Design of RGD-functionalized GSH-responsive pegylated polymeric protacs for selective BRD4 degradation and EndMT-driven cardiac fibrosis inhibition

Abstract

Endothelial cells (ECs) of endothelial-to-mesenchymal transition (EndMT) are drivers of cardiac fibrosis. BRD4 has recently been identified as an epigenetic regulator of EndMT. Proteolysis-targeting chimera (PROTAC) technology has revolutionized targeted protein degradation, offering unprecedented opportunities for BRD4 modulation in diverse pathological contexts. Nevertheless, the non-selective cellular targeting profile of PROTACs poses significant limitations for their therapeutic application in cardiac fibrosis management. To address these limitations, we developed a GSH-responsive nanoscale PROTAC (RGD-PEG-MZ1) that targets activated platelets, leveraging their chemotactic properties to precisely degrade BRD4 in ECs. RGD-PEG-MZ1 exhibits selectivity for ECs and inhibition of EndMT, which can prevent the progression of cardiac fibrosis. The RNA-seq analysis revealed an attenuation of the MAPK signaling pathway following RGD-PEG-MZ1 treatment. The interaction between BRD4 and the MAPK signaling was analyzed through AlphaFold3 and immunoprecipitation assays. The experimental data showed that BRD4 directly interacts with RAF1, a critical effector in MAPK signaling, which suggested that RGD-PEG-MZ1 modulates MAPK signaling by disrupting the BRD4-RAF1 interaction. This innovative GSH-activated PROTAC strategy not only offers a novel therapeutic approach for cardiac fibrosis but also provides insights into the functional role of BRD4 in the disease pathogenesis of cardiac fibrosis.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04036-7.

Introduction

Cardiac fibrosis is a process characterized by inadequate response to chronic tissue injure or stress, and is a common feature of many types of cardiac disease, which potentially distorts architecture of the myocardium, and alters the biological function of extracellular matrix (ECM) [1–6]. The pathological substitution of myocardial tissue with fibrotic deposits is recognized as a pivotal lesion contributing to the pathogenesis of heart failure [7]. Cardiac fibroblasts represent the primary cellular drivers of cardiac fibrosis development [7, 8]. Preclinical studies have demonstrated that pharmacological interventions for cardiac fibroblasts not only mitigate and attenuate cardiac fibrosis, may even induce its regression in certain cases [1]. Furthermore, although conventional therapeutic approaches [9], have been shown to confer partial efficacy in reducing cardiac fibrosis in humans, fibrosis may persist in patients with specific comorbidities despite adherence to these standard treatments [10, 11]. This persistence underscores the imperative for developing more efficacious antifibrotic strategies to enhance therapeutic outcomes.

Selective modulation of endothelial cell (EC) function has been extensively utilized in the therapeutic management of various diseases, emerging as a key strategy in contemporary treatment paradigms [12, 13]. EC dysfunction is thought to be involved in a series of pathological processes, such as diabetes, wound healing and tumor growth [14–18]. Furthermore, endothelial dysfunction could promote cardiovascular disease, including hypertension, coronary artery disease, chronic heart failure and peripheral artery disease through Endothelial-Mesenchymal Transition (EndMT) [19–21]. EndMT refers to the process in which ECs lose their original characteristics and transform into mesenchymal cells (myofibroblasts, smooth muscle cells, etc.) under the action of various stimulation, resulting in significant changes in the polarity, morphology, and function of ECs [22, 23]. ECs from different organs exhibit varying propensities to undergo EndMT, and play a vital role in the pathological process of fibrosis in almost all organs, such as liver, kidney, pulmonary and heart [24–26]. In addition to producing a permanent mesenchymal phenotype, ECs can also remain in an intermediate stage of transdifferentiation for a long time, a process called partial EndMT [27–29]. It should be noted that endothelial cells in partial EndMT can be reversed under certain conditions, making it a potential target for the treatment of cardiac fibrosis.

