The MCP-1/CCR2–CD16⁺ monocyte axis drives thrombus fibrosis in Chronic Thromboembolic Pulmonary Hypertension

Abstract

Objective

This study aimed to determine whether dysregulation of the MCP-1/CCR2 axis and expansion of CD16⁺ monocytes represent a mechanistic link between acute pulmonary embolism (APE) and the chronic fibrotic vascular occlusion characteristic of chronic thromboembolic pulmonary hypertension (CTEPH).

Approach and results

We performed an integrative multi-level analysis combining multicenter clinical cohorts, single-cell transcriptomics, and in vivo modeling. In a multicenter cohort (32 CTEPH patients and 20 healthy controls), circulating CD16⁺ monocytes were selectively expanded in CTEPH and correlated with disease severity, as assessed by pulmonary vascular resistance and mean pulmonary arterial pressure. MCP-1 levels were significantly elevated and positively associated with CD16⁺ monocyte expansion. Single-cell RNA sequencing of human APE thrombi (5 patients; 24,399 cells) revealed that CD16⁺ monocytes exhibit a profibrotic transcriptional program and function as key signaling hubs within CCL-mediated communication networks. Histological analysis further demonstrated spatial colocalization of MCP-1 expression and M2 macrophage infiltration in fibrotic thrombus regions. In a murine model of venous thrombosis recapitulating fibrotic thrombus remodeling, CCR2 inhibition with RS102895 suppressed the expansion of profibrotic monocytes, reduced M2 macrophage infiltration, and attenuated thrombus fibrosis.

Conclusion

These findings identify the MCP-1/CCR2–CD16⁺ monocyte axis as a central driver of fibrotic thrombus remodeling in CTEPH and provide a mechanistic framework linking acute thrombosis to chronic vascular occlusion. Targeting this pathway may represent a promising immunomodulatory strategy to limit pathological vascular remodeling.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-08283-y.

Introduction

The pathological progression of chronic thromboembolic pulmonary hypertension (CTEPH) revolves around two core mechanisms: chronic fibrosis of proximal pulmonary artery thrombi and pathological remodeling of distal pulmonary microvasculature. Anatomically, proximal thrombi manifest as dense networks of collagen fibers mixed with elastic fibers, calcifications, and inflammatory cell infiltration, which starkly contrast with the fresh thrombi seen in acute pulmonary embolism (APE) that are mainly composed of red blood cells and fibrin. The histological analysis of pulmonary endarterectomy (PEA) specimens reveals the complexity of the inflammatory microenvironment in CTEPH. Studies by Quarck et al. [1]. and Bochenek et al. [2]. have found significant macrophage enrichment in fibrotic thrombus regions, suggesting their close association with thrombus organization. In the first part of our study, single-cell RNA sequencing analysis of thrombus tissue from CTEPH patients post-PEA demonstrated that most macrophage subsets exhibit a pronounced pro-inflammatory phenotype, with high expression of inflammatory cytokines (such as IL-1β and TNF-α) and lipid metabolism-related genes (such as APOC1 and CH25H). This finding aligns with previous reports of atherosclerotic features in pulmonary arteries of pulmonary arterial hypertension (PAH) patients [3], suggesting that macrophage-mediated metabolic reprogramming may play a key role in pulmonary artery intimal atherosclerotic lesion progression, leading to vascular wall stiffening and progressive elevation of pulmonary arterial pressure, which may be a critical mechanism underlying hemodynamic deterioration in CTEPH.

From a disease progression perspective, CTEPH is considered a late complication of acute pulmonary thromboembolism. In most APE cases, thrombi can be completely resolved within three months through fibrinolytic processes mediated by immune cells. This process involves sequential changes in thrombus composition: from erythrocyte-rich thrombi gradually transitioning to chronic fibrotic thrombi rich in fibrin and collagen, accompanied by stepwise recruitment of leukocytes (such as neutrophils and monocytes) and neovascularization [4]. In the early stage of thrombus formation, neutrophils are rapidly recruited to the thrombus site through interleukin-8 (IL-8)-mediated chemotaxis and contribute to degradation of fibrin and collagen matrices by releasing matrix metalloproteinases (MMPs) and urokinase-type plasminogen activator (uPA) [4, 5]. Animal studies have shown that neutrophil depletion in inferior vena cava (IVC) ligation-induced venous thrombosis models results in increased thrombus volume and collagen deposition [6]. Additionally, neutrophils further recruit peripheral blood monocytes into thrombi by secreting monocyte chemoattractant protein-1 (MCP-1), which then differentiate into macrophages [7]. Early infiltrating macrophages display a pro-inflammatory phenotype involved in pathogen clearance; as the disease progresses, they gradually transition to a reparative phenotype, promoting extracellular matrix remodeling and blood flow restoration through secretion of MMP-9 and angiogenic factors such as vascular endothelial growth factor (VEGF) [4, 8]. This dynamic evolution suggests that monocytes occupy a pivotal node within the temporal regulatory network of thrombus formation.

Notably, our previous study found a significant increase in the proportion of the CD16⁺ monocyte subset in peripheral blood of CTEPH patients. Transcriptomic features of this subset show significant upregulation of genes related to platelet activation, coagulation response, and migratory capacity. Importantly, this subset demonstrates potential for differentiation into thrombus-infiltrating macrophages, indicating that CD16⁺ monocytes may be a critical cellular source driving pathological remodeling of the pulmonary artery intima.

As a key chemokine of the monocyte/macrophage system, MCP-1, a core member of the C-C chemokine family, drives monocyte migration and functional activation by binding to the C-C chemokine receptor type 2 (CCR2) [9]. In PAH and congestive heart failure, MCP-1 has been demonstrated to exacerbate tissue fibrosis by promoting monocyte/macrophage infiltration. For example, in a pressure overload-induced left ventricular hypertrophy model, MCP-1 is highly expressed in myocardial interstitial macrophages, with expression levels positively correlating with cardiac dysfunction [10]. Studies also report that peripheral blood MCP-1 levels in CTEPH patients correlate significantly with pulmonary vascular resistance (PVR) and decline synchronously with hemodynamic improvement after PEA surgery [11]. Immunohistochemical studies further demonstrate widespread MCP-1 expression in endothelial cells, macrophages, and newly formed intimal smooth muscle cells within CTEPH thrombus tissue, whereas normal pulmonary artery tissue shows little to no active expression, suggesting that upregulation is directly related to pathological mechanical forces (such as high shear stress) or inflammatory stimuli. Furthermore, macrophages may amplify inflammatory signals via autocrine or paracrine release of MCP-1, creating a positive feedback loop that drives vascular remodeling.

Although our previous single-cell sequencing study suggested a potential involvement of CD16⁺ monocytes in CTEPH [12], a critical knowledge gap remains regarding their precise functional role and dynamic evolution during thrombus organization. Specifically, it remains unclear how circulating monocyte subsets are recruited into thrombotic lesions, undergo phenotypic transition, and ultimately differentiate into profibrotic effector cells that drive vascular occlusion in CTEPH. While expansion of CD16⁺ monocytes has been reported in other inflammatory and hypoxia-related vascular diseases [13–16], their functional contribution within the context of immunothrombotic remodeling in CTEPH is poorly defined. In addition, although multiple inflammatory mediators are involved in immunothrombosis, the upstream regulatory pathways governing monocyte recruitment and phenotypic specification remain to be clarified. In this study, we focused on the MCP-1/CCR2 axis due to its well-established role in monocyte chemotaxis and its strong clinical association with pulmonary vascular resistance, suggesting that it may function as a central upstream driver rather than a secondary inflammatory signal.

To address this gap, we hypothesized that dysregulation of the MCP-1/CCR2 signaling pathway selectively drives the recruitment and expansion of CD16⁺ monocytes and promotes their transition toward a profibrotic macrophage phenotype, thereby orchestrating chronic thrombus remodeling in CTEPH. By integrating multicenter clinical cohorts, single-cell transcriptomic profiling, and in vivo experimental models, this study aims to delineate the cellular and molecular mechanisms underlying this process and to provide a conceptual framework for the development of targeted immunomodulatory strategies.

Methods

Study population and clinical criteria

This study was approved by the institutional ethics committees the First Affiliated Hospital of Fujian Medical University, the Affiliated Hospital of Jianghan University, and the China-Japan Friendship Hospital, and written informed consent was obtained from all participants.

