Targeted Delivery of GLUT1 Inhibitor via Macrophage Nanovesicles for Pulmonary Arterial Hypertension Therapy

ABSTRACT

Pulmonary arterial hypertension (PAH) is a lethal disease characterized by inflammation‐driven vascular remodelling, with a lack of precise intervention strategies targeting pro‐inflammatory M1 macrophages' glycolytic metabolic reprogramming. This study introduced Mac@WZB117, a macrophage membrane‐camouflaged nanovesicle delivering the GLUT1 inhibitor WZB117, specifically targeting M1 macrophages for PAH treatment. In mouse models of PAH induced by MCT and SU5416 combined with hypoxia, Mac@WZB117 effectively targeted pulmonary vascular lesions, leading to a significant reversal of arterial thickening and collagen deposition. Molecular analyses indicated a downregulation of GLUT1 expression in M1 macrophages, along with a polarization shift from M1 to M2. Furthermore, intervention with Mac@WZB117 reshaped the immuno‐metabolic lineage by inhibiting glycolysis, lipid metabolism and TGF‐β signalling pathways. Spatial transcriptomics analysis confirmed the spatial reorganization of hyperactive M1 macrophages from lesion regions. Overall, Mac@WZB117 delivers WZB117 precisely to M1 macrophages, suppressing their pro‐inflammatory and metabolic functions to remodel the immune microenvironment and reverse vascular structural changes, suggesting a promising therapeutic approach for PAH and other immuno‐metabolic diseases.

Mac@WZB117 to target M1 macrophage glycolysis, remodel the immune microenvironment, and attenuate vascular remodelling in PAH.

Abbreviations

2‐DG

2‐deoxy‐D‐glucose

α‐SMA

α‐smooth muscle actin

ALT

alanine aminotransferase

ANOVA

analysis of variance

AST

aspartate aminotransferase

BUN

blood urea nitrogen

CLSM

confocal laser scanning microscopy

Cr

creatinine

DEG

differentially expressed gene

DLS

dynamic light scattering

DL

drug loading

DMSO

dimethyl sulfoxide

ECAR

extracellular acidification rate

ECL

enhanced chemiluminescence

EE

encapsulation efficiency

FACS

fluorescence‐activated cell sorting

FBS

foetal bovine serum

FDR

false discovery rate

FMO

fluorescence‐minus‐one

GEO

gene expression omnibus

GSEA

gene set enrichment analysis

H&E

hematoxylin and eosin

HPASMCs

human pulmonary artery smooth muscle cells

IHC

immunohistochemistry

IF

immunofluorescence

iNOS

inducible nitric oxide synthase

Mac@WZB117

macrophage membrane‐camouflaged nanovesicles for glycolysis inhibition through the delivery of the GLUT1 inhibitor WZB117

MCT

monocrotaline

mPAP

mean pulmonary arterial pressure

PCA

principal component analysis

PAH

pulmonary arterial hypertension

PBS

phosphate‐buffered saline

PDI

polydispersity index

PLGA

poly(lactic‐co‐glycolic acid)

PMVECs

pulmonary microvascular endothelial cells

PVA

polyvinyl alcohol

PVR

pulmonary vascular resistance

qPCR

quantitative polymerase chain reaction

RES

reticuloendothelial system

ROS

reactive oxygen species

RVHI

right ventricular hypertrophy index

RVSP

right ventricular systolic pressure

scRNA‐seq

single‐cell RNA sequencing

SD

standard deviation

SM22α

transgelin

SPF

specific pathogen‐free

TEM

transmission electron microscopy

VIP

variable importance in projection

W/O

water‐in‐oil

W/O/W

water‐in‐oil‐in‐water

WT

wall thickness

WT%

wall thickness percentage

1. Introduction

Pulmonary arterial hypertension (PAH) remains a lethal cardiopulmonary disorder characterized by progressive pulmonary vascular remodelling and sustained elevation of vascular resistance, ultimately leading to right ventricular overload and heart failure (Thenappan et al. 2018; Li et al. 2024). Despite advances in pharmacotherapy, current strategies, predominantly focused on vasodilation and anticoagulation, fail to reverse established structural lesions or significantly improve long‐term survival rates, highlighting an urgent need for disease‐modifying interventions that target the root pathological mechanisms (Sharma et al. 2025; Bisserier et al. 2020; Jin et al. 2023). Emerging evidence implicates the immune‐vascular axis as a critical driver of PAH pathogenesis. Specifically, the infiltration and maladaptive polarization of macrophages towards the pro‐inflammatory M1 phenotype orchestrate vascular inflammation and remodelling through the secretion of cytokines and paracrine modulation of smooth muscle cell proliferation (Zhang et al. 2023; Fan et al. 2021). Consequently, correcting the aberrant polarization of M1 macrophages represents a promising therapeutic frontier for remodelling the pulmonary vascular microenvironment (Luo and Qiu 2022; Liu et al. 2023).

Recent advances in immunometabolism have unveiled that macrophage phenotype and function are intrinsically coupled to their metabolic state (Tabas and Bornfeldt 2016; Batista‐Gonzalez et al. 2020). M1 macrophage activation is heavily dependent on aerobic glycolysis (the Warburg effect) for rapid ATP generation and biosynthetic precursor supply, a process rate‐limited by the glucose transporter 1 (GLUT1) (Fan et al. 2022). Upregulation of GLUT1 is a hallmark of M1 polarization, and its inhibition has been shown to markedly suppress pro‐inflammatory responses and attenuate fibrosis in various disease models (O'Rourke et al. 2022; Liang et al. 2024). While the small‐molecule GLUT1 inhibitor WZB117 demonstrates potent glycolytic suppression, its clinical translation for PAH is hindered by poor pulmonary selectivity and potential systemic toxicity arising from off‐target effects on glucose‐dependent organs (Ojelabi et al. 2016; De et al. 2023).

To overcome these delivery barriers and achieve precision immunometabolic modulation, biomimetic nanotechnology offers a compelling solution. By leveraging the natural homing capabilities of immune cell membranes, drug delivery systems can be engineered to specifically target inflamed lesions. In this study, we developed a macrophage membrane‐camouflaged nanovesicle system loaded with WZB117 (Mac@WZB117). This platform aims to precisely target glycolysis‐dependent M1 macrophages within the PAH microenvironment, block their glycolytic flux via GLUT1 inhibition, and induce phenotypic reprogramming from M1 to M2. This strategy provides a novel therapeutic paradigm to reverse pulmonary vascular remodelling and ameliorate PAH progression.

2. Materials and Methods

2.1. Construction and Grouping of Animal Models

The animal experiments were conducted in a specific pathogen‐free (SPF) barrier environment, utilizing 8‐week‐old male C57BL/6J mice (weighing 20–22 g), procured from Beijing Vital River Laboratory Animal Technology Co. Ltd (Beijing, China). Following a 7‐day acclimatization period, the mice were randomly assigned to groups using a random number table to ensure balanced grouping. Five experimental groups were established (15 mice per group): (1) Sham group: Mice were maintained under normoxic conditions (21% O₂) with subcutaneous injection of an equivalent volume of normal saline. (2) PAH + NS group: Mice were subcutaneously injected with SU5416 (20 mg/kg, Sigma) once weekly for 3 weeks and exposed to a hypoxic chamber (10% O₂, 90% N₂) continuously for 21 days. (3) PAH+Free WZB117 group: Following PAH model induction, free WZB117 was administered via tail vein injection (10 mg/kg, dissolved in a 10% dimethyl sulfoxide [DMSO]/PBS), three times per week for two consecutive weeks starting on Day 7. The dosage of 10 mg/kg was determined based on a previous foundational pharmacodynamic study (Liu et al. 2012) and further validated by another study (Wu et al. 2020), which confirmed that this concentration effectively inhibits GLUT1‐mediated transport in vivo while maintaining a favourable safety profile. The intervention commenced on Day 7 post‐induction to specifically target the progressive phase of vascular remodelling and evaluate the therapeutic reversal potential, rather than prophylactic prevention. (4) PAH+Mac@WZB117 group: Mice were administered biomimetic nanovesicles containing the same dosage of WZB117 via tail vein injection, following the identical schedule as the Free WZB117 group (starting on Day 7, 3 times/week for 2 weeks). (5) PAH+Mac@WZB117+2‐deoxy‐D‐glucose (2‐DG) group: in addition to nanovesicle intervention, mice were administered with daily intraperitoneal injections of 2‐DG (500 mg/kg, Sigma). This dosage was selected according to established protocols for systemic glycolysis inhibition in murine models to ensure sufficient metabolic blockade. This group served as a mechanistic validation control to confirm that the therapeutic efficacy is indeed mediated through the inhibition of the glycolytic pathway. A detailed schematic of the experimental timeline and treatment regimens is provided in Figure S1. Hypoxia induction was achieved using the ProOx 110 system (BioSpherix), with automatic real‐time control and daily calibration of oxygen concentration. SU5416 was diluted to a final concentration of 5% DMSO and 0.5% Tween‐80 using sterile DMSO and polysorbate‐80. The animals' body weight, food intake and activity status were monitored daily, and mortality rates were recorded to assess modelling success and intervention tolerance. The successful establishment of the PAH model was rigorously validated prior to therapeutic intervention. As shown in Figure S2A,B, the PAH model group exhibited significantly elevated right ventricular systolic pressure (RVSP) and mean pulmonary arterial pressure (mPAP) compared to controls, alongside a survival rate consistent with severe disease progression. Furthermore, representative H&E staining (Figure S2C) confirmed characteristic vascular remodelling, including medial hypertrophy and luminal narrowing, validating the reliability of the model.

Upon completion of modelling, all mice were maintained under normoxic conditions until Day 28. All procedures during this period were conducted by two independent researchers in a blinded manner to eliminate observer bias. No antibiotics or glucocorticoids were used during the experiment to avoid interference with immune metabolic status. Animal experiment adhered to the International Guidelines for Animal Experimentation and ARRIVE 2.0 reporting standards. The study design, animal welfare considerations, and endpoint management received approval from the Animal Ethics Committee of Hainan Medical University (Ethics Approval No.: HYLL‐2024‐759).

