genes were selected for individual sample through the variance stabilizing transformation method. Principal components analysis (PCA) was performed and unsupervised clustering was visualized in Uniform Manifold Approximation and Projection (UMAP) plots with a resolution of 0.1, based on the top ten principal components. Cell clusters were then annotated based on the well-established markers. Differentially expressed genes among the identified cell clusters were determined by FindMarkers function from the Seurat package. For pseudo-temporal single-cell trajectory analysis, Monocle (Version 2.30.1) and Monocle3 (Version 1.3.7) was utilized to investigate developmental trajectories. CellChat (Version 2.1.2)⁶⁷ was used to identify cell-cell communications, and cell differentiation states were predicted by CytoTRACE package (Version 0.3.3)⁶⁸. For cell-cell communication analysis, the numbers, strengths, and probabilities of cell-cell interactions across cell types were quantitatively computed using CellChat (Version 2.1.2) based on the curated ligand-receptor interaction database (CellChatDB). The information on the analysis results was provided in the Supplementary Data 4 and 5. Lipid metabolism signature score was computed using the function of AddModuleScore. The gene sets used for the pathway enrichment analyses in this study were listed in the Supplementary Data 6 and 7. A QC summary of the single-cell RNA sequencing was provided in Supplementary table 4.

Human study

This study enrolled patients who underwent one-year angiographic follow-up after second-generation DES implantation at the Department of Cardiology, Affiliated Hangzhou First People’s Hospital of Westlake University between July and October 2025. The cohort consisted of 25 patients with in-stent restenosis (ISR), define as ≥50% diameter stenosis in the stented segment or its immediate 5 mm margins, and 25 control patients without ISR. Patients with hepatic, renal disease, malignancy, acute inflammation, or age<18 or >80 years were excluded. All samples were obtained from patients after receipt of written informed consent in accordance with the Declaration of Helsinki. The study protocol was approved by the ethics committee of Affiliated Hangzhou First People’s Hospital of Westlake University (approval ID-2025ZN282-1). The basic demographic information of all participants enrolled in this study iis presented in Supplementary Table 3. Plasma IL-33 and OPN protein levels were determined by a commercially available human IL33 ELISA Kit (#EK133, Multi sciences) and human OPN ELISA Kit (#EK1135, Multi Sciences) according to the manufacturer’s instructions.

Statistical analysis

Statistical analysis was carried out as indicated in the figure legends. All data are presented as the mean ± s.e.m. unless otherwise indicated in the figure legends. Data variance was examined by F test or Bartlett’s test. The data meet the assumptions of the indicated statistical analysis. All tests were performed as two sided. A p value less than 0.05 was considered statistically significant. All statistical analyses were performed using Microsoft Excel or GraphPad Prism 8.0 (GraphPad Software). All statistical information, including p values, sample sizes and repetitions, are provided in Supplementary Data 3.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All reagents generated or used in this study are available on request from the corresponding author with a completed Material Transfer Agreement and the requests of materials should be addressed to Bo Shan. Single-cell RNA-seq and bulk RNA-seq data have been deposited to Gene Expression Omnibus under the accession numbers GSE281008 and GSE281009. Previously published data are accessed in Gene Expression Omnibus under the accession numbers GSE151653, GSE182232, and GSE19136 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc= GSE19136). The proteomics data generated in this study are provided in the Supplementary Information. Source data are provided with this paper. All data are included in the Supplementary Information or available from the authors, as are unique reagents used in this Article. The raw numbers for charts and graphs are available in the Source Data file whenever possible. Model figures were created in BioRender. Shan, B. (2026) https://BioRender.com/7d1cyua and https://BioRender.com/7d1cyua Source data are provided with this paper.