A multicenter case-control study was conducted to enroll patients with CTEPH, patients with APE, and healthy controls. The diagnosis of CTEPH was definitively confirmed according to current guidelines, requiring a mean pulmonary arterial pressure (mPAP) greater than 20 mmHg measured by right heart catheterization at rest, mismatched perfusion defects on ventilation/perfusion (V/Q) lung scans, and specific diagnostic signs (e.g., ring webs, slits, or total chronic occlusions) confirmed by computed tomography pulmonary angiography (CTPA) or digital subtraction angiography following at least three months of effective therapeutic anticoagulation. For the APE group, patients were included if they presented with newly diagnosed symptomatic acute pulmonary embolism objectively confirmed by filling defects on CTPA and were candidates for catheter-directed thrombectomy. Age- and sex-matched healthy volunteers with no history of venous thromboembolism, cardiopulmonary diseases, or chronic illnesses—verified by routine physical examinations, electrocardiograms, and echocardiography—served as controls. Crucially, to eliminate potential confounding effects on peripheral circulating monocyte subsets and systemic inflammatory cytokines such as MCP-1, stringent universal exclusion criteria were applied across all groups. Patients were excluded if they presented with acute or chronic active infections, severe systemic autoimmune or connective tissue diseases, known active malignancies, other classifications of pulmonary hypertension (such as Group 1 PAH, Group 2, or Group 3), hematological disorders affecting leukocyte lineages, or a recent history of systemic corticosteroid or immunosuppressive agent use within the past three months.

Sample preparation and tissue dissociation

Pulmonary thrombus tissues were obtained via catheter-directed aspiration from 5 independent patients with APE and immediately immersed in ice-cold 0.9% saline to preserve cell viability. To minimize ischemic time and rigorously preserve the original in vivo transcriptomic state, tissue dissociation was strictly initiated within 1 h of aspiration. Under sterile conditions, thrombi were rinsed repeatedly with ice-cold PBS to remove residual blood, and subsequently minced into ~ 0.5 mm fragments on ice. Tissue digestion was performed using the GEXSCOPE^® Tissue Dissociation Solution (Singleron Biotechnologies). This optimized enzymatic formulation, combined with a strictly limited digestion window, was specifically employed to maintain high cell viability and minimize dissociation-induced transcriptomic artifacts or cellular stress responses. Briefly, minced samples were incubated at 37 °C with gentle agitation (150 rpm) for 30 min, with intermittent pipetting to facilitate dissociation. Digestion progress was monitored microscopically, and enzymatic treatment was terminated once no visible tissue fragments remained.

The resulting cell suspensions were triturated, passed through a 40 μm filter, and subjected to sequential washes in ice-cold PBS (300 g, 5 min, 4 °C). Red blood cell lysis was performed using pre-chilled RBC lysis buffer for 2 min on ice, followed by additional PBS washes. After centrifugation, cell pellets were resuspended in 500 µL PBS on ice, yielding single-cell suspensions for downstream single-cell RNA sequencing. Cell viability (> 85%) was confirmed by trypan blue exclusion.

Single-cell RNA sequencing and data processing

Raw sequencing reads were processed using CeleScope (v1.5.2, Singleron Biotechnologies) according to the manufacturer’s pipeline. After quality control filtering (cells with < 200 or > 5000 detected genes or > 20% mitochondrial transcripts were excluded), a total of 24,399 high-quality single-cell transcriptomes were retained. The 20% mitochondrial threshold was specifically adopted to preserve metabolically active and stress-responding immune cells inherent to the highly hypoxic thrombus microenvironment, thereby preventing the artificial depletion of disease-relevant populations.

To identify highly variable genes, the FindVariableFeatures function in Seurat (v5.1.0) was applied, returning the top 2000 most variable genes. Principal component analysis (PCA) was then performed using the RunPCA function, and the first 20 principal components were selected for downstream analysis. Batch effects across samples were corrected using Harmony (v1.2.0).

For dimensionality reduction and clustering, we used the FindNeighbors function in Seurat after SCTransform normalization, followed by RunUMAP for visualization. Clustering was performed at multiple resolutions (1.0–1.5), ultimately yielding 21 distinct clusters. To further ensure dataset integrity following computational preprocessing, an additional layer of doublet exclusion was performed during cluster annotation. Cells or clusters exhibiting aberrant co-expression of canonical markers from distinct lineages (e.g., CD3D and CD14) were carefully curated and removed, thereby minimizing the impact of residual doublets not captured by initial filtering.

Differentially expressed genes (DEGs) were identified using Seurat’s FindMarkers function (Wilcoxon likelihood-ratio test, default parameters). Genes expressed in > 10% of cells within a cluster and with log(fold change) > 0.25 were retained as DEGs.

Cell type annotation was based on canonical marker expression within DEGs, cross-referenced with published literature. To ensure robust validation, the top 10 defining differentially expressed genes for each identified cluster were systematically profiled and are provided in Supplementary Data 1. Expression patterns were visualized using DoHeatmap and DotPlot functions in Seurat.

Functional enrichment analyses of DEGs were conducted with clusterProfiler (v4.10.1) using the msigdbr database (https://www.gsea-msigdb.org/gsea/msigdb). Pathways with p.adj < 0.05 were considered significant. Gene Ontology (GO) enrichment included molecular function (MF), biological process (BP), and cellular component (CC), while KEGG pathway analysis was performed in parallel.

Finally, CellChat (v1.6.1) was applied to infer intercellular communication networks, and pseudotime trajectory analysis was performed using the RunSlingshot function from the SCP package. All analyses were performed in R (v4.2.2).

Peripheral blood mononuclear cell (PBMC) isolation and in vitro stimulation

Peripheral blood samples (5 mL) were collected into sterile EDTA anticoagulant tubes and processed within 2 h. PBMCs were isolated by density gradient centrifugation using Ficoll-Hypaque. Briefly, whole blood was diluted 1:1 with PBS, carefully layered onto 5 mL of separation medium, and centrifuged at 800 g for 25 min at room temperature. The mononuclear cell layer was harvested, washed twice with PBS (300 g, 5 min), and resuspended in PBS. Cell viability and counts were determined using an automated cell counter.

For in vitro stimulation assays, PBMCs from 6 independent healthy donors were resuspended in RPMI 1640 medium supplemented with 10% fetal bovine serum at a density of 1 × 10⁶ cells/mL and plated in 12-well plates (n = 6 independent biological replicates per experimental condition). Cells were treated as follows: (i) MCP-1 group: recombinant human MCP-1 (100 ng/mL). This specific concentration was selected based on preliminary dose-response optimizations (0–200 ng/mL) to ensure optimal stimulation within a non-toxic physiological window; (ii) MCP-1 + RS102895 group: pre-incubation with the CCR2 antagonist RS102895 (10 µM, 30 min) followed by MCP-1 (100 ng/mL); (iii) RS102895 group: RS102895 alone (10 µM); (iv) vehicle control: equal volumes of ddH₂O and 0.9% saline (solvent controls). Cells were incubated at 37 °C with 5% CO₂ for 24 h, harvested, and washed with staining buffer (PBS + 0.5% BSA) for downstream flow cytometry.

Flow cytometric analysis of monocyte subsets

PBMCs were stained with APC-conjugated anti-CD16 and FITC-conjugated anti-CD14 monoclonal antibodies for 30 min at room temperature in the dark, washed, and resuspended in PBS. 7-AAD was added immediately before acquisition to exclude dead cells. To strictly control inter-sample variability and ensure quantitative reproducibility, the flow cytometer was calibrated daily using standard tracking beads. Samples were analyzed on a flow cytometer, and data were processed with FlowJo v10.0 software.

The gating strategy for monocyte subsets was strictly standardized. Briefly, single cells were first isolated using forward scatter height versus area (FSC-H vs. FSC-A) to exclude doublets. Live cells were subsequently identified as 7-AAD–negative events on side scatter area (SSC-A) versus 7-AAD plots. Within the live cell population, total monocytes were definitively gated on SSC-A versus CD14 as CD14⁺ events, which effectively excluded other non-target myeloid populations and lymphocytes. Finally, the specific CD16⁺ monocyte subset (CD14⁺CD16⁺) was further delineated within this primary monocyte gate based on SSC-A versus CD16 expression. To precisely define these critical gating boundaries and rigorously eliminate background fluorescence artifacts, Fluorescence Minus One (FMO) controls were strictly utilized alongside appropriate routine controls (unstained, single-stained, and isotype controls).