Bromodomain-containing protein 4 (BRD4), an integral component of the BET family of proteins, serve as epigenetic readers of histone acetylation, which recruit transcriptional regulator complexes to chromatin and bind to acetylated histones [30–32]. BRD4, which possesses epigenetic reader functions and the ability to recruit transcriptional complexes, is involved in the regulation of genes associated with inflammation and profibrotic responses during fibrosis [33]. During cardiac fibrosis, BRD4 becomes significantly enriched at fibrosis-associated promoters, active enhancers, and super-enhancer subgroups under the influence of fibrosis-inducing factors, thereby directly driving the expression of fibrosis-related genes [34]. Through interaction with P-TEFb, BRD4 not only facilitates the expression of cardiac pathological biomarkers but also promotes transcription elongation and the release of RNA polymerase II [35]. Moreover, in response to TNF-α stimulation, activated NF-κB rapidly translocates to the nucleus, where it recruits BRD4 to enhancer and promoter regions within the genome [36]. The subsequent BRD4-NF-κB binding forms a super-enhancer region that, in a BRD4-dependent manner, drives the transcription of pro-inflammatory genes, thereby contributing to the progression of cardiac fibrosis [37]. While the role of BRD4 in modulating fibrosis-related gene expression and the inflammatory response during cardiac fibrosis is well-established, its specific involvement in EndMT remains incompletely understood. Therefore, elucidating the mechanistic role of BRD4 in EndMT regulation is crucial for advancing our understanding of its broader biological function in cardiac fibrosis.

BRD4 plays a critical regulatory role in the progression of cardiac fibrosis, the pharmacological targeting of BRD4 to ameliorate fibrosis-related pathology has been extensively validated in various studies [38–40]. The small molecule inhibitor C-34, targeting BRD4, exerts inhibitory effects on the TGF-β/Smad signaling cascade, attenuates the proliferation and migration of neonatal rat cardiac fibroblasts stimulated by angiotensin II (Ang II), curbs the synthesis of extracellular matrix (ECM) proteins in the context of fibrosis, and ameliorate symptoms associated with cardiac fibrosis [41]. The BRD4 inhibitor JQ1 has been demonstrated to reverse the EndMT in ECs to some extent, both in vivo and in vitro models, effectively blocking the synthesis of ECM proteins, and consequently mitigating the symptoms of cardiac fibrosis [42]. These investigations highlight the therapeutic potential of pharmacologically targeting BRD4 protein as a powerful strategy against cardiac fibrosis. However, traditional therapeutic strategies targeting BRD4 using kinase inhibitors have demonstrated limited efficacy, primarily due to their low efficiency in suppression of BRD4-mediated transcriptional regulation.

The advent of Proteolysis Targeted Chimeras (PROTACs) has revolutionized the paradigm of drug development, introducing a groundbreaking approach to selectively degrade proteins associated with diseases [43–45]. PROTACs are bifunctional molecules capable of simultaneously recruiting both E3 ubiquitin ligase (E3) and proteins of interest (POIs) [6, 46]. Formation of the POI-PROTAC-E3 ternary complex can trigger POI ubiquitination, thereby facilitating the targeted degradation of POI with unprecedented precision [47, 48]. This mechanism enables PROTACs to overcome the inherent limitations of traditional small molecule inhibitors by degrading the entire target protein with a catalytic amount, exemplified by their superior efficacy in modulating target proteins [47, 49]. However, a series of problems, such as systemic toxicity and low solubility usual associate with PROTAC [50]. Furthermore, the indiscriminate degradation of the protein by PROTACs in different cell subpopulations may attenuate the therapeutic effect and, in some case, exacerbate the disease progress. Therefore, the development of effective protein degradation strategies for cell subpopulations under pathological conditions, aimed at overcome the potential adverse effects of PROTACs, represents a formidable challenge.

Recently, a series of PROTAC-based prodrugs were proposed by us and other groups to achieve precise treatment of PROTAC [6, 51–56]. In this study, we developed a new options to improve the action of PROTACs in cardiac fibrosis, focusing on the design of in vivo controlled delivery nanoparticles and providing a novel mechanistic explanation. As illustrated in Scheme 1, RGD-PEG-MZ1 consists of cRGD modified polyethylene glycol (cRGD-PEG) as the hydrophilic and ECs targeting part, an GSH-responsive group for rapid response to GSH, and hydrophobic PROTACs targeting BRD4. The amphiphilic nature of RGD-PEG-MZ1 enabled them to self-assemble into monocomponent nanoparticles in an aqueous solution. After intraperitoneal injection, the RGD-PEG-MZ1 nanoparticles, through its integrin-binding RGD motif, selectively binds to activated platelets within the bloodstream [57, 58]. Subsequently, the activated platelets would precisely navigate RGD-PEG-MZ1 to the vicinity of ECs. Once RGD-PEG-MZ1 are actively internalized into ECs, they can be effectively activated by overexpressed GSH in the ECs undergoing EndMT, resulting in rapid release of the PROTACs molecule MZ1. Then, the resulting PROTACs MZ1 could selectively degrade BRD4 in ECs, which in turn downregulates a series of fibrosis biomarkers by inhibits EndMT.