At the end of the intervention period, right heart catheterization was performed on some mice to assess pulmonary artery pressure parameters and confirm model effects; other animals were utilized according to experimental objectives for lung tissue collection, serum biochemical analysis, micro‐CT imaging and histopathological studies. Lung tissue collection followed standardized operational sequences and anaesthesia methods (induction with inhaled isoflurane at 2.5%, maintenance at 1.5%). After tracheal perfusion with PBS, the left/right lung lobes were isolated for different experimental modules.

To control for individual variability affecting intervention outcomes, baseline body weight and initial pulmonary artery pressure values (n = 10 pre‐experiment) were incorporated as covariates in variance analysis modelling. Specimens and data from each group underwent unified coded management with the establishment of sample‐indicator bidirectional matching tables to ensure data consistency and traceability throughout analysis. All animal samples were stored under cold chain conditions (‐80°C) or paraffin‐embedded according to predefined experimental priority strategies to ensure parallelism in molecular, biochemical, pathological and imaging data (Figure S1).

2.2. Construction and Characterization of Mac@WZB117

In this study, drug‐loaded nanoparticles were prepared using a double emulsion solvent evaporation method (water‐in‐oil‐in‐water [W/O/W]). Specifically, 10 mg of GLUT1 inhibitor WZB117 and 50 mg of poly(lactic‐co‐glycolic acid) (PLGA) were dissolved in 2 mL of dichloromethane to form the oil phase. Under an ice bath, 0.5 mL of distilled water was added. The mixture was then emulsified for 30 s using a probe‐type sonicator (20 kHz, 30% amplitude) to form the primary water‐in‐oil (W/O) emulsion. The primary emulsion was rapidly poured into a 4 mL 2% polyvinyl alcohol (PVA) solution and sonicated again to form the W/O/W system. The organic solvent was removed by rotary evaporation under reduced pressure to obtain the preliminary PLGA nanovesicles. These vesicles were subsequently purified by centrifugation at 10,000 rpm and resuspended in PBS. To improve encapsulation efficiency (EE) and particle size uniformity, the suspension was passed through a 0.22 µm filter, yielding WZB117@PLGA nanoparticles.

To construct biomimetic nanovesicles, cell membranes were first extracted from RAW264.7 murine macrophages (ATCC TIB‐71). After collection, cells were pre‐centrifuged at 1000 g to remove dead cells and debris, followed by cold centrifugation at 10,000 g × 20 min to extract membrane components. The membrane precipitate was resuspended in Tris buffer and subjected to repeated freeze‐thaw cycles and sonication (at 4°C, power 20%, cycle of 10s/10s for six times) to form homogeneous membrane fragments. These were then mixed with PLGA nanovesicles at a protein/vesicle mass ratio of 1:10. Membrane fusion was achieved using a liquid extruder at 37°C to form Mac@WZB117 biomimetic nanovesicles. Successful coating was confirmed via Western blot analysis (Song et al. 2019) of surface membrane protein markers CD11b, F4/80 and CD86 to assess consistency with the original macrophage membrane protein enrichment. A detailed list of all key reagents, antibodies, cell lines and software used in this study, along with their sources and catalogue numbers, is provided in the Key Resources Table (Table S1) in the Supporting Information.

The nanovesicles were physicochemically characterized using multiple techniques. A Zetasizer Nano ZS 90 (Malvern) was used to measure their hydrated particle size, polydispersity index (PDI) and zeta potential to evaluate size stability and charge‐shielding effects. Transmission electron microscopy (TEM, JEOL JEM‐2100) with negative staining was employed to observe their morphology and structure. Ultraviolet‐visible spectroscopy (UV‐Vis, NanoDrop 2000) and Raman spectroscopy (Thermo DXR) were used to qualitatively verify drug loading (DL) and membrane components. The EE and DL of the nanoparticles were quantitatively determined by LC‐MS/MS (Thermo Q Exactive) and calculated using a standard curve method.

To evaluate the microenvironmental responsiveness of the nanovesicles, conditions mimicking physiological pH 7.4 and inflammation‐associated acidic pH 6.5, as well as oxidative stress with 100 µM H₂O₂, were set up as simulated ROS environments. Vesicle samples were incubated in different buffer solutions (PBS or H₂O₂ solution at 37°C), collected at time points of 0, 1, 2, 4, 8, 12 and 24 h to monitor changes in particle size and zeta potential drift. Drug release rates were assessed using a dialysis bag method with PBS (pH 7.4 or pH 6.5) as release media; samples taken at specified intervals were analyzed for remaining drug content via LC‐MS/MS to plot release kinetics curves. The fabrication process and physicochemical characterization workflow of the Mac@WZB117 nanovesicles are schematically illustrated in Figure S3.

2.3. In Vitro M1 Macrophage Model and Pharmacological Intervention

An in vitro M1 polarization model was established using the murine macrophage cell line RAW264.7 (ATCC, TIB‐71). Cells were seeded in 6‐well plates at a density of 1 × 10⁶ cells/mL and, after adherence, were stimulated for 24 h in RPMI‐1640 medium (containing 10% foetal bovine serum [FBS]) supplemented with 100 ng/mL LPS (Sigma) and 20 ng/mL IFN‐γ (PeproTech) to induce M1 phenotype polarization. Polarization efficiency was verified by assessing iNOS and TNF‐α mRNA expression levels via real‐time quantitative polymerase chain reaction (qPCR) and by measuring the proportion of CD86‐positive cells using flow cytometry. To ensure experimental consistency, all polarization procedures were performed under identical conditions, each experiment was independently repeated three times, and unstimulated RAW264.7 cells were used as M0 controls.

DiI‐labelled Mac@WZB117 nanovesicles (final concentration equivalent to WZB117 10 µM) were introduced into polarized M1 cells and incubated for 6 h. Subsequently, the cells were washed with PBS and fixed for analysis using confocal laser scanning microscopy (CLSM, Leica SP8) to observe the intracellular distribution and uptake efficiency of the nanovesicles. To verify subcellular localization, LysoTracker Green DND‐26 (75 nM, Invitrogen) was added post‐vesicle treatment to label lysosomes for colocalization imaging with DiI red fluorescence. Imaging parameters were kept constant, and fluorescence intensity was quantified using ImageJ software to assess uptake efficiency.

To evaluate the regulation of glycolysis function in macrophages upon intervention, extracellular acidification rates (ECAR) were measured using the Seahorse XF Extracellular Flux Analyzer (Agilent XF96), employing a Glycolysis Stress Test kit with glucose, oligomycin and 2‐DG injections set according to manufacturer instructions. Glucose uptake capacity was assessed using the fluorescent glucose analogue 2‐NBDG (100 µM, Thermo Fisher), with fluorescence intensity analyzed by fluorescence‐activated cell sorting (FACS) after a 30‐min incubation. Lactate concentrations were measured in cell culture supernatants using a commercial kit (BioVision, K627). Key factors in the glycolysis pathway (GLUT1, HK2, LDHA) were analyzed at both mRNA and protein levels via qPCR and Western blot. qPCR utilized a SYBR Green system with GAPDH as the reference gene; β‐actin served as a reference in Western blot, with membrane proteins extracted using RIPA lysis buffer supplemented with protease inhibitors.

For functional phenotype analysis, the expression levels of CD86 (APC‐conjugated) and inducible nitric oxide synthase (CD206; PE‐conjugated) were assessed by FACS (BD LSRFortessa) to evaluate changes in the M1/M2 polarization ratio. Cell culture supernatants were collected, and the concentrations of cytokines, including IL‐10, TGF‐β, TNF‐α and IL‐1β, were measured using ELISA kits (all from R&D Systems). All ELISA assays were performed in duplicate wells under standard curve conditions, and absorbance was read at 450 nm to calculate concentration changes. All cell experiments were independently repeated at least three times, and the investigators were blinded to sample identity to ensure the objectivity and reproducibility of data collection.

2.4. Macrophage‐Smooth Muscle Cell Co‐Culture System

To evaluate the paracrine regulatory effects of M1 macrophages on the functional state of HPASMCs (ScienCell) before and after glycolysis intervention, a Transwell non‐contact co‐culture system was established. M1‐polarized RAW264.7 cells, induced by LPS (100 ng/mL) and IFN‐γ (20 ng/mL) and subjected to various interventions (Free WZB117, Mac@WZB117, Mac@WZB117+2‐DG) for 48 h, were seeded at a density of 5 × 10⁵ cells/insert in the upper chamber of a Transwell with a pore size of 0.4 µm (Corning). HPASMCs were seeded in the lower chamber at 1 × 10⁵ cells/well in a 6‐well plate and cultured in a medium containing 1% FBS for another 48 h without changing the medium to ensure stable accumulation of paracrine factors.

The functional state of HPASMCs was assessed using various indicators. Cell proliferation was measured using an EdU cell proliferation kit (Beyotime, C0075S). Four hours before the end of co‐culture, EdU (10 µM) was added to the lower chamber cells. After incubation, fixation, click reaction and DAPI nuclear staining were performed according to the manufacturer's instructions. Imaging was conducted using a fluorescence microscope (Leica DMi8), and ImageJ software randomly selected five fields to count the proportion of EdU‐positive cells. Cell migration ability was evaluated through a scratch assay. Before starting co‐culture, uniform‐width gaps were created in HPASMC monolayers using a 200 µL pipette tip; after removing the supernatant and adding the Transwell system, images were recorded at 0 h and again after 48 h to calculate migration distance (% closure rate).

For synthetic phenotype transition evaluation, qPCR and Western blot were employed to detect expression levels of smooth muscle‐specific synthetic phenotype markers α‐smooth muscle actin (α‐SMA) and transgelin (SM22α). qPCR utilized SYBR Green dye with GAPDH as an internal control; reaction systems and cycling conditions followed standard protocols. A Western blot was performed using RIPA lysis buffer to extract total cellular proteins. Primary antibodies included α‐SMA (1:1000, Abcam) and SM22α (1:1000, Proteintech), with HRP‐conjugated species‐specific IgG as secondary antibodies. Protein bands were visualized using enhanced chemiluminescence (ECL) and quantified with Image Lab software. To further elucidate the signalling pathways activated by the paracrine mechanism, proteins related to the TGF‐β1/Smad2/3 and PDGF‐BB/PDGFR‐β pathways were examined, including TGF‐β1, phosphorylated Smad2/3, PDGF‐BB, PDGFR‐β and their phosphorylated forms.