Enzyme-linked immunosorbent assay (ELISA) for MCP-1

Plasma and pulmonary arterial blood MCP-1 concentrations were measured using a commercially available human MCP-1 ELISA kit (Bio-Swamp, catalog no. HM10894). The detection sensitivity of the assay was ≤ 3 pg/mL, and all samples, standards, and controls were measured in triplicate to ensure intra-assay precision, following the manufacturer’s instructions. For this specific quantitative analysis, samples were collected from a total cohort comprising 32 patients with CTEPH, 30 patients with APE, and 20 age- and sex-matched healthy controls. Briefly, standards and plasma samples were added to pre-coated 96-well plates, followed by incubation with HRP-conjugated detection antibodies at 37 °C for 60 min. After five washes with wash buffer, wells were developed with substrate solution (50 µL each of reagents A and B) at 37 °C in the dark for 15 min, and the reaction was stopped with 50 µL stop solution. Optical density (OD) was measured at 450 nm using a microplate reader, and MCP-1 concentrations were calculated from the standard curve.

Animal experiments

Animal grouping and treatment

Male BALB/c mice (8–10 weeks, n = 6 per group) were utilized in this study. Male mice were specifically selected to eliminate the known confounding effects of the female estrous cycle and fluctuating estrogen levels on coagulation and vascular remodeling [17, 18]. Furthermore, the sample size of 6 mice per experimental group was determined based on established IVC ligation protocols in prior literature. This number has been previously demonstrated to be statistically sufficient to detect significant differences in thrombus weight and fibrotic remodeling outcomes, while strictly adhering to the ethical principle of ‘Reduction‘ [19–22]. The animals were randomly assigned to six groups: (1) Sham; (2) IVC ligation; (3) Sham + 0.9% saline; (4) Sham + RS102895; (5) IVC + 0.9% saline; (6) IVC + RS102895. Sham-operated mice underwent laparotomy with IVC exposure but without ligation. IVC thrombosis was induced by ligating the IVC just below the left renal vein with 7 − 0 non-absorbable sutures, while visible side branches were ligated or cauterized. Following the surgical induction of thrombosis (designated as Day 0), mice allocated to the interventional group received daily intraperitoneal injections of the selective CCR2 antagonist RS102895 at a dose of 5 mg/kg/day. This strict administration protocol commenced exactly 24 hours postoperatively (Day 1) and was maintained continuously throughout the observation period. To avoid acute pharmacological interference during terminal sample collection, the daily regimen was concluded on the day immediately preceding euthanasia, ensuring no drug was administered on the day of tissue harvest. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Fujian Medical University (Approval No. IACUC FJMU 2022 − 0442) and strictly followed institutional and national guidelines for animal welfare and ethical conduct.

IVC ligation procedure

Mice were anesthetized with 1% pentobarbital sodium (0.5 mL/100 g, i.p.) and placed supine. After midline laparotomy, the intestines were exteriorized and covered with warm saline-moistened gauze. The retroperitoneum was gently dissected to expose the IVC and abdominal aorta. The IVC was carefully separated from surrounding connective tissue and ligated below the left renal vein without damaging the aorta. Side branches were tied with 7 − 0 sutures or coagulated by electrocautery. Sham animals underwent the same surgical exposure without IVC ligation. After hemostasis, intestines were repositioned, and the abdomen was closed in layers.

Sample collection

Peripheral blood was collected from the retro-orbital venous plexus at designated time points during the study and transferred into anticoagulant tubes; mice were returned to their cages after hemostasis and monitored until full recovery. At the experimental endpoint, mice were euthanized with an overdose of pentobarbital. The IVC was then exposed distal to the ligation site, and thrombi were carefully excised, immediately placed in cold saline, and measured for length, width, and weight using calipers and a precision balance. To rigorously prevent observer bias, all subsequent downstream outcome measurements, including histological quantifications and molecular analyses, were performed by independent investigators who were strictly blinded to the in vivo treatment allocations.

Flow cytometry of peripheral blood monocytes

Whole blood (100 µL) was stained with fluorochrome-conjugated monoclonal antibodies: FITC–anti-CD11b, PE–anti-Ly6C, and APC-Cy7–anti-CD45 (5 µL each). Samples were incubated for 30 min at room temperature in the dark, followed by red blood cell lysis (10–15 min, RT), washing, and resuspension in PBS containing 7-AAD for viability staining. Data were acquired on a flow cytometer and analyzed with FlowJo v10.0.

Histological and immunohistochemical analyses

Tissue processing and embedding

In this study, 8 human pulmonary endarterectomy PEA specimens were obtained from the patients with a definitive diagnosis of CTEPH. The diagnosis and surgical operability were rigorously established through exhaustive multidisciplinary preoperative evaluations, strictly including right heart catheterization (RHC) and pulmonary angiography. The organized chronic thrombus tissues were subsequently extracted during standard PEA surgeries. These clinical specimens, along with murine IVC thrombi were fixed in 4% paraformaldehyde for ≥ 24 h, dehydrated through graded ethanol, cleared in xylene, and embedded in paraffin. Section  (4 μm) were prepared using a microtome, mounted on glass slides, and baked at 60 °C for subsequent staining.

Hematoxylin and Eosin (H&E) Staining

Paraffin sections were deparaffinized, rehydrated, and stained with hematoxylin for nuclear visualization, followed by eosin counterstaining for cytoplasmic components. Sections were dehydrated, cleared, and mounted with neutral resin.

Masson’s trichrome and sirius red staining

To assess collagen deposition, adjacent sections were stained with Masson’s trichrome (collagen fibers in blue) or Sirius Red (collagen fibers in red under polarized light). Standard protocols were followed, including overnight incubation in primary dyes, differentiation, dehydration, and mounting.

Immunohistochemistry (IHC)

After deparaffinization and rehydration, antigen retrieval was performed in citrate buffer, followed by blocking of endogenous peroxidase with 3% hydrogen peroxide. Sections were blocked with 3% BSA and incubated overnight at 4 °C with primary antibodies against CD68, INOS, ARG1 and CCL2. After washing, HRP-conjugated secondary antibodies were applied, and signal was visualized using DAB substrate. Nuclei were counterstained with hematoxylin.

Immunofluorescence (IF)

Multiplex immunofluorescence was performed using a TSA amplification system. Sections underwent antigen retrieval, BSA blocking, and sequential incubation with primary antibodies (e.g., F4/80 and CD206) and HRP- or fluorophore-conjugated secondary antibodies. Signals were amplified with TSA fluorophores, and nuclei were counterstained with DAPI. Autofluorescence quenching was performed before mounting with antifade medium.

Image acquisition and analysis

Slides were scanned using the Motic EASYSCAN or Pannoramic MIDI digital slide scanners. Quantitative image analysis was initially performed using Motic DSAssistant and CaseViewer software to assess overall thrombus morphology and inflammatory infiltration. For precise quantitative analysis, regions of interest (ROIs) were selected using a randomized approach to ensure representative sampling of the entire thrombus area. Specifically, 5 to 10 non-overlapping high-power fields (200× or 400× magnification) were randomly captured per section, carefully avoiding areas with tissue folds or staining artifacts.

Specific pathological features and cellular subsets were quantified as follows: Collagen deposition and fibrosis area were quantified from Masson’s trichrome and Sirius Red-stained sections using ImageJ software (version 1.53, NIH). The percentage of the fibrotic area was calculated as the ratio of the collagen-positive area (blue for Masson; red for Sirius Red) to the total thrombus area across the randomly selected high-power fields. For macrophage subtyping via immunofluorescence or immunohistochemistry, M1-type macrophages were identified by the expression of iNOS (inducible nitric oxide synthase), while M2-type macrophages were defined by the expression of Arg1 (Arginase-1) and CD206 (mannose receptor). For the human PEA specimens, a qualitative descriptive analysis was employed to assess the spatial architecture of the organized thrombus. Rather than performing digital area quantification, these clinical sections were meticulously observed to descriptively evaluate the spatial distribution, enrichment patterns, and in situ co-localization of macrophage subsets (particularly M2 macrophages) relative to the dense fibrotic core regions. All qualitative assessments and representative image selections were independently verified by two pathologists blinded to the clinical metadata to ensure descriptive objectivity.