References

Wynn, T. A. & Vannella, K. M. Macrophages in tissue repair, regeneration, and fibrosis. Immunity 44, 450–462 (2016).
Article CAS PubMed PubMed Central Google Scholar
Hoeffel, G. et al. Sensory neuron-derived TAFA4 promotes macrophage tissue repair functions. Nature 594, 94–99 (2021).
Article ADS CAS PubMed Google Scholar
Bosurgi, L. et al. Macrophage function in tissue repair and remodeling requires IL-4 or IL-13 with apoptotic cells. Science 356, 1072–1076 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Gieseck, R. L. 3rd, Wilson, M. S. & Wynn, T. A. Type 2 immunity in tissue repair and fibrosis. Nat. Rev. Immunol. 18, 62–76 (2018).
Article CAS PubMed Google Scholar
Watanabe, S., Alexander, M., Misharin, A. V. & Budinger, G. R. S. The role of macrophages in the resolution of inflammation. J. Clin. Invest 129, 2619–2628 (2019).
Article PubMed PubMed Central Google Scholar
Rodriguez-Morales, P. & Franklin, R. A. Macrophage phenotypes and functions: resolving inflammation and restoring homeostasis. Trends Immunol. 44, 986–998 (2023).
Article CAS PubMed PubMed Central Google Scholar
Patsalos, A. et al. A growth factor-expressing macrophage subpopulation orchestrates regenerative inflammation via GDF-15. J. Exp. Med. 219, 10210 (2022).
Tabas, I. & Bornfeldt, K. E. Macrophage phenotype and function in different stages of atherosclerosis. Circ. Res. 118, 653–667 (2016).
Article CAS PubMed PubMed Central Google Scholar
Evans, C. E., Iruela-Arispe, M. L. & Zhao, Y. Y. Mechanisms of endothelial regeneration and vascular repair and their application to regenerative medicine. Am. J. Pathol. 191, 52–65 (2021).
Article CAS PubMed Google Scholar
Luscher, T. F. et al. Drug-eluting stent and coronary thrombosis: biological mechanisms and clinical implications. Circulation 115, 1051–1058 (2007).
Article PubMed Google Scholar
Swirski, F. K. & Nahrendorf, M. Cardioimmunology: the immune system in cardiac homeostasis and disease. Nat. Rev. Immunol. 18, 733–744 (2018).
Article CAS PubMed Google Scholar
Chen, W., Schilperoort, M., Cao, Y., Shi, J., Tabas, I. & Tao, W. Macrophage-targeted nanomedicine for the diagnosis and treatment of atherosclerosis. Nat. Rev. Cardiol. 19, 228–249 (2022).
Article PubMed Google Scholar
Zhu, Q. et al. CIDEC/FSP27 exacerbates obesity-related abdominal aortic aneurysm by promoting perivascular adipose tissue inflammation. Life Metab. 4, loae035 (2025).
Article PubMed Google Scholar
Shang, M. et al. Macrophage-derived glutamine boosts satellite cells and muscle regeneration. Nature 587, 626–631 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Chi, Z. et al. Gasdermin D-mediated metabolic crosstalk promotes tissue repair. Nature 634, 1168–1177 (2024).
Article ADS CAS PubMed Google Scholar
Park, M. D., Silvin, A., Ginhoux, F. & Merad, M. Macrophages in health and disease. Cell 185, 4259–4279 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ghamangiz, S., Jafari, A., Maleki-Kakelar, H., Azimi, H. & Mazloomi, E. Reprogram to heal: macrophage phenotypes as living therapeutics. Life Sci. 371, 123601 (2025).
Article CAS PubMed Google Scholar
Li, N. & Horng, T. A metabolic conspiracy drives anti-tumorigenic macrophages. Life Metab. 2, load009 (2023).
Article PubMed PubMed Central Google Scholar
Dwyer, G. K., D’Cruz, L. M. & Turnquist, H. R. Emerging Functions of IL-33 in Homeostasis and Immunity. Annu Rev. Immunol. 40, 15–43 (2022).
Article CAS PubMed Google Scholar
Morimoto, Y. et al. Amphiregulin-Producing Pathogenic Memory T Helper 2 Cells Instruct Eosinophils to Secrete Osteopontin and Facilitate Airway Fibrosis. Immunity 49, 134–150 e136 (2018).
Article CAS PubMed Google Scholar