All experiments were performed within the same cell passage and independently repeated three times to ensure data consistency. Image and Western blot band analyses were conducted in a blinded manner to guarantee objectivity and reproducibility. The coculture system was designed to mimic the pathophysiological process in which M1 macrophages regulate vascular smooth muscle function through paracrine signalling under inflammatory conditions, providing a basis for elucidating the in vitro mechanisms of glycolysis‐targeted therapy.

2.5. Multi‐Omics Sequencing and Metabolomics Analysis

To elucidate the impact of glycolysis intervention on the immune microenvironment of lung tissue and macrophage functional states, we conducted scRNA‐seq analysis on lung tissues from treated mice. Fresh lung tissues were sectioned into 1 mm³ pieces and digested enzymatically with a mixture of IV collagenase D (2 mg/mL) and DNase I (50 U/mL) at 37°C in a water bath for 30 min to prepare single‐cell suspensions. The cells were filtered through a 40 µm mesh, red blood cells removed, and libraries constructed using the 10× Genomics Chromium Next GEM Single Cell 3' Reagent Kit v3.1. Each sample aimed to capture between 6000 and 10,000 cells. Library quality was assessed using an Agilent Bioanalyzer for fragment analysis, with paired‐end sequencing (PE150) performed on an Illumina NovaSeq 6000 platform, ensuring a minimum of 50,000 reads per cell.

The sequencing data were aligned to the reference genome (mm10), gene quantification, cell identification and quality control filtering were performed using Cell Ranger v7.0. Strict quality control (QC) was applied to filter low‐quality cells: cells with mitochondrial gene content >15% or unique feature counts (nFeature_RNA) <200 or >6000 were excluded to remove dead cells and doublets (Figure S4A,B). Downstream analysis with Seurat v4.4. To correct for batch effects between biological replicates, the Harmony algorithm was employed on the PCA embeddings (dims = 1:30) to integrate datasets. Low‐quality cells with mitochondrial gene content >15% or fewer than 200 genes were excluded. Data normalization was conducted using ‘LogNormalize’, and the first 30 principal components from PCA were selected for t‐SNE dimensionality reduction clustering. Immune cell subtypes were annotated based on the Mouse Cell Atlas database standards, further identifying M1 and M2 macrophages through canonical markers such as F4/80, CD11b, CD206, iNOS, CD86, and so forth.

To determine the spatial localization and functional state of macrophages within the lung tissue microenvironment, fresh‐frozen OCT‐embedded lung tissue sections (10 µm thick) underwent spatial transcriptomics analysis. Tissue preprocessing, in situ cDNA synthesis, and library construction were performed using the 10× Visium Spatial Gene Expression platform. Imaging was conducted using a Leica Aperio AT2 platform for H&E staining imaging, with spatial cluster mapping, feature gene identification, and spatial annotation facilitated by Space Ranger software. Spatial expression data were further anchored to the annotated macrophage subpopulations from the scRNA‐seq dataset. This integration was performed using Seurat's anchor‐based transfer workflow. Specifically, FindTransferAnchors was run using canonical correlation analysis as the reduction method with default parameters (normalization.method = ‘LogNormalize’), followed by TransferData to predict the spatial location of scRNA‐seq‐defined cell types (Figure S4C). The Seurat HTO integration function was applied to combine the spatial expression data with the scRNA‐seq annotations, enabling detailed analysis of the distributional changes of M1 and M2 macrophages.

To explore specific metabolic pathways involved in glycolysis and energy metabolism remodelling, lung tissue samples and magnetically sorted F4/80⁺ pulmonary macrophages (MACS method by Miltenyi) were subjected to metabolomics analysis. Samples were treated with pre‐cooled 80% methanol; supernatants collected after centrifugation were dried under nitrogen gas and resuspended in a 50% acetonitrile aqueous solution before targeted metabolomic detection via high‐resolution mass spectrometry (Thermo Fisher Q Exactive). Raw peak areas processed by Compound Discoverer were imported into MetaboAnalyst 5.0 for data normalization, PCA analysis, and differential metabolite screening (variable importance in projection [VIP] > 1, p < 0.05), followed by enrichment and network analyses of significant metabolic pathways.

2.6. In Vivo Pharmacodynamics and Histological Analysis

To assess the pulmonary targeting capability of biomimetic nanovesicles, Mac@WZB117, DiR near‐infrared fluorescent dye (Ex/Em = 748/780 nm, Invitrogen) was incorporated into the lipid phase of the vesicles during preparation at a final concentration of 2 µg/mg PLGA. The mixture was incubated in the dark for 30 min to remove free dye, followed by three washes with PBS. DiR‐Mac@WZB117 was administered via tail vein injection into SU5416+hypoxia‐induced PAH model mice, with a dose calculated as WZB117 equivalent to 10 mg/kg of WZB117. At 24 h post‐administration, whole‐body fluorescence signals were captured using an IVIS Spectrum in vivo imaging system (PerkinElmer) with an exposure time of 1–2 s. After standard background correction, semi‐quantitative analysis of pulmonary fluorescence intensity was performed using ROI measurements. Upon completion of the experiment, animals were immediately dissected to obtain major organs, including lungs, liver, spleen and kidneys, for ex vivo imaging to evaluate vesicle distribution and specificity of pulmonary accumulation.

To evaluate the effects of therapeutic intervention on haemodynamic improvement in PAH, right heart catheterization was performed at the end of treatment to measure RVSP, mPAP and right ventricular hypertrophy index (RVHI). After anaesthetizing the mice, a window was opened beside the sternum to insert a microcatheter (Millar SPR‐1000) into the right ventricle and pulmonary artery. Pressure curves were recorded and synchronously collected using a PowerLab system, with averages from three stable respiratory cycles analyzed. For quantifying pulmonary vascular remodelling extent, high‐resolution three‐dimensional pulmonary vascular imaging (resolution 9 µm) was conducted using a SkyScan 1176 microCT system. Following contrast‐enhanced X‐ray scanning of lung tissues, CTAn software reconstructed the pulmonary vasculature network to analyze metrics such as vessel density, diameter distribution and small artery branching frequency.

For histological analysis, the left lung was harvested for paraffin embedding and sectioned at a thickness of 5 µm. H&E staining was performed to examine pulmonary vascular morphology and the degree of muscularization, while Masson's trichrome staining was used to assess collagen deposition around the vessels and within the lung interstitium. Pulmonary arteriole wall thickness (WT) was calculated using wall thickness percentage (WT%) = (outer membrane thickness‐inner membrane thickness)/outer membrane thickness × 100%, with 10 randomly selected arterioles per section measured for average analysis. Immunohistochemistry (IHC) detection involved primary antibodies, including CD206 (Abcam, 1:200), GLUT1 (Cell Signalling, 1:500), Ki67 (Proteintech, 1:300), α‐SMA (Sigma, 1:300), with HRP‐labelled IgG as secondary antibodies using the DAB chromogenic method. All stained slides were dehydrated and mounted before scanning and digitization under a Pannoramic SCAN system. Immunofluorescence (IF) staining for dual‐label colocalization analysis utilized Alexa Fluor series‐labelled secondary antibodies with DAPI counterstaining nuclei; high‐resolution image acquisition was performed using a confocal microscope (Zeiss LSM980). ImageJ software facilitated grayscale semi‐quantitative analysis across randomly selected fields, standardized for comparing protein expression levels and spatial distribution differences among groups. All tissue quantitative analyses were conducted under blind conditions by two independent observers to enhance data objectivity and consistency.

2.7. Validation of Macrophage Subpopulations and Glycolytic Function

To quantitatively assess the regulatory effects of glycolysis intervention on the polarization spectrum and metabolic status of lung tissue macrophages, we first prepared a single‐cell suspension from mouse lung tissue. Fresh lung tissue was minced and placed in an enzyme digestion solution containing collagenase IV (1 mg/mL), DNase I (50 U/mL) and Hank's buffer, followed by incubation in a 37°C water bath with shaking for 30 min. The suspension was then filtered through a 40 µm mesh and treated with RBC lysis buffer to remove red blood cells, resulting in a high‐viability single‐cell suspension. Cells were pre‐incubated with Fc Block (CD16/CD32) before being stained with fluorescently labeled antibodies F4/80‐APC (BioLegend), CD206‐PE (BioLegend) and iNOS‐FITC (Invitrogen) for surface and intracellular staining. Data were collected using a BD LSRFortessa flow cytometer. M1 macrophages were defined as F4/80⁺iNOS⁺, while M2 macrophages were defined as F4/80⁺CD206⁺. FlowJo v10 software was used for gating analysis to calculate changes in the proportion of each subpopulation within the total macrophage population. To ensure staining specificity and data accuracy, fluorescence‐minus‐one (FMO) controls and isotype controls were set up for fluorescence compensation and background correction.

Simultaneously, to verify the expression of metabolism‐related key genes and polarization phenotype factors, total RNA and protein from the right lung tissue were extracted for subsequent molecular detection. RNA extraction was performed using the TRIzol method, reverse transcription was conducted with the PrimeScript RT Kit (Takara) and qPCR utilized the SYBR Green dye system. qPCR primers were designed for genes such as GLUT1, HK2, LDHA, ARG1, iNOS, TGF‐β1, with GAPDH serving as an internal control. Reaction conditions included an initial denaturation at 95°C for 5 min, followed by 40 cycles of 95°C for 15 s and 60°C for 30 s. Protein samples were extracted using RIPA lysis buffer and quantified via the BCA method. After SDS‐PAGE electrophoresis and membrane transfer, blocking was performed with 5% BSA. Primary antibodies against GLUT1 (1:1000), HK2 (1:1000), LDHA (1:1000), ARG1 (1:2000), iNOS (1:1000) and TGF‐β1 (1:1000) were incubated overnight, followed by HRP‐conjugated secondary antibody incubation and ECL development. Band intensity was semi‐quantitatively analyzed using Image Lab software.