To ensure analytical objectivity and rigorously prevent observer bias, all downstream outcome measurements, including histological quantifications, positive area percentages, and cell counts, were performed by two independent investigators who were strictly blinded to both the clinical metadata of human subjects and the specific treatment allocations of the in vivo murine models. Crucially, in cases of significant discrepancy between the two independent observers (defined as an inter-observer variation > 10%), a third senior investigator was consulted to evaluate the slides and reach a final consensus.

Statistical analysis

Continuous variables were tested for normality and expressed as mean ± standard deviation (SD) when normally distributed, or as median (interquartile range, IQR) when non-normally distributed. Categorical variables were presented as counts and percentages. For comparisons between two groups, either independent-sample t test (normal distribution) or Mann–Whitney U test (non-normal distribution) was applied. Multiple group comparisons were conducted using the Kruskal–Wallis test for non-parametric data. Categorical variables were analyzed using the χ² test or Fisher’s exact test as appropriate. Correlations were assessed using Pearson’s correlation coefficient for normally distributed data or Spearman’s rank correlation coefficient for non-parametric data.

To evaluate the directional associations among variables, mediation analysis was performed. The significance of the mediation effect was fundamentally determined using a bootstrapping approach with 1,000 resamples to calculate the 95% confidence intervals (CI) for the indirect effect; an indirect effect was considered statistically significant if its 95% CI did not cross zero. A two-sided p value < 0.05 was considered statistically significant. Statistical analyses and graphing were performed using GraphPad Prism 9.5, SPSS 26.0, and Origin 2024 software.

Result

Elevated CD16⁺ monocyte proportions and MCP-1 levels are associated with hemodynamic dysfunction in CTEPH patients

To validate our previous single-cell transcriptomic findings indicating expansion of CD16⁺ monocyte subsets in CTEPH, we conducted a multicenter case-control study. Thirty-two CTEPH patients confirmed by right heart catheterization and 20 age- and sex-matched healthy controls were enrolled. Baseline characteristics revealed that CTEPH patients had significantly lower lymphocyte counts, hemoglobin, platelet levels, and serum albumin compared to controls (all p < 0.05), while total leukocyte count, neutrophil percentage, and total monocyte count showed no significant differences. Notably, serum NT-proBNP was significantly elevated in the CTEPH group, consistent with increased right ventricular strain. Hemodynamic assessment confirmed that patients exhibited marked pulmonary hypertension and increased vascular resistance (mean pulmonary artery pressure [mPAP]: 54.91 ± 19.54 mmHg; pulmonary vascular resistance [PVR]: 6.48 ± 4.54 Wood units), with most patients (84.4%) classified as NYHA functional class II (Supplementary Table 1).

Flow cytometric analysis revealed a substantial increase in CD16⁺ monocyte subset proportion among total monocytes in CTEPH patients (29.44% ± 4.27%) compared to healthy controls (10.65% ± 2.38%, p < 0.0001), despite no difference in total monocyte counts, indicating a shift in monocyte subset distribution rather than absolute monocytosis (Fig. 1A-B).

Fig. 1.

Expansion of circulating CD16⁺ monocytes and their association with MCP-1 signaling in CTEPH. (A) Flow cytometry gating strategy for monocyte subset identification. (i) Single cells were gated on FSC-H versus FSC-A to exclude doublets. (ii) Live cells were identified as 7-AAD–negative on SSC-A versus 7-AAD plots. (iii) Monocytes were gated on SSC-A versus CD14, as CD14⁺ events. (iv) CD16⁺ monocytes were further defined within the monocyte gate on SSC-A versus CD16 expression. (B) Quantitative comparison of circulating CD16⁺ monocyte proportions between CTEPH patients and healthy controls. (C) Plasma MCP-1 concentrations across healthy controls, acute pulmonary embolism (APE), and CTEPH groups, measured by ELISA. (D) Comparison of MCP-1 levels in peripheral blood (PB) and pulmonary arterial (PA) samples from CTEPH patients. (E) Correlation of CD16⁺ monocyte proportions and plasma MCP-1 levels with hemodynamic parameters assessed by right heart catheterization in CTEPH patients. (F) In vitro induction assay demonstrating MCP-1–mediated regulation of CD16⁺ monocytes via CCR2. Representative gating strategy: (i) single cells selected by FSC-A versus FSC-H; (ii) live cells gated as 7-AAD–negative; (iii) monocytes gated by SSC-A versus CD14; (iv) CD16⁺ monocyte subset quantified by SSC-A versus CD16. (G) Bar graph summarizing the effects of MCP-1 stimulation and CCR2 antagonist RS102895 on the proportion of CD16⁺ monocytes in PBMCs from healthy donors

To investigate upstream inflammatory mediators, we quantified plasma MCP-1 levels across three groups. MCP-1 was highest in APE patients, intermediate in CTEPH, and lowest in controls (all intergroup p < 0.05) (Fig. 1C). Importantly, CTEPH patients also exhibited significantly higher MCP-1 levels in pulmonary artery blood (obtained via right heart catheterization) compared to peripheral venous blood (p < 0.05), suggesting site-specific enrichment of MCP-1 in the pulmonary vascular bed (Fig. 1D).

Spearman correlation analysis demonstrated a strong positive correlation between the proportion of CD16⁺ monocytes and both PVR (r = 0.86, p < 0.0001) and mPAP (r = 0.59, p < 0.001). Moreover, peripheral MCP-1 levels correlated with CD16⁺ monocyte proportion (r = 0.36, p < 0.05) and PVR (r = 0.45, p < 0.05), while no significant correlation was found between pulmonary artery MCP-1 levels and hemodynamic parameters (Fig. 1E).

To further elucidate the mechanistic interplay between MCP-1 and CD16⁺ monocytes, we performed mediation analysis. MCP-1 levels showed a near-significant association with CD16⁺ monocyte proportions (a = 0.036, p = 0.064). CD16⁺ monocytes were strongly associated with both PVR (b = 0.827, p < 0.001) and mPAP (b = 2.858, p < 0.001). Indirect effects of MCP-1 on PVR and mPAP through CD16⁺ monocytes were significant (PVR: ab = 0.030, 95% CI [0.004, 0.073]; mPAP: ab = 0.102, 95% CI [0.017, 0.265]), whereas direct effects were nonsignificant. In contrast, when CD16⁺ monocyte proportion was treated as the independent variable and MCP-1 as mediator, no significant indirect effects were observed, suggesting MCP-1 may function upstream of monocyte subset reprogramming (Supplementary Table 2).

To validate this hypothesis, we performed in vitro stimulation of healthy PBMCs with recombinant MCP-1. Initial dose-response experiments (0–200 ng/mL) clearly demonstrated that 100 ng/mL induced the most robust expansion of CD16⁺ monocytes without compromising cell viability (Supplementary Fig. 1). Consequently, cells in the main experimental group were stimulated with this optimal concentration of 100 ng/mL for 24 h (Fig. 1F). MCP-1 significantly increased the proportion of CD16⁺ monocytes compared to vehicle control. This effect was attenuated by co-treatment with the CCR2 antagonist RS102895, while RS102895 alone had no effect (Fig. 1G). These results suggest MCP-1-mediated CD16⁺ monocyte expansion is dependent on CCR2 signaling.

Collectively, these findings reveal that MCP-1 is enriched in the pulmonary vasculature and may act via CCR2 to promote CD16⁺ monocyte expansion, which in turn is closely associated with elevated pulmonary vascular resistance and pressure in CTEPH patients.

Single-cell transcriptomic landscape reveals the cellular heterogeneity and monocyte differentiation trajectory within acute pulmonary arterial thrombi

To elucidate the early immunopathological features of APE and the involvement of monocyte subsets in thrombus formation and progression, we performed single-cell RNA sequencing (scRNA-seq) on pulmonary arterial thrombi freshly obtained from five patients undergoing catheter-directed thrombectomy. The baseline clinical characteristics of these patients are summarized in Supplementary Table 3, while representative imaging findings from CT pulmonary angiography and digital subtraction angiography are shown in Supplementary Fig. 2. Macroscopic appearances of freshly aspirated thrombi are presented in Supplementary Fig. 3. The experimental workflow is illustrated in Fig. 2A, encompassing thrombus collection, tissue processing, library construction, sequencing, and bioinformatic analysis. In total, 24,399 high-quality single-cell transcriptomes were obtained across the five samples, with detailed quality control metrics summarized in Supplementary Table 4.