Faas, M. et al. IL-33-induced metabolic reprogramming controls the differentiation of alternatively activated macrophages and the resolution of inflammation. Immunity 54, 2531–2546 e2535 (2021).
Article CAS PubMed PubMed Central Google Scholar
Shan, B., Shao, M., Zhang, Q., An, Y. A., Vishvanath, L. & Gupta, R. K. Cold-responsive adipocyte progenitors couple adrenergic signaling to immune cell activation to promote beige adipocyte accrual. Genes Dev. 35, 1333–1338 (2021).
Article CAS PubMed PubMed Central Google Scholar
Spallanzani, R. G. et al. Distinct immunocyte-promoting and adipocyte-generating stromal components coordinate adipose tissue immune and metabolic tenors. Sci. Immunol. 4, 3658 (2019).
Mahlakoiv, T. et al. Stromal cells maintain immune cell homeostasis in adipose tissue via production of interleukin-33. Sci. Immunol. 4, 0416 (2019).
Yang, X. et al. Very-low-density lipoprotein receptor-enhanced lipid metabolism in pancreatic stellate cells promotes pancreatic fibrosis. Immunity 55, 1185–1199 (2022).
Article CAS PubMed Google Scholar
Dahlgren, M. W. et al. Adventitial stromal cells define group 2 innate lymphoid cell tissue niches. Immunity 50, 707–722 e706 (2019).
Article CAS PubMed PubMed Central Google Scholar
Soliman, H. et al. Multipotent stromal cells: one name, multiple identities. Cell Stem Cell 28, 1690–1707 (2021).
Article CAS PubMed Google Scholar
Wang, Y., Fang, J., Liu, B., Shao, C. & Shi, Y. Reciprocal regulation of mesenchymal stem cells and immune responses. Cell Stem Cell 29, 1515–1530 (2022).
Article CAS PubMed Google Scholar
Buechler, M. B., Fu, W. & Turley, S. J. Fibroblast-macrophage reciprocal interactions in health, fibrosis, and cancer. Immunity 54, 903–915 (2021).
Article CAS PubMed Google Scholar
Tang, J. et al. Arterial Sca1(+) vascular stem cells generate de novo smooth muscle for artery repair and regeneration. Cell Stem Cell 26, 81–96 (2020).
Article CAS PubMed Google Scholar
Jolly, A. J. et al. Heterogeneous subpopulations of adventitial progenitor cells regulate vascular homeostasis and pathological vascular remodelling. Cardiovasc. Res. 118, 1452–1465 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kuwabara, J. T. & Tallquist, M. D. Tracking adventitial fibroblast contribution to disease: a review of current methods to identify resident fibroblasts. Arterioscler Thromb. Vasc. Biol. 37, 1598–1607 (2017).
Article CAS PubMed PubMed Central Google Scholar
Abedin, M., Tintut, Y. & Demer, L. L. Mesenchymal stem cells and the artery wall. Circ. Res. 95, 671–676 (2004).
Article CAS PubMed Google Scholar
Plikus, M. V. et al. Fibroblasts: origins, definitions, and functions in health and disease. Cell 184, 3852–3872 (2021).
Article CAS PubMed PubMed Central Google Scholar
Gao, L. et al. A resident stromal cell population actively restrains innate immune response in the propagation phase of colitis pathogenesis in mice. Sci. Transl. Med. 13, 5071 (2021).
Tellides, G. & Pober, J. S. Inflammatory and immune responses in the arterial media. Circ. Res. 116, 312–322 (2015).
Article CAS PubMed Google Scholar
Tinajero, M. G. & Gotlieb, A. I. Recent developments in vascular adventitial pathobiology: the dynamic adventitia as a complex regulator of vascular disease. Am. J. Pathol. 190, 520–534 (2020).
Article CAS PubMed Google Scholar
Jiang, L. et al. Nonbone marrow CD34(+) Cells are crucial for endothelial repair of injured artery. Circ. Res. 129, e146–e165 (2021).
Article CAS PubMed Google Scholar
Sata, M. et al. A mouse model of vascular injury that induces rapid onset of medial cell apoptosis followed by reproducible neointimal hyperplasia. J. Mol. Cell Cardiol. 32, 2097–2104 (2000).
Article CAS PubMed Google Scholar