To obtain structural fingerprint information on glycolysis metabolism, a Raman confocal microscopy system (Horiba LabRAM HR Evolution) was employed for label‐free detection of macrophage metabolic status in lung tissue. Tissue sections with a thickness of 10 µm underwent preprocessing, including PBS washing, air drying and short dehydration without staining. The excitation wavelength was set at 785 nm with a scan range of 400–1800 cm⁻¹, focusing on intensity changes at glycolysis‐specific peaks such as 1120 cm⁻¹ (C‐C stretching) and 1450 cm⁻¹ (CH₂ bending). Single‐cell level spectra underwent background subtraction and baseline correction before normalization and PCA using LabSpec 6 software. These results were spatially matched with M1/M2 cell regions identified by FACS to quantitatively evaluate the degree of metabolic reprogramming in macrophage subgroups across different groups. This spectroscopic analysis provides structural and functional complementarity to traditional protein/nucleic acid methods, aiding in non‐invasive assessment of dynamic shifts in immunometabolic states.

2.8. Remedial Experiments and Toxicological Analysis

To verify the functional importance of GLUT1 in the therapeutic effect and to confirm the specificity of the intervention pathway, a rescue experiment was designed. In the Mac@WZB117 treatment group, mice additionally received intraperitoneal injections of either the glycolytic substrate 2‐DG (500 mg/kg, Sigma) or the GLUT1 agonist Bay‐876 (10 mg/kg, MedChemExpress) once daily for seven consecutive days. The rescue treatments began on Day 7 after PAH model induction and were administered concurrently with the primary intervention, with 15 mice per group. At the end of treatment, pulmonary arterial pressure (RVSP, mPAP), immune phenotypes (CD206⁺/iNOS⁺ ratio and CD206⁺ M2 proportion), vascular smooth muscle cell synthetic phenotype markers (α‐SMA and SM22α expression) and pulmonary glycolytic metabolites (lactate and pyruvate) were assessed as described above to determine whether the rescue interventions reversed the therapeutic effects of Mac@WZB117, thereby further validating the causal role of glycolysis inhibition in improving PAH.

In terms of toxicological safety evaluation, peripheral blood was collected from all treatment groups post‐therapy via orbital bleeding. Serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN) and creatinine (Cr) were measured using an automated biochemical analyzer (Hitachi 7020) to assess potential hepatic and renal impairment. Major organs (heart, liver, lungs, kidneys) were simultaneously harvested and fixed in 4% paraformaldehyde for paraffin embedding and H&E staining to observe pathological changes such as structural integrity, cellular morphology, inflammatory infiltration and necrotic areas. All pathological slides were evaluated in a double‐blind manner by two experienced pathologists who quantitatively graded tissue damage according to standardized pathological scoring criteria. Integrating serum indicators with histological results allowed for a comprehensive assessment of the systemic toxicity and organ‐targeting safety of the nanoplatform, providing a basis for subsequent clinical translation.

2.9. Statistical Analysis

All experimental data are expressed as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism v9.0 and R v4.2.2. For comparisons between two groups that met the assumptions of normality and homogeneity of variance, two‐tailed Student's t‐tests were applied. For comparisons among multiple groups, one‐way analysis of variance (ANOVA) was used, followed by Tukey's post hoc test when significant differences were detected. Data that were not normally distributed were analyzed with the Mann‐Whitney U test. For multi‐omics differential gene and metabolite analyses, multiple hypothesis testing was corrected using the Benjamini‐Hochberg method to control the FDR. Time‐series data were analyzed by repeated‐measures ANOVA. Correlation analyses were performed using either Pearson or Spearman correlation coefficients. A p value < 0.05 was considered statistically significant.

3. Results

3.1. Characterization and In Vitro Targeting of Mac@WZB117 Nanovesicles

The synthesized Mac@WZB117 nanovesicles exhibited a uniform spherical morphology with an average diameter of ∼164 nm, a PDI of 0.18, and a negative surface charge of ‐22.6 mV, confirming high colloidal stability (Figure S5A‐C). Successful encapsulation of WZB117 and fusion of macrophage membrane proteins were verified by UV‐Vis spectroscopy and Western blot analysis, respectively (Figure S5D‐E). In drug release assays, the nanovesicles demonstrated significant pH‐ and ROS‐responsive behaviour, with cumulative release rates at 24 h reaching ∼70% under simulated inflammatory conditions (pH 6.5 or 100 µM H₂O₂), compared to only 28.6% in neutral PBS (Figures S5F and S6A,B). Furthermore, cellular uptake experiments revealed that Mac@WZB117 was efficiently internalized by M1‐polarized macrophages via the endolysosomal pathway, showing a 3.2‐fold higher intracellular accumulation compared to free WZB117 and significantly superior uptake relative to uncoated PLGA vesicles (Figure S7A‐D). These data collectively validate the physicochemical integrity, stimuli‐responsive release and active targeting capability of the Mac@WZB117 platform.

3.2. Mac@WZB117 Significantly Reduces Glycolysis Activity in M1 Macrophages

To systematically elucidate the mechanism by which Mac@WZB117 intervenes in the metabolic programming of M1 macrophages, ECAR was assessed using the Seahorse XF Analyzer to evaluate glycolysis function. The results demonstrated that glycolysis levels in LPS/IFN‐γ‐induced M1 macrophages were significantly elevated, whereas treatment with Mac@WZB117 markedly inhibited both basal glycolysis and maximum glycolysis rates, resulting in a 48.1% reduction in overall ECAR within 24 h (Figure 1A, p  < 0.001, n = 3). Furthermore, glycolysis reserve was notably decreased, indicating that its inhibitory effect extended to adaptive metabolic responses under energy stress conditions.

FIGURE 1.

Inhibition of glycolytic activity in M1 macrophages by Mac@WZB117. (A) Seahorse analysis demonstrated that Mac@WZB117 markedly reduced the basal glycolysis rate, maximal glycolytic capacity and glycolytic reserve of M1 macrophages (p < 0.001, n = 3). (B) 2‐NBDG fluorescence imaging revealed a pronounced decrease in intracellular fluorescence following Mac@WZB117 treatment, indicating diminished glucose uptake capacity. (C) Quantitative fluorescence analysis showed a ∼35.6% reduction in glucose uptake in the Mac@WZB117 group compared with the M2 control group (p < 0.01, n = 3). (D) Colourimetric assay of lactate secretion indicated that lactate concentration in the Mac@WZB117 group was reduced by 42.9% relative to the M1 control (p < 0.001). (E) qPCR analysis revealed significant downregulation of the key glycolytic genes GLUT1, HK2 and LDHA (p < 0.01). (F) Western blot confirmed corresponding changes at the protein level, demonstrating that Mac@WZB117 significantly suppresses glycolytic pathway activity. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Subsequently, glucose uptake capacity was monitored in real‐time using the glucose fluorescent probe 2‐NBDG. Confocal imaging and fluorescence quantification revealed a significant reduction in glucose uptake intensity in the Mac@WZB117 group, with average fluorescence intensity decreasing by 35.6% compared to the M2 control (Figure 1B,C, p  < 0.01). Additionally, colourimetric analysis of lactate secretion indicated that Mac@WZB117 reduced lactate production from 2.31 ± 0.18 mmol/L to 1.32 ± 0.12 mmol/L (Figure 1D, a 42.9% decrease, p < 0.001, n = 3), suggesting a substantial restriction in the synthesis of glycolytic end products.

At the molecular level, qPCR and Western blot analyses were employed to examine changes in the expression of key glycolytic pathway molecules. Results indicated that Mac@WZB117 significantly downregulated the transcription and protein levels of GLUT1, HK2 and LDHA, with mRNA levels decreasing by 50%–65%, and consistent downregulation observed at the protein level (Figure 1E,F, p  < 0.01, n = 3). Taken together, these findings demonstrate that Mac@WZB117 effectively blocks the glycolytic metabolic flux of M1 macrophages by inhibiting glucose uptake and the expression of key glycolytic enzymes.

3.3. Mac@WZB117 Promotes Reprogramming of Macrophages From M1 to M2 Phenotype

To investigate the role of Mac@WZB117 in regulating the immune polarization direction of macrophages, an M1‐type RAW264.7 model was established using LPS/IFN‐γ induction, and changes in phenotype marker expression were assessed via FACS. Results indicated that in the untreated M1 model, the proportion of CD86⁺ cells was as high as 82.3%. However, after 24 h of Mac@WZB117 treatment, this proportion significantly decreased to 38.7% (Figure 2A, p  < 0.001, n = 3). Concurrently, the proportion of CD206⁺ cells markedly increased from 12.5% to 44.6% (Figure 2B), indicating that Mac@WZB117 effectively suppresses the pro‐inflammatory M1 phenotype and promotes the expression of the anti‐inflammatory/repair‐associated M2 phenotype.

FIGURE 2.

Mac@WZB117 promotes M1‐to‐M2 repolarization of macrophages. (A) Flow cytometry showed that Mac@WZB117 significantly reduced the proportion of M1 macrophages (p < 0.001, n = 3). (B) Flow cytometry detected a marked increase in CD206⁺ M2 cells (p < 0.001), indicating enhanced anti‐inflammatory phenotype. (C) qPCR analysis of M1/M2 marker gene expression (p < 0.01, p < 0.001). (D) IF double staining shows changes in iNOS and CD206 protein expression. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

To further confirm the molecular basis of phenotype remodelling, qPCR analysis was conducted. The results demonstrated a significant 61.8% reduction in the mRNA levels of the M1 marker iNOS (Nos2) compared to the control group, while the expression of the M2 marker Arg1 increased nearly 2.9‐fold (Figure 2C, p  < 0.01‐0.001), showing a trend highly consistent with FACS results.

In addition, to further confirm the protein‐level changes associated with this polarization shift, IF double staining was performed to assess the spatial localization of CD206 and iNOS expression. The results showed markedly reduced iNOS fluorescence and enhanced CD206 fluorescence in the Mac@WZB117‐treated group, with CD206 displaying a diffuse cytoplasmic distribution, visually demonstrating the functional transition of macrophages from a pro‐inflammatory to a reparative state (Figure 2D).

Collectively, these findings suggest that Mac@WZB117 not only intervenes in glycolysis metabolism but also directs macrophage immune fate transformation through multidimensional molecular and phenotypic mechanisms.