Fig. 2.

Single-cell transcriptomic landscape of acute pulmonary arterial thrombi and functional characterization of monocyte subsets. (A) Schematic workflow illustrating the experimental pipeline, from thrombus collection during catheter-directed thrombectomy to tissue processing, library preparation, sequencing, and bioinformatic analysis. (B) UMAP visualization of single-cell transcriptomes from thrombus tissues, showing unsupervised clustering. (C) Projection of five major immune cell types onto the UMAP plot. (D) Pie chart showing the relative proportions of the five major immune cell populations within the total single-cell dataset. (E) Bar plots depicting the composition of the five immune cell populations across thrombus samples from five patients with acute pulmonary embolism (APE). (F) UMAP projection of the monocyte/macrophage compartment after subclustering. (G) Annotation of monocyte/macrophage subsets, showing three populations: CD14⁺ monocytes, CD16⁺ monocytes, and macrophages. (H) Pie chart showing the relative distribution of the three monocyte/macrophage subsets within the total monocyte/macrophage compartment. (I) Bar plots showing patient-level distribution of CD14⁺ monocytes, CD16⁺ monocytes, and macrophages across the five APE thrombus samples. (J) GO enrichment analysis of genes significantly upregulated in CD16⁺ monocytes compared with CD14⁺ monocytes. (K) KEGG enrichment analysis of genes significantly upregulated in CD16⁺ monocytes compared with CD14⁺ monocytes. (L) GO enrichment analysis of genes significantly downregulated in CD16⁺ monocytes compared with CD14⁺ monocytes. (M) KEGG enrichment analysis of genes significantly downregulated in CD16⁺ monocytes compared with CD14⁺ monocytes. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes. APE, acute pulmonary embolism

Unsupervised clustering and dimensionality reduction via UMAP revealed 19 transcriptionally distinct clusters, which were annotated into five major immune and vascular cell types based on canonical marker gene expression: neutrophils, monocytes/macrophages (MPs), NK/T cells, B cells, and endothelial cells (Fig. 2B–C). Feature plots confirmed the expression specificity of representative markers (e.g., CD14, FCGR3A, CD68, CD3D, CD79A, PECAM1), validating the robustness of cell-type annotation (Supplementary Fig. 4A-B). Quantitative cell proportion analysis indicated that neutrophils (48.88%) and monocytes/macrophages (48.67%) overwhelmingly dominated the thrombus immune microenvironment, whereas adaptive immune cells and endothelial cells were sparse (< 3%), consistent across all individual samples (Fig. 2D–E).

Focusing on mononuclear phagocytes, further reclustering identified 15 subclusters within the monocyte/macrophage compartment, which were aggregated into three major subsets: CD14⁺ classical monocytes, CD16⁺ non-classical monocytes, and macrophages (Fig. 2F–G). Expression patterns of CD14, FCGR3A, CD68, and CD163 supported clear separation among subsets (Supplementary Fig. 5A-B). Among these, CD16⁺ monocytes constituted the dominant population (63.93%), followed by macrophages (20.84%) and CD14⁺ monocytes (15.23%), with highly consistent distribution across individual samples (Fig. 2H–I), suggesting that CD16⁺ monocytes are a central player in the early APE thrombus niche.

To further investigate functional differences between CD16⁺ and CD14⁺ monocytes in acute pulmonary arterial thrombi, we performed GO and KEGG enrichment analyses on genes differentially up- or downregulated between the two subsets. Upregulated genes in CD16⁺ monocytes were enriched in pathways related to glycolysis/gluconeogenesis, chemotaxis, hypoxia response, extracellular matrix organization, and cell-substrate junction assembly (Fig. 2J-K), indicating metabolic reprogramming, enhanced recruitment to hypoxic thrombus regions, and potential involvement in thrombus structural remodeling. Conversely, downregulated genes were significantly enriched in immune-related processes, including leukocyte cell-cell adhesion, regulation of T cell activation, leukocyte-mediated immunity, and antigen processing and presentation (Fig. 2L). KEGG analysis similarly highlighted suppression of cell adhesion molecules, chemokine signaling, antigen presentation, and NK cell-mediated cytotoxicity pathways (Fig. 2M), suggesting that CD16⁺ monocytes in APE exhibit reduced intercellular interactions, impaired T cell activation, and attenuated immune effector functions relative to CD14⁺ monocytes.

To investigate intercellular communication within pulmonary arterial thrombi, we applied CellChat analysis to construct cell-cell interaction networks. Significant interactions were observed among neutrophils, monocytes, and macrophages, with CD16⁺ monocytes exhibiting markedly stronger communication with other cell types compared to CD14⁺ monocytes (Fig. 3A). Given the established role of MCP-1 in CTEPH and its membership in the CCL chemokine family, we further focused on CCL signaling. The analysis revealed enhanced CCL-mediated communication among CD14⁺ monocytes, CD16⁺ monocytes, and macrophages (Fig. 3B). Within this pathway, CD16⁺ monocytes predominantly functioned as signal senders, whereas macrophages acted mainly as signal receivers and effectors (Fig. 3C). Moreover, CCL signaling between these cell types was largely mediated through autocrine and paracrine interactions (Fig. 3D). Collectively, these findings highlight the central role of CD16⁺ monocytes in coordinating CCL-dependent intercellular signaling within pulmonary arterial thrombi and underscore their close communication with macrophages.

Fig. 3.

Cell–cell communication and differentiation dynamics of monocyte/macrophage subsets in acute pulmonary thrombi. (A) Overview of cell–cell communication networks inferred by CellChat. Left: global interaction network among all immune cell populations; middle: network centered on CD16⁺ monocytes as signal senders; right: network centered on CD14⁺ monocytes as signal senders. Edge thickness reflects interaction strength. (B) Chord diagram illustrating intercellular interactions mediated by the CCL signaling pathway. Each arc represents a cell type, and ribbons denote signal transmission between source and target cells, with ribbon width proportional to interaction strength. (C) Heatmap summarizing the roles of different cell types in the CCL signaling pathway, including sender, receiver, mediator, and influencer. Darker colors indicate higher relative importance in each role. (D) Bubble plot depicting major CCL signaling interactions among CD14⁺ monocytes, CD16⁺ monocytes, and macrophages. Source and target cell pairs are indicated; bubble color intensity reflects interaction strength. (E) Pseudotime trajectory analysis of monocyte/macrophage subsets within acute thrombus tissue, visualized on a UMAP plot. Analyses were performed using CellChat and Slingshot algorithms. UMAP, Uniform Manifold Approximation and Projection

Finally, pseudotime trajectory inference reconstructed a continuous monocyte-to-macrophage differentiation path, originating from CD14⁺ monocytes, transitioning through CD16⁺ monocytes, and culminating in mature macrophages (Fig. 3E). This trajectory reveals a potential developmental cascade within the thrombus, where recruited CD14⁺ monocytes undergo phenotypic reprogramming toward CD16⁺ non-classical monocytes and ultimately differentiate into macrophages, aligning with their spatial and functional distribution.