Chen, R. et al. Macrophages in cardiovascular diseases: molecular mechanisms and therapeutic targets. Signal Transduct. Target Ther. 9, 130 (2024).
Article PubMed PubMed Central Google Scholar
Yap, J., Irei, J., Lozano-Gerona, J., Vanapruks, S., Bishop, T. & Boisvert, W. A. Macrophages in cardiac remodelling after myocardial infarction. Nat. Rev. Cardiol. 20, 373–385 (2023).
Article PubMed Google Scholar
Lim, H. Y. et al. Hyaluronan receptor LYVE-1-expressing macrophages maintain arterial tone through hyaluronan-mediated regulation of smooth muscle cell collagen. Immunity 49, 326–341 (2018).
Article CAS PubMed Google Scholar
Yaghi, O. K. et al. A discrete ‘early-responder’ stromal-cell subtype orchestrates immunocyte recruitment to injured tissue. Nat. Immunol. 24, 2053–2067 (2023).
Article CAS PubMed PubMed Central Google Scholar
Shan, B. et al. Perivascular mesenchymal cells control adipose-tissue macrophage accrual in obesity. Nat. Metab. 2, 1332–1349 (2020).
Article CAS PubMed PubMed Central Google Scholar
Molofsky, A. B., Savage, A. K. & Locksley, R. M. Interleukin-33 in Tissue Homeostasis, Injury, and Inflammation. Immunity 42, 1005–1019 (2015).
Article CAS PubMed PubMed Central Google Scholar
Liew, F. Y., Girard, J. P. & Turnquist, H. R. Interleukin-33 in health and disease. Nat. Rev. Immunol. 16, 676–689 (2016).
Article CAS PubMed Google Scholar
Konishi, H. et al. Platelets activated by collagen through immunoreceptor tyrosine-based activation motif play pivotal role in initiation and generation of neointimal hyperplasia after vascular injury. Circulation 105, 912–916 (2002).
Article CAS PubMed Google Scholar
Frismantiene, A., Philippova, M., Erne, P. & Resink, T. J. Smooth muscle cell-driven vascular diseases and molecular mechanisms of VSMC plasticity. Cell Signal 52, 48–64 (2018).
Article CAS PubMed Google Scholar
Deng, Y. et al. Alternate day fasting aggravates atherosclerosis through the suppression of hepatic ATF3 in Apoe(-/-) mice. Life Metab. 3, loae009 (2024).
Article PubMed PubMed Central Google Scholar
Farooq, V., Gogas, B. D. & Serruys, P. W. Restenosis: delineating the numerous causes of drug-eluting stent restenosis. Circ. Cardiovasc Inter. 4, 195–205 (2011).
Article Google Scholar
Weinberger, T. et al. Ontogeny of arterial macrophages defines their functions in homeostasis and inflammation. Nat. Commun. 11, 4549 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Al-Rifai, R. et al. JAK2V617F mutation drives vascular resident macrophages toward a pathogenic phenotype and promotes dissecting aortic aneurysm. Nat. Commun. 13, 6592 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Adachi, Y. et al. Beiging of perivascular adipose tissue regulates its inflammation and vascular remodeling. Nat. Commun. 13, 5117 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Ensan, S. et al. Self-renewing resident arterial macrophages arise from embryonic CX3CR1(+) precursors and circulating monocytes immediately after birth. Nat. Immunol. 17, 159–168 (2016).
Article CAS PubMed Google Scholar
Swirski, F. K., Robbins, C. S. & Nahrendorf, M. Development and function of arterial and cardiac macrophages. Trends Immunol. 37, 32–40 (2016).
Article CAS PubMed PubMed Central Google Scholar
Liu, P., Patil, S., Rojas, M., Fong, A. M., Smyth, S. S. & Patel, D. D. CX3CR1 deficiency confers protection from intimal hyperplasia after arterial injury. Arterioscler. Thromb. Vasc. Biol. 26, 2056–2062 (2006).
Article CAS PubMed Google Scholar
Giannakis, N. et al. Dynamic changes to lipid mediators support transitions among macrophage subtypes during muscle regeneration. Nat. Immunol. 20, 626–636 (2019).