3.4. Significant Alteration in Macrophage Secretory Profile and Enhanced Immune Activation

To comprehensively assess the immunomodulatory effects of Mac@WZB117 on macrophage secretion functions, we employed an ELISA assay to measure changes in the concentrations of key cytokines in the supernatant of LPS/IFN‐γ‐induced M1‐polarized RAW264.7 cells before and after treatment. After 24 h of treatment, a significant reduction in the secretion of anti‐inflammatory cytokines TNF‐α and IL‐β was observed, dropping from 212.6 ± 14.3 pg/mL and 185.4 ± 12.1 pg/mL to 95.2 ± 10.9 pg/mL and 70.8 ± 8.6 pg/mL, respectively, representing decreases of 55.2% and 61.8% (Figure 3A, p  < 0.01, n = 3). These findings suggest that Mac@WZB117 effectively suppresses the classical pro‐inflammatory response of M1 macrophages.

FIGURE 3.

Mac@WZB117 reprograms macrophage secretory profile towards an anti‐inflammatory phenotype. (A) ELISA results showed that anti‐inflammatory factors TNF‐α and IL‐1β secretion decreased significantly by 55.2% and 61.8% (p < 0.01, n = 3). (B) Levels of pro‐inflammatory factors IL‐10 and TGF‐β increased by 43.9% and 39.7% (p < 0.01, n = 3), indicating enhanced immune activation. (C) Radar chart analysis of the four‐cytokine secretion profile. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Conversely, the levels of pro‐inflammatory cytokines IL‐10 and TGF‐β significantly increased, rising from 138.5 ± 9.2 pg/mL and 126.4 ± 10.1 pg/mL to 199.3 ± 11.5 pg/mL and 176.6 ± 9.8 pg/mL, respectively, indicating increases of 43.9% and 39.7% (Figure 3B, p  < 0.01, n = 3), which reflecting a secretory profile characteristic of a shift towards the M2 anti‐inflammatory/repair phenotype.

To provide an overall representation of changes in immune activation status, a radar chart integrating the normalized data of the cytokines above was constructed. The results showed that Mac@WZB117 treatment markedly reshaped the cytokine profile, shifting the overall secretion pattern from a pro‐inflammatory profile dominated by TNF‐α and IL‐1β to an anti‐inflammatory profile characterized by IL‐10 and TGF‐β (Figure 3C).

Taken together, these data demonstrate that Mac@WZB117 not only regulates macrophage metabolism and phenotype but also exerts immunomodulatory effects by reprogramming their secretory function.

3.5. Enhanced Inhibition of Smooth Muscle Cell Proliferation by Macrophages Post‐Intervention

To determine whether Mac@WZB117‐induced reprogramming of M1 macrophages confers paracrine ability to inhibit pathological proliferation of HPASMC, a Transwell co‐culture model was established with M1‐polarized RAW264.7 cells and HPASMC. EdU assays revealed that in the PAH + NS group, the proportion of EdU‐positive cells reached 67.2%, indicating that the smooth muscle cells were in a markedly hyperproliferative state. In contrast, co‐culture with Mac@WZB117‐treated macrophages reduced this proportion to 31.4%, a decrease of 53.3% (p < 0.001, n = 3) (Figure 4A), indicating that the intervention significantly mitigated the pro‐proliferative stimuli on downstream cells.

FIGURE 4.

Mac@WZB117‐treated macrophages suppress HPASMC proliferation via paracrine modulation. (A) EdU incorporation assay showed that HPASMCs in the PAH + NS group exhibited a high proliferative state, with an EdU‐positive rate of 67.2%, whereas the rate was markedly reduced to 31.4% in the Mac@WZB117 group (p < 0.001, n = 3). (B) Flow cytometric cell cycle analysis revealed a significant increase in the proportion of G1‐phase cells in the Mac@WZB117 group, reaching 71.2%, indicating G1 phase arrest. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Flow cytometric cell‐cycle analysis further showed that in the Mac@WZB117 treatment group, the proportion of cells in the G1 phase increased significantly from 52.7% in the control to 71.2% (p < 0.01), with a corresponding decrease in the S‐phase population (Figure 4B). This finding indicates that HPASMCs experienced G1‐phase cell‐cycle arrest and reduced proliferative capacity. These results suggest that, under intervention conditions, macrophages indirectly modulate vascular remodelling by altering their metabolic and immune regulatory effects on smooth muscle cells through paracrine signalling.

3.6. Significant Restriction in Smooth Muscle Cell Migration Capability

To investigate whether the paracrine effects of macrophages (macrophage) post‐Mac@WZB117 intervention could inhibit the migratory behaviour of HPASMC, a scratch assay co‐culture model of macrophage and HPASMC was established. In the experiment, HPASMCs were grown to confluence in the lower chamber of a Transwell setup, following which RAW264.7 cells from different treatment groups were introduced into the upper chamber. Scratch areas were imaged at 0 and 24 h, with migration areas and rates quantitatively analyzed using ImageJ (Figure 5A).

FIGURE 5.

Mac@WZB117‐treated macrophages suppress HPASMC migration capacity. (A) Representative images and quantitative analysis of scratch assays showed that the wound closure area at 24 h in the Mac@WZB117‐treated group was significantly smaller than that in the PAH + NS group (24.5% vs. 63.8%, p < 0.001, n = 3). (B) Calculation of cell migration rate revealed that the mean migration speed in the intervention group decreased to 0.11 µm/min (p < 0.001). Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

The results indicated that in the PAH + NS group, the scratch wound healing area reached 63.8% within 24 h, whereas in the Mac@WZB117 group, it was only 24.5%, a decrease of 61.6% (p < 0.001, n = 3), demonstrating that functional reprogramming of macrophages significantly weakens their promotion of smooth muscle cell migration. Further analysis of cell average migration rates (Figure 5B) showed that the Mac@WZB117 group had a rate of 0.11 µm/min, significantly lower than the 0.28 µm/min observed in the PAH + NS group (p < 0.001), indicating that the intervention effectively limits the dynamic migratory capacity of smooth muscle cells.

3.7. Downregulation of Synthetic Markers and Dedifferentiation of Smooth Muscle Cells

To further evaluate whether Mac@WZB117‐mediated macrophage reprogramming can modulate the phenotypic state of HPASMCs, we constructed a Transwell co‐culture system to analyze the expression changes of synthetic markers α‐SMA and SM22α (Figure 6A,B). In the model group (PAH + NS), qPCR results indicated that the mRNA levels of α‐SMA and SM22α were significantly elevated, displaying the typical ‘synthetic’ phenotype characteristic of PAH. In contrast, in the Mac@WZB117 intervention group, the levels were reduced by 62.3% and 51.9%, respectively (both p < 0.01, n = 3), suggesting a significant suppression of the synthetic phenotype.

FIGURE 6.

Mac@WZB117‐modulated macrophages suppress the synthetic phenotype of HPASMCs. (A) qPCR analysis of HPASMCs showed that mRNA expression levels of the synthetic markers α‐SMA and SM22α were reduced by 62.3% and 51.9%, respectively, in the Mac@WZB117 group compared with the PAH + NS group (both p < 0.01, n = 3). (B) Representative qPCR bar plots demonstrated that the intervention group markedly suppressed the expression of synthetic phenotype‐related genes. (C) Western blot analysis and densitometric quantification revealed decreased protein levels of α‐SMA and SM22α in the Mac@WZB117 group, consistent with the mRNA results, indicating a phenotypic shift of HPASMCs toward a functional state. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Further analysis via Western blot revealed a consistent trend at the protein level. The bands for α‐SMA and SM22α were notably diminished in the Mac@WZB117 group, with grayscale values decreased by 58.2% and 47.5%, respectively (Figure 6C), corroborating the biological consistency of RNA downregulation. This phenomenon indicates that macrophages, intervened with Mac@WZB117, can effectively suppress the synthetic phenotype induction on HPASMCs, potentially promoting a shift back to a contractile functional state of the smooth muscle cells.

3.8. Significant Inhibition of TGF‐β1 and PDGF‐BB Pathway Expression

To determine whether the reprogrammed M2 macrophages Mac@WZB117 can modulate critical vascular remodelling signalling pathway in smooth muscle cells through paracrine mechanisms, the classical signalling pathways TGF‐β1/Smad2/3 and PDGF‐BB/PDGFR‐β were assessed in a Transwell co‐culture system.

Western blot analysis revealed that in the PAH + NS group, the expression of TGF‐β1 protein in HPASMC was significantly elevated, with concurrent increases in the phosphorylation levels of p‐Smad2 and p‐Smad3, indicating an active signalling axis (Figure 7A). In contrast, in the Mac@WZB117 intervention group, TGF‐β1 protein expression decreased by approximately 54.8%, with phosphorylated Smad2 and Smad3 levels reduced by 46.1% and 49.7%, respectively (all p < 0.01, n = 3), suggesting a marked suppression of TGF‐β1/Smad signalling pathway activity (Figure 7B).

FIGURE 7.

Mac@WZB117 suppresses TGF‐β1 and PDGF‐BB signalling activity in HPASMCs. (A) Western blot analysis showed that the expression of TGF‐β1, p‐Smad2 and p‐Smad3 was markedly upregulated in HPASMCs from the PAH + NS group, whereas all three were suppressed in the Mac@WZB117 group. (B) Densitometric quantification of the bands revealed decreases of approximately 54.8%, 46.1% and 49.7% for TGF‐β1, p‐Smad2 and p‐Smad3, respectively (all p < 0.01). (C) The protein levels of PDGF‐BB and PDGFR‐β were elevated in the PAH + NS group, indicating a pro‐migratory state of HPASMCs; following intervention, they were reduced by 57.2% and 44.5%, respectively. (D) Bar graphs from densitometric analysis further quantified the extent of PDGF pathway suppression, indicating that Mac@WZB117 substantially attenuates the activation of HPASMCs by exogenous remodelling signals. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Simultaneously, the protein expression of PDGF‐BB and its receptor PDGFR‐β also showed high levels in the PAH + NS group, whereas in the Mac@WZB117 group, expressions were reduced by 57.2% and 44.5%, respectively (Figure 7C,D), indicating effective inhibition of the PDGF signalling pathway.

These results demonstrate that Mac@WZB117 reprograms macrophage function to suppress the release of pro‐vascular‐remodelling factors such as TGF‐β1 and PDGF‐BB, thereby blocking the pathological activation of HPASMCs at the signalling level.