Scheme 1.

Schematic illustration of GSH activated RGD-PEG-MZ1 for the treatment of cardiac fibrosis by suppress MAPK signaling pathway

In this study, our synthesized RGD-PEG-MZ1 effectively degrades BRD4 protein, thereby exerting therapeutic effects in the prevention of cardiac fibrosis. However, the precise molecular mechanisms underlying BRD4-mediated regulation of EndMT during cardiac fibrosis remain incompletely elucidated. To address this gap, we performed RNA-seq analysis to assess the impact of RGD-PEG-MZ1 on key regulatory nodes within the EndMT-associated signaling pathways, thereby providing further insights into the role of BRD4 in this pathological process. Our experimental results revealed that RGD-PEG-MZ1 potently attenuated the activation of the MAPK signaling cascade, a pathway in the EndMT process [59]. Although we have confirmed that the degradation of BRD4 inhibits the MAPK signaling pathway, the precise molecular mechanism by which BRD4 regulates this pathway remains incompletely understood. To further elucidate this mechanism, we employed AlphaFold3 and immunoprecipitation assays to assess the interaction between BRD4 and Raf1, a critical transducer in the MAPK signaling cascade. The results of these experiments demonstrate a stable interaction between BRD4 and Raf1, suggesting that the inhibitory effect of RGD-PEG-MZ1 on MAPK signaling is likely mediated through the disruption of the BRD4-Raf1 complex. Consequently, this study not only expands the understanding of the molecular mechanisms underlying EndMT but also provides a novel therapeutic approach for cardiac fibrosis, offering valuable insights into the development of precise therapies for cardiovascular diseases.

Experimental section

Chemicals, reagents and antibodies

Angiotensin II (HY-13948), isoprenaline hydrochloride (HY-B0468), sorafenib (HY-10201), JQ-1 (HY-13030), Adenosine 5’-diphosphate (HY-W010918) and MZ1 (HY-107425) were all procured from MedChemExpress (MCE). Additionally, 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt (DiD’ solid, C1039), and 4′,6-diamidino-2-phenylindole (DAPI, C1002) were sourced from Beyotime (Table 1).

Table 1.

Antibodies used in this study

Name	Citation (PMID)	Supplier	Cat no.	Clone no.
FITC Rat Anti-Mouse CD62P	8,562,500	BD Pharmingen	561,923	RB40.34
Anti-Brd4	36,720,918	Abcam	ab128874	EPR5150(2)
Anti-GAPDH	31,666,698	Proteintech	60004-1-Ig	1E6D9
Anti-α-SMA	37,962,473	Proteintech	14395-1-AP	Polyclonal
Anti-Fibronectin	36,480,284	Proteintech	15613-1-AP	Polyclonal
Anti-Collagen I	37,901,206	Abcam	ab138492	EPR7785
Anti-FSP 1	38,198,846	Proteintech	20886-1-AP	Polyclonal
Anti-VE-Cadherin	39,129,277	CST	2500T	D87F2
Anti-CD31	38,295,159	Proteintech	66065-2-Ig	3F8E2
Anti-MYC	37,433,812	Proteintech	60003-2-Ig	1A5A2
Anti-mCherry	38,413,626	Proteintech	26765-1-AP	Polyclonal
Anti-Raf1	36,759,880	Abcam	ab181115	EP4969
Anti-Raf1 (phospho S259)	38,020,551	Abcam	ab173539	EPR3433(2)
Anti-ERK1 + ERK2	38,124,661	Abcam	ab184699	EPR17526
Anti-ERK1 (phospho T202 + Y204) + ERK2 (phospho T185 + Y187)	37,798,786	Abcam	ab278538	ERK12T202Y204-A11
Anti-ELK1	35,614,034	Abcam	ab32106	E277
Anti-ELK1 (phospho S383)	31,602,408	Abcam	ab218133	Polyclonal
Goat Anti-Rabbit IgG (Alexa Fluor® 488)	38,124,020	Abcam	ab150081	Polyclonal
Goat Anti-Mouse IgG (Alexa Fluor® 555)	37,555,936	Abcam	ab150114	Polyclonal
HRP-conjugated Goat Anti-Rabbit IgG(H + L)	37,286,607	Proteintech	SA00001-2	Polyclonal
HRP-conjugated Goat Anti-Mouse IgG(H + L)	37,098,345	Proteintech	SA00001-1	Polyclonal