Article CAS PubMed PubMed Central Google Scholar
Burgess, M., Wicks, K., Gardasevic, M. & Mace, K. A. Cx3CR1 expression identifies distinct macrophage populations that contribute differentially to inflammation and repair. Immunohorizons 3, 262–273 (2019).
Article CAS PubMed Google Scholar
Xiong, D., Wang, Y. & You, M. A gene expression signature of TREM2(hi) macrophages and gammadelta T cells predicts immunotherapy response. Nat. Commun. 11, 5084 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Hulsmans, M. et al. Recruited macrophages elicit atrial fibrillation. Science 381, 231–239 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Dib, L. et al. Lipid-associated macrophages transition to an inflammatory state in human atherosclerosis increasing the risk of cerebrovascular complications. Nat. Cardiovasc. Res. 2, 656–672 (2023).
Article CAS PubMed PubMed Central Google Scholar
Hong, Y. et al. A strongly adhesive hemostatic hydrogel for the repair of arterial and heart bleeds. Nat. Commun. 10, 2060 (2019).
Article ADS PubMed PubMed Central Google Scholar
Saludas, L., Pascual-Gil, S., Prosper, F., Garbayo, E. & Blanco-Prieto, M. Hydrogel based approaches for cardiac tissue engineering. Int. J. Pharm. 523, 454–475 (2017).
Article CAS PubMed Google Scholar
Shan, B. et al. The metabolic ER stress sensor IRE1alpha suppresses alternative activation of macrophages and impairs energy expenditure in obesity. Nat. Immunol. 18, 519–529 (2017).
Article CAS PubMed Google Scholar
Satija, R., Farrell, J. A., Gennert, D., Schier, A. F. & Regev, A. Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol. 33, 495–502 (2015).
Article CAS PubMed PubMed Central Google Scholar
Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods 16, 1289–1296 (2019).
Article CAS PubMed PubMed Central Google Scholar
Jin, S. et al. Inference and analysis of cell-cell communication using CellChat. Nat. Commun. 12, 1088 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Gulati, G. S. et al. Single-cell transcriptional diversity is a hallmark of developmental potential. Science 367, 405–411 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This work was supported by the grants from the National Key Research and Development Program of China (2022YFA1104102) to B. Shan, Leading Innovative and Entrepreneur Team Introduction Program of Zhejiang (2022R01002, 2024R01003) to B. Shan, the National Natural Science Foundation of China (32571287, 32371235, 81972674, 31900543, 82170891, 82270926, 32241008, 62161160312, 32471175) to B. Shan, Y. Wu, M. Shao and J. Huang, the National Natural Science Foundation of China (82188102), the Fundamental Research Funds for the Central Universities (226-2024-00024) to B. Shan, Zhejiang Provincial Program for the Cultivation of High-level Innovative Health Talents and Zhejiang Provincial Natural Science Foundation of China (LMS26H030004) to Y. Wu, and Shanghai Municipal Science and Technology Project (22140903200) to M. Shao, the Science and Technology Foundation of Zhejiang Province of China (2020C03018), the Science and Technology Foundation of Hangzhou of China (Z20210019) to J. Huang, the Noncommunicable Chronic Disease-National Science and Technology Major Project (2024ZD0530100) to M. Shao and B. Shan, and the Talent Plan of Shanghai Branch-Chinese Academy of Sciences (CASSHB-QNPD-2023-019) to M. Shao, the Tianfu Jincheng Laboratory (No. TFJCPI20250049). We thank Dr. Peng Shi (Zhejiang University) for providing the mouse strains of Cx3cr1^GFP, Cx3cr1^CreER, Rosa26-iDTR, Dr. Yao Xie (Zhejiang University) for providing MOVAS cell line and Chao Bi from the Core Facility, Zhejiang University School of Medicine for their technical support.