3.9. Mac@WZB117 Demonstrates Enhanced Pulmonary Accumulation and Reduced Hepatic and Splenic Clearance

To systematically evaluate the in vivo targeting and distribution characteristics of the Mac@WZB117 nanodelivery platform, a PAH mouse model was established using SU5416 combined with hypoxia. Mice were injected with DiR‐labelled Mac@WZB117, uncoated PLGA@WZB117 or free WZB117, and fluorescence distribution was monitored 24 h later using an IVIS Spectrum in vivo imaging system.

The results showed that the Mac@WZB117 group exhibited markedly enhanced DiR fluorescence in the lung region (Figure 8A), with an average fluorescence intensity of (3.26 ± 0.41) × 10⁷ photons/s—3.5 times and 2.8 times higher than the free WZB117 and PLGA@WZB117 groups, respectively—and significantly superior to the PLGA@WZB117 group (p < 0.001, n = 15) (Figure 8B). These findings indicate that macrophage membrane modification markedly improves the active targeting and enrichment of the nanosystem in the lungs. Quantitative fluorescence analysis further confirmed that Mac@WZB117 possesses excellent pulmonary homing capability, likely due to the combined effects of macrophage membrane‐derived recognition mechanisms and passive targeting enrichment.

FIGURE 8.

Mac@WZB117 selectively accumulates in lung tissue while avoiding hepatic and splenic clearance. (A) IVIS imaging recorded organ fluorescence distribution in mice 24 h after treatment, showing that the Mac@WZB117 group exhibited the strongest signal in the lung region, markedly higher than that in the control group. (B) Quantitative analysis of pulmonary fluorescence intensity revealed a mean photon flux of (3.26 ± 0.41) × 10⁷ photons/s in the Mac@WZB117 group, which was 3.5‐fold greater than that of the free drug group (p < 0.001, n = 15). (C) Fluorescence intensity in the liver and spleen was significantly reduced by 42.7% and 37.9%, respectively (p < 0.01), indicating a pronounced advantage of this nanoplatform in evading RES uptake. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Additionally, in organs enriched by the reticuloendothelial system (RES), the Mac@WZB117 group showed significantly lower fluorescence signals compared to the non‐encapsulated and free groups. In the liver and spleen regions, fluorescence intensity decreased by 42.7% and 37.9%, respectively (p < 0.01) (Figure 8C), suggesting that this system can extend circulation half‐life while partially avoiding RES clearance, thereby delaying peripheral clearance rates.

3.10. Mac@WZB117 Significantly Improves Right Heart Pressure Metrics

In the SU5416 plus hypoxia‐induced PAH mouse model, right‐heart catheterization was performed to assess the effects of Mac@WZB117 on haemodynamics and ventricular remodelling.

Compared to the Sham group, mice in the PAH + NS group exhibited a marked increase in RVSP, reaching 43.2 ± 2.5 mmHg, characteristic of classic PAH. Following intervention with Mac@WZB117, RVSP significantly decreased to 27.4 ± 2.0 mmHg, representing a reduction of 36.5% (p < 0.001, n = 15) (Figure 9A). Similarly, the mPAP was reduced from 26.0 ± 1.7 mmHg to 16.0 ± 1.4 mmHg, a decrease of 38.5% (p < 0.001) (Figure 9B), indicating a significant reduction in pulmonary vascular resistance (PVR).

FIGURE 9.

Mac@WZB117 significantly improves right ventricular haemodynamic in PAH mice. (A) RVSP changes measured by Millar catheter measurements showed that RVSP was markedly elevated in the PAH + NS group but was significantly reduced to 27.4 ± 2.0 mmHg in the Mac@WZB117 group, approaching the level observed in the Sham group (p < 0.001). (B) mPAP changes showed a similarly significant decrease, dropping from 26.0 ± 1.7 mmHg to 16.0 ± 1.4 mmHg (p < 0.001). (C) RVHI changes showed that the RVHI declined from 0.52 ± 0.04 to 0.27 ± 0.03 (p < 0.001), indicating relief of right ventricular load and reduced pulmonary circulatory pressure. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Regarding ventricular remodelling, the RVHI in the Mac@WZB117‐treated group decreased from 0.52 ± 0.04 to 0.27 ± 0.03, a reduction of 47.8% (p < 0.001) (Figure 9C), indicating a marked alleviation of right‐heart load.

These results demonstrate that Mac@WZB117 effectively improves haemodynamic abnormalities and ventricular remodelling in PAH by targeting macrophage metabolic reprogramming in the pulmonary vascular microenvironment.

3.11. Significant Alleviation of Pulmonary Vascular Remodelling and Reduced Fibrosis

To systematically evaluate the in vivo effects of Mac@WZB117 on PAH‐associated pulmonary vascular remodelling and tissue fibrosis, a comprehensive histological analysis was conducted on lung tissues from various treatment groups. H&E staining revealed a notable thickening of the pulmonary arteriole walls and narrowing of the vascular lumen in the PAH + NS group, with a WT% of 35.1% ± 2.7%. Following intervention with Mac@WZB117, this metric significantly decreased to 17.6% ± 1.9% (p < 0.001, n = 6), indicating effective inhibition of medial hypertrophy and smooth muscle proliferation (Figure 10A,B).

FIGURE 10.

Mac@WZB117 alleviates pulmonary vascular remodelling and fibrosis in PAH mice. (A) H&E staining revealed marked thickening of the pulmonary arteriole walls in the PAH + NS group, with a WT% of 35.1%, whereas the Mac@WZB117 group exhibited substantially reduced WT% (= 17.6%). (B) Quantitative analysis of WT% is presented as bar graphs (n = 6, p < 0.001). (C) Masson staining showed intensified blue staining in the lung parenchyma and perivascular regions of the model group, indicative of collagen deposition; in contrast, the Mac@WZB117 group displayed a markedly reduced stained area. (D) Quantification of fibrotic area using Masson images demonstrated a 49.3% reduction in fibrotic area in the Mac@WZB117 group (p < 0.001). (E) Statistical analysis showing changes in WT% and fibrotic area across groups. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Furthermore, Masson's trichrome staining was employed to assess collagen fibre deposition within lung tissues. In the PAH + NS group, there was a marked increase in blue‐stained areas within the lung parenchyma and perivascular regions, with fibrosis occupying 21.4% ± 1.6% of the area. In contrast, the Mac@WZB117 group exhibited a significant reduction in this phenomenon, with collagen deposition area reduced to 10.9% ± 1.3% (p < 0.001, n = 6), achieving an inhibition rate of nearly 50% (Figure 10C,D), indicating a pronounced attenuation of collagen accumulation and pulmonary fibrosis.

Taken together, these morphological parameters (Figure 10E) demonstrate that Mac@WZB117 effectively alleviated pulmonary vascular remodelling and fibrosis in PAH mice by reducing small‐artery medial thickness and smooth muscle hypertrophy, as well as by inhibiting fibrin deposition.

3.12. Downregulation of GLUT1 Expression in Macrophages

To further confirm the in vivo targeting effect of Mac@WZB117 on the glycolytic pathway of macrophages, IHC and IF co‐localization were performed to analyze the expression and cellular localization of GLUT1 in lung tissue. IHC results indicated a marked enhancement of GLUT1‐positive staining in the perivascular regions of the PAH + NS model group, characterized by dark brown patches primarily concentrated in areas rich in perivascular inflammatory cells. In contrast, post‐treatment with Mac@WZB117, GLUT1 expression was significantly attenuated, with a notable reduction in staining area, suggesting effective metabolic intervention by this delivery system in vivo (Figure 11A).

FIGURE 11.

Mac@WZB117 suppresses GLUT1 expression in M2 macrophages within the pulmonary vascular niche. (A) IHC revealed markedly enhanced GLUT1‐positive signals in the perivascular regions of the PAH + NS group, whereas staining intensity and distribution were substantially reduced in the Mac@WZB117‐treated group, indicating effective in vivo suppression of GLUT1 expression. (B) IF co‐staining images showed strong colocalization of GLUT1 (red) with CD206 (green) in the PAH + NS group, whereas the colocalized regions were notably reduced in the Mac@WZB117 group. (C) Quantitative analysis of colocalized areas demonstrated that the proportion of GLUT1/CD206 double‐positive cells decreased from 65.4% to 21.3% (p < 0.001, n = 15), and the Pearson's correlation coefficient dropped from 0.78 to 0.32. These findings suggest that Mac@WZB117 selectively modulates the glycolytic phenotype of M1 macrophages. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Double IF staining for GLUT1 and the M1 marker CD86 revealed strong co‐localization in the PAH + NS group, with the double‐positive area accounting for 65.4% ± 3.2% of the total fluorescence and a Pearson correlation coefficient of 0.78. These findings indicate that macrophages in this group exhibit high GLUT1 expression and an activated glycolytic metabolism. However, in the Mac@WZB117 intervention group, this co‐expression area significantly decreased to 21.3% ± 2.4%, with the Pearson coefficient dropping to 0.32, indicating a significant suppression of GLUT1 expression in M1 macrophages (Figure 11B,C).

These results reveal at the spatial tissue level that the Mac@WZB117 nanoplatform achieves targeted recognition through macrophage membrane modification and successfully delivers a GLUT1 inhibitor to glycolysis‐dependent M1 macrophages, specifically downregulating GLUT1 expression.

3.13. Correction of Macrophage Polarization Ratios In Vivo

To assess whether Mac@WZB117 can modulate macrophage polarization states in a PAH model in vivo, FACS was employed to analyze different phenotypic macrophage populations in single‐cell suspensions from lung tissue. F4/80 was used as a marker for total macrophages, while CD86 and CD206 were employed to identify M1 and M2 subpopulations, respectively. This approach enabled a quantitative evaluation of the drug's impact on the immune microenvironment.

In the PAH + NS model group, the proportion of F4/80⁺CD86⁺ M1 macrophages reached 47.6% ± 3.8%, indicating a strong enrichment of pro‐inflammatory macrophages in the lung tissue. After Mac@WZB117 treatment, this population decreased to 18.9% ± 2.3% (p < 0.01, n = 5) (Figure 12A,B). Conversely, the proportion of F4/80⁺CD206⁺ M2 macrophages, which was only 7.3% in the PAH + NS group, rose significantly to 20.8% following Mac@WZB117 treatment (p < 0.01), nearly restoring it to the level observed in the Sham control group (Figure 12C). These changes indicate a pronounced reversal of macrophage polarization, with a shift away from the pro‐inflammatory activation pathway toward an anti‐inflammatory, reparative phenotype.