Preparation of RGD-PEG-MZ1 nanoparticles

A solution of 50 µL RGD-PEG-MZ1 in DMSO (50 mg/mL) was added to 3 mL deionized water while stirred at 1500 rpm. After stirred at room temperature for 10 min, the above mixture was dialyzed with deionized water for 24 h (MWCO = 2000) to remove DMSO to obtain RGD-PEG-MZ1 nanoparticles.

Mice

All animal experimental procedures were conducted in strict accordance with the guidelines set forth by the National Institutes of Health and were duly approved by the Animal Ethics Committee at Southwest Medical University in Luzhou, China (Approval No. SWMU20240084). The study utilized male C57Bl/6j mice, which were sourced from GemPharmatech Co., Ltd (Chengdu, China). These mice were maintained and nourished under standard laboratory conditions, with a consistent 12-hour light/dark cycle, from 6 a.m. to 6 p.m., to ensure optimal living standards.

Establishment of isoprenaline animal models and treatment of JQ1, MZ1 and RGD-PEG-MZ1

As previously delineated, a murine model of ISO-induced cardiac fibrosis was meticulously established. The experimental cohort was subjected to daily subcutaneous administrations of isoproterenol at a concentration of 50 mg/kg for a continuous 14-day period [60]. Concurrently, the control cohort was treated with an equivalent volume of isotonic saline solution. A total of 30 rodents were systematically stratified into five discrete cohorts, each consisting of six individuals. The treatment paradigms administered to these cohorts were as delineated: isotonic saline solution as the control placebo, ISO as the pathogenic inducer, ISO in combination with 20 mg/kg of JQ1, ISO in conjunction with 20 mg/kg of the MZ1, and ISO co-administered with 20 mg/kg of RGD-PEG-MZ1 conjugate. The administration of JQ1, MZ1, and RGD-PEG-MZ1 was initiated on day 8 following the onset of ISO treatment. All compounds were delivered via intraperitoneal injection in accordance with a rigorously controlled and predefined dosing schedule. ISO injections were executed at 11:00 AM daily, whereas the adjunctive compounds JQ1, MZ1, and RGD-PEG-MZ1 were administered at 17:00 PM daily. Following the ultimate administration, the murine subjects were subjected to euthanasia on the succeeding day, and comprehensive cardiac tissue specimens were procured for subsequent experimental analyses.

Histological assessment of cardiac fibrosis

Cardiac tissues were postfixed in a 4% paraformaldehyde solution diluted in phosphate-buffered saline (PBS). Following postfixation, the tissues underwent a dehydration process utilizing a graduated ethanol series. The dehydrated cardiac tissues were then embedded in paraffin and sectioned to a uniform thickness of 4 μm. Prior to histopathological staining, the paraffin-embedded sections were subjected to a reverse ethanol gradient for rehydration. Standard protocols were employed for Hematoxylin and Eosin (H&E), Masson’s Trichrome, and Sirius Red staining. Quantitative assessment of cardiac fibrosis in the stained sections was conducted utilizing Image Pro Plus (IPP) version 6.0 software.

Immunofluorescence staining of cardiac tissue slides

For the immunofluorescence staining procedure, sections of frozen cardiac tissue, each 4 μm thick, were immersed in a solution containing 4% paraformaldehyde. The slides were initially stabilized with a 5% goat serum solution, then exposed to primary antibodies—α-SMA (Proteintech, 14395-1-AP), Collagen-I (Proteintech, 14695-1-AP), CD31 (Proteintech, 66065-2-Ig), Raf1 (Abcam, ab181115), Raf1 (phospho S259) (Abcam, ab173539), ERK1 + ERK2 (Abcam, ab184699) and ERK1 (phospho T202 + Y204) + ERK2 (phospho T185 + Y187) (Abcam, ab278538)—and subsequently incubated at 4 °C for an extended period. Following a rinse with PBS, the slides were incubated with a secondary antibody, Goat Anti-Rabbit IgG H&L (Abcam, ab150081) and Goat Anti-Mouse IgG (Abcam, ab150114), for a duration of 1 h. Ultimately, the sections were stained with DAPI, cleaned with PBS, and preserved in a solution containing 20% glycerol prior to imaging.