Author information

Authors and Affiliations

Zhejiang Key Laboratory of Pancreatic Disease, The First Affiliated Hospital, Zhejiang Key Laboratory of Frontier Medical Research on Cancer Metabolism, and Institute of Translational Medicine, Zhejiang University School of Medicine, Hangzhou, Zhejiang, China
Yan Ping, Zhewen Qin, Xingxiao Huang, Qiuli Sun, Tao Yao, Xin Tian, Bowen Diao, Lin Zhao & Bo Shan
Department of Cardiology, Westlake laboratory of Life Sciences and Biomedicine, Affiliated Hangzhou First People’s Hospital, Westlake University School of Medicine, Hangzhou, Zhejiang, China
Yan Ping, Xingxiao Huang, Qiuli Sun, Long Wang & Jinyu Huang
Department of Endocrinology, Children’s Hospital, Zhejiang University School of Medicine, National Clinical Research Center for Child Health, Hangzhou, Zhejiang, China
Zhewen Qin
Tongde Hospital of Zhejiang Province, Zhejiang Chinese Medical University, Hangzhou, Zhejiang, China
Tao Yao, Xin Tian, Xintong Xia & Ying Wu
University of Chinese Academy of Sciences, Beijing, China
Siqi Li & Danni Wei
Shanghai Institute of Immunity and Infection, Chinese Academy of Sciences, Shanghai, China
Siqi Li & Danni Wei
School of Basic Medical Sciences and Forensic Medicine, Hangzhou Medical College, Hangzhou, Zhejiang, China
Xinyuan Zhang
Key Laboratory of Immune Response and Immunotherapy, Shanghai Institute of Immunity and Infection, Chinese Academy of Sciences, Shanghai, China
Mengle Shao
Tianfu Jincheng Laboratory, 610212, Chengdu, China
Mengle Shao & Ying Wu
Institute of Fundamental and Transdisciplinary Research, Cancer Center, Zhejiang University, Hangzhou, Zhejiang, China
Bo Shan

Authors

Yan Ping
View author publications
Search author on:PubMed Google Scholar
Zhewen Qin
View author publications
Search author on:PubMed Google Scholar
Xingxiao Huang
View author publications
Search author on:PubMed Google Scholar
Qiuli Sun
View author publications
Search author on:PubMed Google Scholar
Tao Yao
View author publications
Search author on:PubMed Google Scholar
Siqi Li
View author publications
Search author on:PubMed Google Scholar
Xin Tian
View author publications
Search author on:PubMed Google Scholar
Xintong Xia
View author publications
Search author on:PubMed Google Scholar
Bowen Diao
View author publications
Search author on:PubMed Google Scholar
Lin Zhao
View author publications
Search author on:PubMed Google Scholar
Xinyuan Zhang
View author publications
Search author on:PubMed Google Scholar
Danni Wei
View author publications
Search author on:PubMed Google Scholar
Long Wang
View author publications
Search author on:PubMed Google Scholar
Mengle Shao
View author publications
Search author on:PubMed Google Scholar
Ying Wu
View author publications
Search author on:PubMed Google Scholar
Jinyu Huang
View author publications
Search author on:PubMed Google Scholar
Bo Shan
View author publications
Search author on:PubMed Google Scholar

Contributions

Y.P., Z.Q., Y.W., and B.S,. designed the study and performed most of the experiments and analyzed the data. X.H., Q.S., T.Y., S.L., X.T., X.X., B.D., L.Z., D.W., L.W., and X.Z. participated in collection and analysis of the data. J.H. and M.S. contributed reagents, technical support and provided scientific insight and discussion. B.S., M.S., and Y.W. conceptualized the study, interpreted the experiments and wrote the manuscript. B.S. and Y.W. supervised the work. All authors contributed to the manuscript and approved the submitted version.

Corresponding authors

Correspondence to Mengle Shao, Ying Wu, Jinyu Huang or Bo Shan.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Finosh Thankam and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. [A peer review file is available].

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Transparent Peer Review file (download PDF )

Reporting Summary (download PDF )

Supplementary Data 1 (download XLSX )

Supplementary Data 2 (download XLSX )

Supplementary Data 3 (download XLSX )

Supplementary Data 4 (download XLSX )

Supplementary Data 5 (download XLSX )

Supplementary Data 6 (download XLSX )

Supplementary Data 7 (download XLSX )

Source data

Source Data (download ZIP )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Ping, Y., Qin, Z., Huang, X. et al. Perivascular mesenchymal cells instruct ST2+ reparative macrophages to promote endovascular injury-induced neointimal hyperplasia in mice. Nat Commun 17, 3635 (2026). https://doi.org/10.1038/s41467-026-68587-x

Download citation

Received: 27 May 2025
Accepted: 08 January 2026
Published: 07 March 2026
Version of record: 21 April 2026
DOI: https://doi.org/10.1038/s41467-026-68587-x