FIGURE 12.

Mac@WZB117 reverses macrophage polarization imbalance in lung tissues of PAH mice. (A) Flow cytometry gating strategy: after FSC/SSC selection, F4/80⁺ was used as a macrophage marker, with CD86 and CD206 labelling M1 and M2 phenotypes, respectively. (B) The proportion of M1 macrophages decreased significantly to 9.2% ± 1.5%, then rose to 28.7% ± 2.6% after Mac@WZB117 intervention (p < 0.01, n = 5). (C) The proportion of F4/80⁺CD206⁺ M2 macrophages increased to 45.3% ± 3.7% in the PAH + NS group and decreased to 20.1% ± 2.2% after Mac@WZB117 treatment (p < 0.01). Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Additionally, the proportion of double‐positive (F4/80⁺CD86⁺CD206⁺) cells notably declined, indicating a clear and stable trend toward M1 to M2 reprogramming. This result aligns closely with transcriptional and protein expression changes observed in in vitro M1 polarization models, further corroborating that Mac@WZB117, through targeted metabolic intervention, effectively remodels immune cell lineages in the complex lung microenvironment. This intervention weakens the immunosuppressive state and enhances pro‐inflammatory functions, potentially impacting downstream fibrosis and vascular remodelling processes.

3.14. scRNA‐seq Reveals Characteristics of Immune Metabolic Outcomes

To further elucidate the regulatory mechanisms by which Mac@WZB117 affects immune cell lineages and metabolic features in lung tissue, scRNA‐seq was performed on lung samples from PAH model mice using the 10× Genomics platform. Two groups were included (PAH + NS and PAH + Mac@WZB117) with biological replicates. After initial quality control with Cell Ranger and standard processing with Seurat, the datasets were integrated and subjected to clustering analysis. A total of nine distinct cell populations were identified, including macrophages, T cells, B cells, neutrophils, NK cells and endothelial cells (Figure 13A,B).

FIGURE 13.

scRNA‐seq revealed immune‐metabolic characteristics of macrophage subsets. (A) t‐SNE plot of lung tissue cell clustering. (B) Bubble plot of characteristic gene expression across cell populations. (C) t‐SNE plot showing macrophage M1‐like and M2‐like subclusters. (D) t‐SNE plot of macrophage marker gene distribution (Chil3, Arg1, Mrc1, Nos2, Il1b, Cd86). (E) Bar chart of M1‐like and M2‐like cluster proportions in different treatment groups. (F) Heatmap of expression changes in glycolysis, lipid metabolism, TGF‐β pathway and inflammatory genes.

Further subclustering of the macrophage population based on marker genes identified two subsets: M1‐like and M2‐like cells (Figure 13C). Feature gene distribution analysis showed that M2‐like macrophages were highly associated with Chil3, Arg1 and Mrc1, whereas M1‐like macrophages predominantly expressed inflammatory molecules such as Nos2, Il1b and Cd86 (Figure 13D). Comparison of sample proportions revealed that Mac@WZB117 treatment significantly increased the M2‐like cluster while decreasing the M1‐like cluster (Figure 13E). Differential gene expression analysis further demonstrated that in the M1‐like cluster, the expression levels of glycolytic molecules (Slc2a1/GLUT1, Hk2, Ldha), lipid metabolism genes (Acly, Fasn), TGF‐β pathway components (Tgfb1, Smad2) and pro‐inflammatory factors (Nos2, Cxcl10, Il1b) all showed a downward trend (Figure 13F). Taken together, these findings indicate that Mac@WZB117 effectively remodels the immunometabolic profile of macrophages, promoting a shift from an inflammatory to a reparative phenotype.

3.15. Spatial Transcriptomics Reveals Restructuring of Macrophage Spatial Distribution

To further dissect the regulatory effects of Mac@WZB117 on macrophages within the lung microenvironment at the tissue level, spatial transcriptomics was performed on lung tissue sections using the 10× Genomics Visium platform. By aligning with H&E‐stained images and annotating tissue structures, multiple functional regions and cell populations were successfully identified (Figure 14A). The results revealed heterogeneous spatial distributions of M1 and M2 macrophages in lung tissue (Figure 14B). In the PAH + NS group, strong dual‐positive signals for GLUT1 (Slc2a1) and the M1 marker CD86 were highly enriched, whereas these signals were markedly reduced in the Mac@WZB117 group (Figure 14C), indicating correction of inflammatory clustering. Further spatial transcriptomic analysis showed that several genes associated with immunometabolism and tissue remodelling—such as Smad3 (a downstream effector of TGF‐β signalling), Col1a1 (collagen synthesis) and Ldha (lactate metabolism)—were significantly downregulated in the Mac@WZB117 group (Figure 14D,E). These findings demonstrate that local fibrosis and metabolic remodelling networks were suppressed, and the lung tissue microenvironment exhibited a distribution pattern consistent with a shift towards an anti‐inflammatory, low‐metabolic steady state.

FIGURE 14.

Spatial transcriptomics analysis revealed macrophage spatial distribution and functional pathway alterations. (A) Spatial clustering map identifying functional regions of lung tissue sections. (B) Distribution patterns of M1 and M2 macrophages within the tissue. (C) Spatial expression of Slc2a1 and Cd86. (D) Expression of Smad3, Col1a1 and Ldha in lung tissues from different treatment groups. (E) Expression levels of Smad3, Col1a1 and Ldha genes across groups.

3.16. Targeted Metabolomics Reveals Systematic Downregulation of Glycolytic Pathways

To further elucidate the mechanism by which Mac@WZB117 modulates immunometabolism, a targeted metabolomics strategy was employed to systematically analyze glycolysis and aerobic metabolic fluxes in macrophages. The metabolic profiles of lung tissues from various treatment groups and F4/80⁺ macrophages isolated via magnetic bead sorting were collected using the LC‐MS/MS platform. These were interpreted through PCA and pathway enrichment. PCA results distinguished samples from the Mac@WZB117 group and the PAH + NS model group, indicating significant differentiation in overall metabolic characteristics (Figure 15A). Volcano plot analysis revealed that key metabolites of glycolysis and the TCA cycle, such as lactate, pyruvate and citrate, were generally downregulated in the Mac@WZB117 intervention group (Figure 15B), indicating a systemic inhibitory trend at the metabolic level.

FIGURE 15.

Mac@WZB117 suppresses glycolysis and the TCA cycle in lung tissue and macrophages. (A) PCA plot showing a clear separation between PAH + NS and Mac@WZB117 groups in the metabolic profile dimension, indicating a marked shift in overall metabolic status. (B) Volcano plot illustrating downregulation of multiple key metabolites in carbohydrate metabolism, including lactate, pyruvate and citrate, in the intervention group (n = 6). (C) Metabolic heatmap displaying expression patterns of 11 glycolysis‐ and TCA cycle‐related metabolites across groups, with the Mac@WZB117 group showing an overall suppression trend. (D) Bar graphs showing changes of lactate, pyruvate and citrate. (E) Metabolic pathway enrichment plot generated using MetaboAnalyst showing significant enrichment of the ‘Glycolysis/Gluconeogenesis’ and ‘TCA Cycle’ pathways in the Mac@WZB117 group.

The metabolic heatmap further revealed that the expression levels of 11 core glycolytic and carbohydrate‐metabolism molecules were broadly reduced in the intervention group, reflecting a wide‐ranging metabolic suppression effect of Mac@WZB117 (Figure 15C). Quantitative analysis showed that lactate, pyruvate and citrate concentrations were all significantly lower in the Mac@WZB117‐treated group compared with the PAH + NS group (Figure 15D). Metabolic pathway interaction network analysis indicated that Mac@WZB117 treatment primarily affected energy and carbohydrate metabolism pathways. The most significantly enriched pathways included the TCA cycle, pyruvate metabolism, glycolysis/gluconeogenesis, starch and sucrose metabolism, and alanine, aspartate, and glutamate metabolism (Figure 15E).

These data collectively suggest that Mac@WZB117 effectively reduces the glycolysis dependency of immune cells by systematically inhibiting the sugar metabolism network.

3.17. Rescue Experiment Indicates Glycolysis Intervention as a Core Mechanism for Treatment

To ascertain the pivotal role of glycolysis inhibition in the efficacy of Mac@WZB117 treatment for PAH, a rescue metabolic activation experiment was designed. Building on the Mac@WZB117 intervention group, 2‐DG (a glucose analogue) was administered intraperitoneally, or a GLUT1 agonist (such as the inhibitory antagonist BAY‐876) was applied. The effects on pulmonary tissue immune phenotype, haemodynamics and tissue remodelling were observed.

At the immune‐cell level, flow cytometry revealed that in the rescue group the proportion of F4/80⁺CD86⁺ M1 macrophages decreased from 28.7% in the Mac@WZB117‐only group to 20.3% (p < 0.01), while the proportion of F4/80⁺CD206⁺ M2 macrophages increased correspondingly (Figure 16A). Consistently, qPCR results showed upregulation of the M1 marker gene iNOS and downregulation of the M2 marker genes Arg1 and Mrc1, indicating that glycolysis reactivation reversed the macrophage phenotype reprogramming induced by Mac@WZB117 (Figure 16B).

FIGURE 16.

Glycolytic reactivation partially attenuates the therapeutic effects of Mac@WZB117. (A) Changes in macrophage phenotype proportions, FACS results showed an increased proportion of M2 macrophages and a decreased proportion of M1 macrophages in the rescue group, indicating reversal of macrophage polarization. (B) qPCR data showed iNOS expression was upregulated, while ARG1 and Mrc1 expression was downregulated, confirming that phenotype remodelling depends on glycolysis inhibition. (C) Haemodynamic parameters (RVSP and mPAP) were significantly elevated after rescue treatment, approaching levels observed in the PAH model. (D) Representative histological staining images and quantitative analysis showed that vascular WT% and α‐SMA‐positive area were markedly increased in the rescue group, suggesting partial reversal of pulmonary vascular structural remodelling. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Functionally, the RVSP in the rescue group mice increased from 27.4 ± 2.0 mmHg in the Mac@WZB117 group to 37.5 ± 2.2 mmHg (p < 0.001), and mPAP also significantly rebounded, suggesting interference with the relief of pulmonary arterial pressure (Figure 16C). Additionally, H&E and Masson staining results showed that the arterial WT% increased from 17.6% in the intervention group to 26.8%, and the α‐SMA‐positive area significantly increased, approaching the median range of the un‐intervened model level, indicating partial reversal of tissue structural remodelling effects (Figure 16D).