Cell culture and in vitro stimulation of EndMT

Human umbilical vein endothelial cells (HUVECs, RRID: CVCL_9Q53) and Human microvascular endothelial cell line (HMEC-1, RRID: CVCL_0307) were procured from iCell Bioscience Inc., located in Shanghai, China. These cells were cultured in a proprietary epithelial cell medium, designated as iCell-h110-001b, which is also provided by iCell Bioscience Inc. The HUVECs/HMEC-1 were maintained under controlled conditions at 37 °C with 5% CO₂ in a humidified incubator.

Optimize the proliferation of HUVECs/HMEC-1 to attain a 50% confluent monolayer. Subsequently, impose a 12 h period of serum deprivation in a medium enriched with 3% fetal bovine serum (FBS). Thereafter, subject HUVECs to a 48 h continuous exposure to Ang II at a concentration of 3.0 µM. At the 24 h juncture of Ang II administration, incorporate the RGD-PEG-MZ1 therapeutic agent exclusively into the intervention cohort as a pharmacological intervention.

Plasmid transfection

For the purpose of plasmid-mediated transfection, plasmids engineered for the overexpression of Raf1 (NCBI Gene ID: 5894, pcDNA 3.1) were procured from the Public Protein/Plasmid Library (PPL, Nanjing, China). pEnCMV-mCherry-Linker-BRD4 (human)-SV40-Neo (P30990) was obtained from MiaoLingBio (China). For the transfection of a single well within a six-well plate, a quantity of 3 µg of plasmid DNA and 0.75 µL of Xfect RNA Transfection Polymer (Catalog No. 631317, Takara) were aliquoted into 100 µL of Xfect Reaction Buffer, subsequently mixed thoroughly, and incubated at ambient temperature for a period of 10 min to facilitate the formation of a transfection complex. Prior to the transfection procedure, the culture medium was replaced with an antibiotic-free and serum-free DMEM. Following a 6 h incubation period, the aforementioned transfection complex was introduced to the cells. The cell culture dishes were then incubated at a controlled temperature of 37 °C for a duration of 12 h. Subsequently, the transfection complex was aspirated, and the cells were replenished with fresh culture medium to proceed with subsequent intervention treatments.

Co-immunoprecipitation

HUVECs were subjected to protein extraction using RIPA buffer. Post-lysis, the supernatant was obtained via centrifugation to achieve protein clarification. Subsequently, 1 µg of a specific primary antibody or its corresponding nonspecific isotype control immunoglobulin was introduced to a 600 µg aliquot of the total cellular protein, reconstituted in a final volume of 500 µL. This mixture was then incubated under conditions of slow agitation at 4 °C overnight to facilitate antigen-antibody complex formation. On the subsequent day, 20 µL of Protein A Agarose Beads (Beyotime, Catalog Number P2051, China) was incorporated into the mixture and allowed to rotate at 4 °C for an additional 2 h period to enhance the immunoprecipitation efficiency. The agarose beads, now enriched with the protein-antibody complexes, were pelleted by centrifugation and subjected to a series of four washes with 1 mL of RIPA buffer to remove unbound proteins. Following the final centrifugation, 30 µL of 1× Sodium Dodecyl Sulfate (SDS) loading buffer was added to the beads, and the mixture was heated to boiling point for 5 min to elute the bound proteins from the beads. This preparation was then utilized for subsequent Western blot analysis to assess the protein expression levels.

Protein-protein complex structure prediction with alphafold 3.1

The protein-protein complex structure prediction was performed using AlphaFold3 (AF3). The input FASTA sequences were derived from the Ras-Binding Domain of c-Raf-1 (PDB ID: 1RFA) and the Bromo Domain 1 of human Bromodomain Containing Protein 4 (BRD4, PDB ID: 2PSS). For the genetic searching step, we followed the official AF3 guidelines to ensure the optimal multiple sequence alignment (MSA) generation process. Detailed MSA parameters, including database and search configurations, are available at https://github.com/google-deepmind/alphafold3 (accessed on 2024/12/9). Here, we briefly describe the databases used for protein chain searches. Five databases were utilized: UniRef90, UniProt, Reduced BFD, MGnify, and Uniclust30 + BFD. The searches were performed using the Jackhmmer and HHBlits tools. To ensure reproducibility and robust predictions, five distinct random seeds were applied, generating a total of 25 structural models across five independent runs (five models per seed). After the model inference step, the highest-ranked model was selected for downstream analysis based on AlphaFold3’s well-calibrated confidence measures, which closely align with prediction accuracy.