Histopathological and immunophenotypic characterization of pulmonary endarterectomy tissues reveals progressive macrophage infiltration and M2 polarization along the chronic thrombus remodeling axis in CTEPH

In our preceding study, single-cell RNA sequencing of acute pulmonary arterial thrombi revealed a dominant presence of CD16⁺ monocytes and a clear pseudotime trajectory of monocyte differentiation from CD14⁺ monocytes through CD16⁺ monocytes toward terminal macrophages. This dynamic suggests that during the early phase of thrombus formation, monocytes undergo distinct phenotypic transitions, culminating in macrophages that may contribute to thrombus fibrosis and vascular remodeling. Based on this, and considering the macrophage-rich nature of PEA specimens from CTEPH patients, we conducted a systematic analysis of PEA-obtained tissues to investigate the localization and functional phenotypes of monocyte-derived macrophages in chronic thrombi. The pulmonary endarterectomy specimens obtained after surgery exhibited grossly intact structural integrity (Fig. 4A), while histological staining with H&E and Masson trichrome revealed distinct regional features corresponding to different stages of thrombus organization, including fresh thrombus, organized thrombus, fibroblast-enriched zones, and end-stage fibrotic tissue. The fresh thrombus regions were characterized by loosely arranged blood cells and fibrin meshwork, indicative of early thrombus formation. Organized thrombus regions showed signs of initial tissue remodeling, while fibroblast-rich areas exhibited increased fibrin and collagen deposition. The end-stage fibrotic regions were dominated by dense connective tissue, reflecting advanced fibrosis (Fig. 4B). To assess macrophage distribution across these stages, we performed CD68 immunohistochemical staining. Results showed minimal macrophage presence in fresh thrombi, whereas organized thrombi, fibroblast-enriched zones, and fibrotic regions exhibited substantial CD68⁺ macrophage infiltration in a progressively increasing pattern, suggesting a potential role for macrophages in thrombus maturation and tissue remodeling (Fig. 4C). To explore the chemotactic signals involved, we further stained for CCL2 (MCP-1), a key monocyte-recruiting chemokine, and observed strong spatial colocalization with areas of macrophage infiltration, implicating MCP-1 in local monocyte/macrophage recruitment and chronic inflammation (Fig. 4D). In our prior single-cell analysis, macrophages within CTEPH tissues—originating from peripheral CD16⁺ monocytes—displayed transcriptomic features consistent with an M2 phenotype. To validate this at the protein level, we performed IHC staining for iNOS (M1 marker) and M2 macrophage markers. The results demonstrated that macrophages infiltrating the organized and fibroblast-rich regions were predominantly M2-type (Fig. 4D), suggesting a profibrotic macrophage polarization pattern potentially contributing to extracellular matrix deposition and chronic thrombus remodeling in CTEPH. Together, these findings reveal a spatially and temporally regulated infiltration of monocyte-derived M2 macrophages in chronic thrombotic tissue, highlighting their likely pathological role in promoting fibrosis and vascular obstruction in CTEPH.

Fig. 4.

Histopathological and immunohistochemical characterization of pulmonary endarterectomy (PEA) specimens from CTEPH patients. (A) Gross morphology of PEA tissue obtained from CTEPH patients (specimen provided by China-Japan Friendship Hospital). The specimen shows intact structural integrity of the excised pulmonary arterial intima. (B) Histological features of CTEPH thrombus tissue. Main panel: Masson’s trichrome staining illustrating structural heterogeneity within the PEA specimen. Insets: representative high-magnification views of distinct regions, including fresh thrombus, organized thrombus, fibrotic area, and fibroblast-rich area. (C) Spatial distribution of macrophages and MCP-1 (CCL2) in CTEPH thrombus tissue. Representative sections from fresh thrombus, organized thrombus, fibroblast-rich regions, and fibrotic areas stained with H&E, Masson’s trichrome, CD68 immunohistochemistry (macrophages), and CCL2 immunohistochemistry. (D) Polarization of macrophages in CTEPH thrombus tissue. Representative sections from organized thrombus and fibrotic regions stained with H&E, Masson’s trichrome, CD68 (pan-macrophage marker), iNOS (M1 macrophage marker), and ARG1 (M2 macrophage marker), highlighting regional enrichment of M2 macrophages in fibrotic areas

Inhibition of the MCP-1/CD16⁺ monocyte axis attenuates thrombus fibrosis through reduction of M2 macrophage accumulation in a murine IVC ligation model

To recapitulate the dynamic evolution of thrombus fibrosis observed in CTEPH, we employed a widely accepted murine IVC ligation model. This model induces venous stasis and endothelial injury, leading to the formation of inflammatory-rich thrombi that progressively remodel into fibrotic structures—closely mirroring the pathophysiological trajectory in human CTEPH. Based on our prior findings in patients, which revealed elevated circulating CD16⁺ monocytes and their positive correlation with plasma MCP-1 levels, we hypothesized that the MCP-1/CD16⁺ monocyte axis plays a key role in promoting thrombus fibrosis. Thus, we leveraged the IVC ligation model to characterize the spatial-temporal progression of thrombus remodeling and dissect the immunological underpinnings mediated by this axis.

Histological analyses of thrombi harvested at serial time points (Days 1, 3, 5, 7, and 14 post-ligation) using H&E, Masson’s trichrome, and Sirius Red staining revealed a continuum of thrombus organization. In early phases (Days 1–3), thrombi were predominantly composed of red blood cells and fibrin meshwork, accompanied by robust neutrophil infiltration indicative of acute inflammation. Over time, neutrophils waned while monocyte/macrophage infiltration increased, with evident collagen deposition and fibroblast-like cells appearing by Day 7. By Day 14, thrombi exhibited dense collagen fibers with layered concentric organization, consistent with advanced fibrosis (Fig. 5A). Quantitative analysis confirmed a sharp increase in fibrotic area percentage between Days 7 and 14 (Fig. 5B). Meanwhile, measurements of thrombus dimensions showed a steady decline in length, while width and weight followed a biphasic trend—initial reduction (Days 1–3), transient rebound (Days 3–5), and subsequent decrease (Days 7–14)—reflecting complex tissue remodeling and resorption dynamics (Supplementary Fig. 6A-C).

Fig. 5.

Dynamic changes of Ly6C⁻ monocytes and fibrosis in the IVC ligation mouse model and effects of CCR2 blockade. (A) Histological features of thrombus tissue in the IVC ligation model at Days 1, 3, 5, 7, and 14 after ligation. Representative H&E, Masson’s trichrome, and Sirius Red staining are shown. Scale bar: 250 μm. (B) Quantitative analysis of thrombus fibrosis in IVC-ligated mice. Masson’s staining–based measurements demonstrate progressive increases in fibrotic area percentage over time. (C) Immunofluorescence staining of thrombus tissue at Days 3, 7, and 14, showing accumulation of F4/80⁺CD206⁺ M2 macrophages. Nuclei are stained with DAPI (blue), macrophages with F4/80 (green), and M2 macrophages with CD206 (red). (D) Flow cytometry gating strategy for Ly6C⁻ monocytes in mouse peripheral blood. Monocytes were first gated on FSC-A versus SSC-A, doublets excluded on FSC-A versus FSC-H, and dead cells excluded using 7-AAD. Live leukocytes were gated as CD45⁺, monocytes as CD11b⁺, and Ly6C expression was used to define Ly6C⁻ monocytes. (E) Time-course line graph showing peripheral Ly6C⁻ monocyte proportions in sham and IVC-ligated mice before and at multiple time points (Days 1, 3, 5, 7, 10, 12, and 14). (F) Bar plots illustrating group differences in Ly6C⁻ monocyte proportions at each time point. (G) Time-course line graph comparing Ly6C⁻ monocyte dynamics between sham groups (sham + saline; sham + RS102895) and IVC-ligated groups (IVC + saline; IVC + RS102895). (H) Bar plots summarizing group-specific differences in Ly6C⁻ monocyte proportions at each time point. (I) Histological assessment of thrombus fibrosis at Day 14 in IVC + saline versus IVC + RS102895 groups. Representative H&E, Masson’s trichrome, and Sirius Red staining images are shown. Scale bar: 250 μm. (J) Quantification of fibrotic area percentage at Day 14, showing reduced fibrosis in RS102895-treated mice compared with saline controls. (K) Immunofluorescence staining of thrombus tissue at Day 14 demonstrating effects of RS102895 on M2 macrophage accumulation. Columns represent treatment groups (IVC + saline vs. IVC + RS102895). Rows show DAPI (blue, nuclei), F4/80 (purple, macrophages), CD206 (red, M2 macrophages), and merged images

Given the observed enrichment of M2 macrophages in fibrotic thrombus regions in CTEPH patients, we examined whether a similar trend occurs in the mouse model. Immunofluorescence staining for F4/80 and CD206 revealed a progressive accumulation of F4/80⁺CD206⁺ M2 macrophages from Day 3 to Day 14, which closely correlated with collagen deposition, supporting a functional role for M2 macrophages in driving thrombus fibrosis (Fig. 5C).