In summary, the results of this rescue experiment demonstrate that the glycolysis inhibition pathway plays a potential supporting role in the effects of Mac@WZB117 intervention.

3.18. Mac@WZB117 Treatment Does Not Induce Significant Systemic Toxicity

Building on the validation of the therapeutic efficacy of Mac@WZB117, we further conducted a comprehensive assessment of its in vivo safety to confirm its favourable biocompatibility and potential for clinical translation. Using serum biochemical assays, histological staining and organ toxicity scoring, we evaluated the impact of this nanomedicine on the function and structure of major organs (Figure 17A).

FIGURE 17.

Mac@WZB117 treatment does not elicit significant systemic toxicity. (A) Experimental workflow illustrating the evaluation of Mac@WZB117 for systemic toxicity in major organs of PAH model mice. (B) Serum biochemical parameters (ALT, AST, BUN, Cr) showed no significant elevation in the Mac@WZB117 group, indicating preserved hepatic and renal function. (C) H&E staining images of the heart, liver, kidney and lung revealed intact tissue architecture with no pathological alterations in the Mac@WZB117 group. (D) Histopathological scoring demonstrated that all organ scores in the Mac@WZB117 group were below 1, markedly better than those in the free drug group. (E) Multi‐parameter radar plot summarising the distribution of toxicity indicators, showing that Mac@WZB117 exhibited a favourable safety profile. Data are presented as mean ± SD. Statistical significance is indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Serum biochemical analysis revealed that the levels of ALT and AST in mice treated with Mac@WZB117 were 36.1 ± 4.7 IU/L and 72.4 ± 6.2 IU/L, respectively, while BUN and Cr levels were 14.7 ± 1.3 mg/dL and 0.34 ± 0.04 mg/dL, all within the normal physiological range and not significantly different from the Sham group (p > 0.05, Figure 17B). In contrast, individual AST levels were elevated in the free WZB117 group, indicating a potential hepatic metabolic burden from the unencapsulated drug form.

H&E staining confirmed the structural integrity of organs. The major organs—including the heart, liver, kidneys and lungs—exhibited intact structures, and tissue sections showed no signs of inflammatory cell infiltration, cellular necrosis or architectural disruption; hepatic cell cords were orderly, renal tubular structures were distinct, cardiac tissue was continuous and alveolar walls were intact without thickening (Figure 17C). The lung and kidney tissues displayed normal airspace structure and glomerular morphology, with no toxicity‐related morphological abnormalities observed.

In terms of histopathological scoring (evaluating pathological features such as inflammation, haemorrhage and degeneration on a scale of 0–5), the Mac@WZB117 treated group scored below 1 for all organs, significantly lower than the free WZB117 group (Figure 17D), with differences being statistically significant (p < 0.01). Furthermore, no significant covariant correlations were found among cardiac, hepatic and renal function indices, further affirming the absence of subclinical toxic burden. A multi‐parameter radar chart was constructed to comprehensively display the horizontal comparison of toxicity‐related indicators, highlighting the low toxicity profile of Mac@WZB117 across five organ dimensions (Figure 17E).

In summary, Mac@WZB117 significantly mitigates PAH symptoms without causing hepatic, renal, pulmonary or cardiac dysfunction, maintaining normal tissue morphology and functional parameters.

4. Discussion

Targeting the metabolic vulnerabilities of pathological cells has emerged as a transformative strategy in cardiovascular medicine, yet its application in PAH remains largely confined to vascular resident cells (Pi et al. 2024; Cuthbertson et al. 2023). Unlike previous studies that primarily focused on the metabolic states of endothelial cells and smooth muscle cells (Lu et al. 2022; Gomes et al. 2022; Cuthbertson et al. 2023), this study identified the glycolytic pathway of infiltrating M1 macrophages as a critical, yet actionable, driver of PAH progression. Here, we developed and validated a macrophage membrane‐camouflaged nanoplatform, Mac@WZB117, which integrates precise disease targeting with potent immunometabolic modulation. By specifically intercepting the glycolytic supply line of pro‐inflammatory macrophages in the pulmonary vascular microenvironment, our findings provide a novel proof‐of‐concept for intervening in the ‘immune‐metabolic axis’ to reverse maladaptive vascular remodelling.

A major bottleneck in metabolic therapy is the lack of specificity, as systemic metabolic inhibitors, such as 2‐DG, often indiscriminately affect healthy tissues, leading to dose‐limiting toxicities (Laussel and Léon 2020; Jin et al. 2019). Our biomimetic strategy overcomes this barrier by leveraging the natural homing capability of macrophage membranes. Mac@WZB117 demonstrated superior accumulation in pulmonary lesions and preferential uptake by activated M1 macrophages, thereby enhancing the therapeutic index of the GLUT1 inhibitor WZB117 while minimizing off‐target effects. This ‘Trojan horse’ approach not only resolves the pharmacokinetic limitations of small‐molecule metabolic modulators but also highlights the potential of synthetic biology‐driven nanomedicine in managing complex vascular diseases.

Mechanistically, our multi‐omics analyses revealed that Mac@WZB117 orchestrates a comprehensive reprogramming of the macrophage landscape. The study demonstrated that Mac@WZB117 not only suppresses the glycolytic activity and pro‐inflammatory cytokine secretion of M1 macrophages but also promotes their transition to an anti‐inflammatory/repair‐oriented M2‐like phenotype, thereby alleviating perivascular inflammation and fibrosis. Crucially, spatial transcriptomics unveiled a novel dimension of this regulation: the ‘spatial reorganization’ of immune cells. We observed that GLUT1/CD86‐positive macrophages, originally clustered around remodelled arterioles, were dispersed following treatment, potentially migrating to less inflamed alveolar regions. Concurrently, the downregulation of pro‐fibrotic modules (e.g., Tgfb1, Col1a1) and lactate metabolism genes (Ldha) within lesion niches suggests that metabolic intervention successfully disrupted the local ‘pro‐fibrotic loop’. This evidence strongly supports the hypothesis that correcting immunometabolic dysregulation can dismantle the cellular and molecular foundations of vascular remodelling.

Compared with previous in vitro macrophage metabolic models (Liu and Ho 2018), this study extended validation to the animal level by employing spatial transcriptomics and scRNA‐seq to demonstrate the spatial, lineage and functional reprogramming of the macrophage metabolism‐signalling axis. GLUT1/CD86 double‐positive M1 macrophages were highly clustered around diseased small arteries, whereas Mac@WZB117 treatment induced their dispersal in a manner that suggests a potential migration to less inflammatory regions, such as the alveolar septa. This spatial reorganization implies the platform's capacity to correct pathological immune cell aggregation, highlighting its strong potential for integrated microenvironmental regulation.

Notably, within the spatial co‐expression modules of the lesion regions, Mac@WZB117 markedly suppressed collagen production (Col1a1), lactate metabolism (Ldha) and TGF‐β signalling activity, further highlighting its important role in shaping an ‘inflammation‐suppressed, low‐metabolic’ microenvironment. This mechanism—interrupting the pathogenic cascade at its ‘metabolic stress’ source—differs from traditional anti‐fibrotic or anti‐remodelling therapies and may offer a new therapeutic window for refractory PAH. While spatial transcriptomics provides a static snapshot, the observed downregulation of remodelling‐associated genes in specific lesion areas strongly supports the hypothesis that metabolic intervention can disrupt the local pro‐fibrotic niche.

Beyond the specific therapeutic findings, this study highlights the value of an integrative analytical framework that couples cellular metabolism, lineage plasticity and spatial topology to evaluate immunometabolic interventions. Rather than relying on single‐dimensional metrics (Hong et al. 2021), this multi‐scale approach allows for a comprehensive dissection of how metabolic modulation translates into tissue‐level remodelling. This methodological paradigm offers a reference for exploring other pathologies driven by immune‐metabolic imbalances, such as autoimmune disorders and organ fibrosis, providing a robust strategy to decode the complex interplay between metabolic reprogramming and disease progression.

In summary, Mac@WZB117 represents a precision nanomedicine strategy designed to dismantle the ‘metabolic checkpoint’ of pathogenic M1 macrophages in PAH. By selectively intercepting GLUT1‐dependent glycolytic flux, this platform deprives pro‐inflammatory cells of their energy source, thereby enforcing a phenotypic shift toward repair and disrupting the paracrine signalling loops that drive vascular remodelling. This work not only validates the ‘glycolysis‐inflammation‐remodelling’ axis as a viable therapeutic target but also demonstrates the potential of biomimetic nanotechnology to overcome the delivery barriers of metabolic inhibitors. However, bridging the gap to clinical application requires addressing limitations such as the long‐term immunogenicity of biomimetic carriers and the heterogeneity of human PAH pathology. Future research should prioritize validating these mechanistic insights in patient‐derived tissues and optimizing the carrier design to ensure safety and efficacy for clinical translation.

Author Contributions

Y.G. and W.W. conceived and designed the study. Y.G., S. L. and Y.H. performed the experiments. S.L., Y.H. and W. W. analyzed the data. Y.G. and S.L. wrote the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

This study was supported by South Sea New Star Science and Technology Innovation Project of Hainan Provincial of China (NHXXKJCX202514), Youth Fund Project of Natural Science Foundation of Hainan Province of China (826QN0881)and the National Natural Science Foundation of China (No. 82460538), South Sea New Star Project of Hainan Provincial of China (NHXX‐WJW‐2025011), Hainan Provincial Natural Science Foundation of China (326MS0414).

Ethics Statement

Animal experiments adhered to the International Guidelines for Animal Experimentation and ARRIVE 2.0 reporting standards. The study design, animal welfare considerations and endpoint management received approval from the Animal Ethics Committee of Hainan Medical University (Ethics Approval No.: HYLL‐2024‐759).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Material: jev270294‐sup‐0001‐SuppMat.docx

Data Availability Statement

All data generated or analyzed during this study are included in this article and/or its supplementary material files. Further enquiries can be directed to the corresponding author.