The resulting protein-protein complex was thoroughly analyzed using PLIP (Protein-Ligand Interaction Profiler), with the Raf1 chain designated as the ligand chain to facilitate the identification of hydrogen bonds (H-bonds), hydrophobic interactions, and salt bridges. The structural visualization of the complex was carried out in PyMOL (Fig. 6), and a 2D interaction diagram highlighting the binding interface was generated using LigPlot⁺ (Fig. 6). This systematic workflow integrated advanced computational modeling with detailed interaction profiling, ensuring a rigorous and comprehensive evaluation of the predicted complex for structural validation and interpretation.

Fig. 6.

RGD-PEG-MZ1 exhibits a salutary therapeutic efficacy of cardiac fibrosis. (A) RGD-PEG-MZ1 attenuated myocardial damage and fibrosis after ISO-induced myocardial injury. H&E, Sirius Red, Masson and immunofluorescence staining were conducted on mousemyocardial sections (n = 6 mice/group). Scale bar, 100 μm. (B-C) α -SMA, Fibronectin, Collagen-Ⅰ and FSP1 levels were assessed via western blot following exposure to RGD-PEG-MZ1 (n = 3). The data are presented as the Mean ± SD of three independent experiments. ^#p < 0.05 and ^##p < 0.01 vs. Sham group; *p < 0.05 and **p < 0.01 vs. ISO group (one-way ANOVA followed by Bonferroni post hoc test)

Western blot analysis

For the purpose of cellular lysis, cells were initially rinsed with ice-cold PBS to remove extracellular debris. Subsequently, cells were subjected to lysis in an ice-cold lysis buffer supplemented with 1% protease and phosphatase inhibitors to prevent enzymatic degradation of proteins. The resulting lysates were subjected to vigorous vortexing to ensure thorough cell disruption and then clarified by centrifugation at 11,000 rpm for 15 min at 4 °C to sediment insoluble material. The concentration of the extracted proteins was quantified using the Bradford assay, as provided by Solarbio. The protein samples were then resolved by electrophoresis through a 4–12% Bis-Tris SurePAGE gel, which was sourced from Shanghai Epizyme Biomedical Technology Co., Ltd. Following electrophoresis, the proteins were transferred onto a 0.25 μm PVDF membrane, procured from Merck Millipore, to facilitate blot analysis. The PVDF membrane was pre-incubated with QuickBlock Western blocking buffer, as supplied by Beyotime, for 20 min at ambient temperature to minimize non-specific antibody binding. The blots were then incubated with the primary antibody overnight at 4 °C with slow rotation to allow for optimal antigen-antibody interaction. On the subsequent day, blots were further incubated with goat HRP-conjugated secondary antibodies for 1 h at room temperature with rotation to enhance signal detection. After a series of washes with TBST to remove unbound antibodies, the protein bands were visualized using a ChemiScope 6200 Touch imaging system. The band intensities were quantified using ImageJ software, with data normalization relative to GAPDH as a loading control and the control group within each experiment. The reported data represent the mean values obtained from three independent biological replicates. The quantitative analysis of Western blot was performed as described previously [61, 62].

Statistical analysis

Statistical analyses were performed using GraphPad Prism 9 software. Data normality and variance homogeneity were assumed prior to analysis. Experimental results are presented as the mean ± standard deviation (mean ± SD), with all experiments independently repeated at least three times. Differences between two independent groups were assessed using a two-tailed Student’s t-test. For comparisons involving more than two groups, one-way analysis of variance (ANOVA) was applied, followed by appropriate post hoc multiple-comparison tests to identify statistically significant differences. In addition, two-way ANOVA was employed to evaluate the effects of two independent variables and their potential interactions. For two-group comparisons, two-tailed unpaired Student’s t-test was applied. ‘ns’ indicates p > 0.05. Symbol definitions are provided in each figure legend (e.g., * vs. model group; # vs. control group) [63].