To further investigate the role of CD16⁺ monocyte equivalents in mice (Ly6C⁻ monocytes), we conducted flow cytometric analyses of peripheral blood Ly6C⁻ monocyte proportions across time points post-ligation. While both sham and IVC groups exhibited a transient early drop (Day 1–3), only the IVC group demonstrated a sustained and significant rebound beginning at Day 5 and persisting through Day 14 (Fig. 5D-F). This mirrored the elevated CD16⁺ monocyte levels seen in CTEPH patients and suggested a pathogenic role for this monocyte subset in chronic thrombus remodeling.

RS102895 is a small-molecule chemical inhibitor that interacts with critical binding sites on the CCR2 receptor, thereby competitively blocking the interaction between MCP-1 and CCR2 [23]. In our previous work, we demonstrated that RS102895 suppresses MCP-1–induced expansion of the CD16⁺ monocyte subset in human peripheral blood PBMCs. Building on these findings, we administered RS102895 via intermittent intraperitoneal injection in the IVC ligation mouse model to further assess its impact on circulating Ly6C⁻ monocytes. In RS102895-treated mice, the post-ligation rise in Ly6C⁻ monocytes was markedly suppressed relative to vehicle controls, particularly at Days 5, 7, and 14 (Fig. 5G-H). Importantly, Masson and Sirius Red staining revealed significantly reduced thrombus fibrosis in RS102895-treated animals at Day 14 (Fig. 5I-J). Consistently, immunofluorescence demonstrated diminished F4/80⁺CD206⁺ M2 macrophage accumulation in RS102895-treated thrombi (Fig. 5K).

Together, these results reveal that MCP-1–driven recruitment of Ly6C⁻ monocytes and their differentiation into M2 macrophages is a key contributor to thrombus fibrosis. Pharmacological blockade of this axis effectively mitigates M2 macrophage accumulation and attenuates fibrotic remodeling, highlighting a potential therapeutic avenue for preventing chronic thromboembolic vascular remodeling in CTEPH.

Discussion

Previous studies have established a close association between pulmonary hypertension (PH) and infiltration of myeloid cells into the pulmonary vasculature [24–26]. In hypoxia-induced PH, nonclassical monocytes were shown to be recruited into small pulmonary arteries and subsequently differentiate into interstitial macrophages, thereby driving vascular remodeling [16, 27]. Similarly, in high-altitude PH, circulating CD16⁺ monocytes (including both intermediate and nonclassical subsets) were markedly elevated compared with healthy controls, suggesting that specific monocyte subsets may represent pivotal regulators of hypoxia-induced vascular remodeling [16]. However, in CTEPH, a distinct subtype of PH, the phenotypic features and mechanistic roles of monocyte subsets remain poorly characterized. In our earlier single-cell transcriptomic analysis, we identified alterations in circulating CD16⁺ monocyte populations in CTEPH patients. To validate these observations, the present multicenter study employed flow cytometry to quantify monocyte subsets, confirming a selective expansion of circulating CD16⁺ monocytes in CTEPH. Notably, this expansion was not attributable to an overall increase in total monocyte counts, indicating preferential enrichment of this specific subset. Importantly, the proportion of CD16⁺ monocytes correlated positively with core hemodynamic parameters—mean pulmonary arterial pressure and pulmonary vascular resistance—implicating CD16⁺ monocytes in the pathogenic progression of CTEPH, potentially via sustained inflammatory activation and maladaptive vascular remodeling.

In recent years, the concept of thrombosis has evolved from a classical hemostatic cascade abnormality toward a broader paradigm of innate immune-driven biology, encapsulated in the notion of “immunothrombosis” [28, 29]. Inflammation and thrombosis are now recognized as bidirectionally reinforcing processes: inflammatory microenvironments promote thrombus formation, whereas developing thrombi further amplify inflammation, thereby establishing a self-perpetuating loop. Neutrophils are the earliest immune responders recruited to sites of thrombus formation, yet as thrombogenesis progresses, neutrophils are progressively replaced by monocytes, which emerge as the dominant immune cell population within thrombi [30]. Importantly, most thrombus-infiltrating monocytes are not derived from resident tissue pools but rather from circulating populations [31], underscoring the central role of blood-borne monocytes in both thrombosis and vascular immune regulation.

Recruitment of monocytes from the circulation to thrombus sites is a tightly regulated process mediated by activated endothelial cells secreting chemokines and monocyte expression of corresponding receptors [32]. Among these pathways, the CCL2/CCR2 axis is regarded as a principal regulator of monocyte trafficking. CCL2 (MCP-1) binds CCR2 on monocytes, promoting their directed chemotaxis [33]. Our previous work demonstrated that circulating MCP-1 levels are significantly elevated in CTEPH and correlate with adverse hemodynamics [34]. In this study, comparative analyses revealed that MCP-1 levels were increased in both CTEPH and APE relative to healthy controls, with the highest levels observed in APE. This gradient suggests that acute thrombosis is accompanied by intense inflammatory activation, whereas persistent MCP-1 elevation in CTEPH reflects ongoing chronic inflammation. MCP-1 levels also correlated positively with pulmonary vascular resistance, consistent with findings by Kimura et al. [11]. Furthermore, Kimura and colleagues demonstrated that MCP-1 levels decreased following pulmonary endarterectomy. Interestingly, in our present study, comparative analyses confirmed significantly higher MCP-1 expression within the pulmonary arteries compared to peripheral venous blood. However, despite this localized pulmonary arterial enrichment, an notable compartment-specific divergence emerged regarding clinical correlations. A robust linear correlation with macroscopic hemodynamic derangements, such as PVR, was uniquely observed with peripheral venous MCP-1 rather than the pulmonary arterial pool. We propose two primary pathophysiological mechanisms to explain this divergence. First, from a statistical perspective, the profoundly elevated and biologically saturated MCP-1 concentrations within the localized pulmonary arterial microenvironment likely create a physiological ‘ceiling effect’; this inherent data clustering restricts the statistical variance necessary to establish a linear correlation with hemodynamics. Second, from a dynamic metabolic standpoint, the intensely high levels of localized pulmonary artery MCP-1 are presumably subjected to rapid receptor binding, target consumption, and degradation by the massive influx of infiltrating immune cells directly within the thrombus lesion. Consequently, the localized pulmonary pool reflects a state of rapid consumption and active tissue turnover, whereas circulating peripheral venous MCP-1 serves as a more stable, unconsumed systemic barometer of the overall active monocyte recruitment cascade. This systemic reflection of generalized immunothrombotic activity ultimately provides a much stronger and more accurate linear correlation with global hemodynamic severity in CTEPH.

We further observed a strong positive correlation between circulating CD16⁺ monocytes and MCP-1 levels in CTEPH. In vitro, MCP-1 stimulation significantly increased CD16⁺ monocyte proportions in PBMCs from healthy donors, whereas this effect was abrogated by CCR2 antagonism, confirming the central regulatory role of the MCP-1/CCR2 axis. Together, these findings support a model wherein MCP-1 not only recruits circulating monocytes to thrombi but also drives their phenotypic shift toward the CD16⁺ subset, thereby amplifying inflammatory signaling and vascular remodeling in CTEPH.

Our integrative analyses highlight CD16⁺ monocytes as critical effector cells in orchestrating vascular remodeling, inflammatory cascades, and thrombus organization in CTEPH. Although PEA tissue studies have primarily emphasized macrophages rather than monocytes, this may reflect a continuum wherein recruited monocytes differentiate into macrophages, obscuring their earlier contributions. To resolve this, we constructed the first single-cell transcriptomic atlas of acute pulmonary thrombus tissue, revealing that neutrophils and monocytes together comprised over 95% of immune infiltrates, with CD16⁺ monocytes as the predominant monocyte subset. Pseudotime analysis delineated a trajectory from CD14⁺ monocytes to CD16⁺ monocytes and ultimately to macrophages. These findings suggest that CD16⁺ monocytes participate early in thrombus formation and subsequently differentiate into macrophages, thereby driving chronic fibrotic remodeling. This work thus expands the current framework of CTEPH pathogenesis by providing high-resolution insight into the immunothrombotic continuum linking acute thrombus formation with chronic vascular occlusion.

Histopathologically, organized thrombi in CTEPH induce intimal remodeling through dysregulated fibroblast/myofibroblast and endothelial proliferation [35]. However, the initiating steps of thrombus organization remain incompletely defined. Our single-cell data demonstrated progressive differentiation of CD16⁺ monocytes into macrophages within acute thrombi. Complementary histological analysis of PEA specimens revealed enrichment of M2 macrophages within regions of organized thrombus and fibroblast aggregation, coinciding with high MCP-1 expression. M2 macrophages, while classically involved in resolution and tissue repair, can drive pathological fibrosis when excessively activated, through secretion of profibrotic mediators including PDGF, CCL18, and TGF-β [36–39]. Consistent with prior evidence that murine Ly6C⁻ monocytes (human CD16⁺ equivalents) serve as precursors for M2 macrophages during myocardial infarction [40, 41], our findings support a model in which MCP-1 recruits circulating monocytes to thrombi, promotes their differentiation into CD16⁺ subsets, and drives subsequent polarization toward M2 macrophages, thereby exacerbating thrombus fibrosis and vascular occlusion.

Despite the integrative multi-level evidence provided in this study, several limitations should be acknowledged. First, the clinical component of this work is constrained by the relatively small sample size, which reflects not only the rarity of CTEPH but also its underdiagnosis in real-world settings. Epidemiological studies estimate that the cumulative incidence of CTEPH following acute pulmonary embolism ranges from approximately 0.4% to 6.2%, with a pooled estimate of around 3.4% [42], posing inherent challenges for large-scale patient recruitment. In addition, definitive diagnosis of CTEPH relies on invasive hemodynamic assessment by right heart catheterization, which may limit patient acceptance and further contribute to underdiagnosis. Although patients were consecutively enrolled across three independent centers using standardized diagnostic criteria, minor heterogeneity in baseline characteristics remains unavoidable. More importantly, the cross-sectional design inherently limits causal inference. While strong associations were observed between circulating CD16⁺ monocytes, MCP-1 levels, and hemodynamic parameters, these relationships may also reflect reverse causality. For example, elevated pulmonary arterial pressure and shear stress could induce endothelial activation and MCP-1 secretion, thereby secondarily promoting monocyte recruitment in a compensatory or feedback-driven manner. In addition, due to the acute and emergent nature of APE, comprehensive immunophenotyping of monocyte subsets was not feasible, and only plasma MCP-1 levels were analyzed in this cohort. Future large-scale, prospective, and longitudinal studies—particularly those tracking patients following acute pulmonary embolism—will be essential to determine whether dynamic changes in MCP-1 or CD16⁺ monocytes can serve as predictive biomarkers for CTEPH development.

At the mechanistic level, several limitations remain. Although our in vitro experiments demonstrate that MCP-1 promotes the expansion of CD16⁺ monocytes via CCR2, the downstream signaling pathways governing CD16 expression, functional polarization, and profibrotic activity remain undefined. Moreover, direct functional studies using isolated primary human CD16⁺ monocytes were not performed, largely due to their low abundance in peripheral blood and their rapid loss of viability following ex vivo isolation, which currently limits their applicability in sustained functional assays. It should also be emphasized that the immunothrombotic microenvironment is orchestrated by a complex cytokine network, and while the present study highlights the MCP-1/CCR2 axis as a dominant driver, other inflammatory mediators are likely to contribute synergistically to monocyte recruitment and macrophage polarization. In parallel, several findings in this study rely on computational inference, including cell–cell communication analysis and pseudotime trajectory modeling. These approaches, while informative, are inherently based on transcriptomic abundance and algorithmic assumptions and should therefore be interpreted as hypothesis-generating rather than definitive evidence. Future studies integrating high-resolution multi-omics approaches, lineage-tracing strategies, and functional validation will be required to fully delineate the molecular and cellular programs underlying CD16⁺ monocyte biology in CTEPH.

Animal models remain essential for mechanistic validation. Although several experimental models of CTEPH have been proposed, each has inherent limitations in fully recapitulating the complex human phenotype [12, 43–47]. In this context, we employed the murine IVC ligation model, which reliably induces thrombus formation and subsequent chronic fibrotic remodeling, thereby providing a robust and experimentally tractable platform to investigate the immunothrombotic mechanisms underlying thrombus organization [48–53]. In this model, we observed dynamic changes in Ly6C⁻ monocytes that paralleled the expansion of CD16⁺ monocytes in human CTEPH, and demonstrated that CCR2 inhibition attenuated monocyte accumulation, M2 macrophage infiltration, and thrombus fibrosis, collectively supporting a causal role for the MCP-1/CCR2/CD16⁺ monocyte axis in chronic thrombus remodeling. Nevertheless, these in vivo findings should be interpreted in the context of model-specific limitations. The IVC ligation model represents a systemic venous thrombosis model and does not fully recapitulate the unique high-shear pulmonary hemodynamics or distal arteriopathy characteristic of human CTEPH. Furthermore, interspecies differences between murine Ly6C⁻ monocytes and human CD16⁺ monocytes, particularly at the level of receptor expression and transcriptional regulation, may limit direct translational extrapolation. Although pharmacological inhibition of CCR2 provided functional support for this pathway, the potential for off-target effects cannot be completely excluded. Future studies employing genetic models, such as CCR2- or MCP-1-deficient mice, in combination with fate-mapping approaches, will be critical to establish more definitive causal relationships. Conceptually, it is also important to recognize that this study primarily addresses the macrophage-driven fibrotic phase of immunothrombosis, whereas earlier events—such as neutrophil extracellular trap formation and platelet activation during acute embolism—were not directly investigated. Moreover, thrombus organization represents only one dimension of CTEPH pathogenesis; complex interactions among macrophages, fibroblasts, endothelial cells, and vascular cells, including processes such as endothelial-to-mesenchymal transition, remain to be systematically explored.

From a translational perspective, while circulating CD16⁺ monocytes and MCP-1 show promise as biomarkers reflecting upstream immunothrombotic activity, the current cohort size limits their immediate clinical applicability. Future large-scale, multicenter studies will be required to validate these findings, and more importantly, prospective longitudinal cohorts following patients after acute pulmonary embolism will be essential to determine whether persistent elevation of CD16⁺ monocytes or MCP-1 can serve as reliable predictors of progression to CTEPH. In parallel, further mechanistic investigations are needed to dissect the downstream signaling pathways of the MCP-1/CCR2 axis and to better define its role in monocyte functional programming and fibrotic remodeling. Although therapeutic targeting of this pathway is conceptually attractive, its central role in innate immune surveillance raises important concerns regarding systemic immunologic side effects. Long-term systemic inhibition of MCP-1/CCR2 signaling may therefore carry risks of impaired host defense. Accordingly, future therapeutic strategies should prioritize approaches that achieve spatial or tissue-specific modulation, such as localized pulmonary delivery, to maximize antifibrotic efficacy while minimizing systemic immune suppression.

In summary, this study identifies significant elevation of both MCP-1 and circulating CD16⁺ monocytes in CTEPH, and delineates their mechanistic interplay in promoting chronic thrombus fibrosis. These findings expand our understanding of the immunopathology of CTEPH and highlight the MCP-1/CD16⁺ monocyte axis as a potential therapeutic target. Although limited by sample size and cross-sectional design, our integrated clinical, experimental, and preclinical analyses provide compelling evidence that immune-driven thrombus remodeling underpins CTEPH progression. Future longitudinal and mechanistic studies will be crucial to validate these findings and to translate them into precision immunomodulatory therapies for CTEPH.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

M.C., X.L., F.L., and J.L. contributed equally to this work. M.C. and C.D. designed the study. M.C, Q.W., N.S., F.L., and J.L. analyzed and interpreted the data. M.C., X.L., F.L., J.L., collected clinical samples. M.C., Q.W., N.S., X.L., and F.L. conducted the experiments. M.C., Q.W., N.S., X.L., and C.D. wrote the manuscript. M.C. and C.D. edited the manuscript. M.C., X.L., F.L., J.L., and C.D. final reviewed the manuscript. All authors read and approved the final manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki for human research and the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals for animal experiments. The clinical protocol was approved by the Ethics Committees of Wuhan No. 6 Hospital (No. WHSHIRB-K-2020005), the First Affiliated Hospital of Fujian Medical University (No. [2023]401), and China-Japan Friendship Hospital (No. 2022-KY-085), and informed consent was obtained from all individual participants. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Fujian Medical University (Approval No. IACUC FJMU 2022 − 0442).

Conflict of interest

The authors declare no conflict of interest.