Results and discussion

Design and synthesis of RGD-PEG-MZ1 nanoparticles

The structure and the synthetic route for RGD-PEG-MZ1 is outlined in Scheme1. As shown in Fig. 1A, RGD-PEG-MZ1 consists of a cRGD fragment that targets ECs [58], a PEG2000 chain that increases water solubility [64], a GSH response fragment [63], and a PROTAC molecule MZ1 that degrades BRD4 [56]. The synthetic route for RGD-PEG-MZ1 is depicted in Fig. 1B. The chemical structure of the key intermediate was thoroughly characterized by proton nuclear magnetic resonance (¹H NMR), carbon-13 nuclear magnetic resonance (¹³C NMR) (Fig. S1-S5), and high-resolution mass spectrometry (HR-MS).

Fig. 1.

Characterization and activation study of RGD-PEG-MZ1.(A) The structure of RGD-PEG-MZ1. (B) synthesis of RGD-PEG-MZ1. (a) triethylamine, DMAP, 1,2-Dichloroethane; TBAF, THF. (b) 4-nitrophenyl carbonochloridate, DCM. (c) triethylamine, DMF. (C) Observation of the Tyndall effect for RGD-PEG-MZ1. (D) Transmission electron microscopy (TEM) images of RGD-PEG-MZ1, illustrating its morphology. (E) Particle size distribution and polydispersity index (PDI) analysis of RGD-PEG-MZ1. (F) Highperformance liquid chromatography (HPLC) profiles of S2 (red line, 5.0 µM) incubated with GSH (1 mM) for 4 h (black line)

Characterization and in vitro release study of self-assembly RGD-PEG-MZ1

The self-assembly properties of RGD-PEG-MZ1 were characterized using dynamic light scattering (DLS) and transmission electron microscopy (TEM). As shown in Fig. 1C, RGD-PEG-MZ1 demonstrated obvious Tyndall effect in aqueous solutions. TEM analysis revealed that RGD-PEG-MZ1 assembled into monodispersed and uniformly spherical nanoparticles (Fig. 1D). The hydrodynamic diameter of RGD-PEG-MZ1 was determined to be 136 nm with a narrow size distribution, as indicated by a polydispersity index (PDI) of 0.06 (Fig. 1E). The RGD-PEG-MZ1 nanoparticles was negatively charged with a Zeta potential of −18.8 mv. The critical micelle concentration (CMC) of RGD-PEG-MZ1 was determined using a fluorescence probe scattering assay, yielding a CMC value of 0.069 µg/mL (Fig. S6A). This exceptionally low CMC indicates that RGD-PEG-MZ1 can readily undergo spontaneous self-assembly and form stable nanostructures even under highly dilute conditions, underscoring its excellent amphiphilic organization and assembly efficiency. Subsequently, we performed a comprehensive characterization of its photophysical properties (Fig. S6B-D). Fluorescence spectroscopy revealed a maximum excitation wavelength at approximately 350 nm and a maximum emission wavelength around 400 nm, corresponding to a Stokes shift of ~ 50 nm. This pronounced shift confirms the presence of a distinct and readily traceable fluorescence signature intrinsic to the material (Fig. S6B-D). Furthermore, to assess the colloidal stability and long-term structural integrity of RGD-PEG-MZ1, we monitored temporal variations in hydrodynamic diameter and zeta potential under storage and incubation conditions (Fig. S6E-F). Over a continuous 15-day evaluation period, the zeta potential consistently remained within the range of − 20 to − 30 mV (Fig. S6E), indicative of substantial electrostatic repulsion that effectively mitigates nanoparticle aggregation. Concurrently, the hydrodynamic diameter remained stable at approximately 140 nm without detectable drift or an increase in polydispersity (Fig. S6F). Collectively, these findings demonstrate that RGD-PEG-MZ1 possesses robust colloidal stability and structural reliability, thereby providing a solid physicochemical foundation for its application as a nano-delivery platform.

To further elucidate the release behavior of RGD-PEG-MZ1 under physiologically relevant glutathione (GSH) concentrations, we first examined its GSH-triggered morphological evolution using TEM. As shown in Fig. S7A, exposure to GSH resulted in a pronounced disintegration of the self-assembled RGD-PEG-MZ1 nanoparticles in the aqueous phase, with the nanostructures exhibiting clear features of structural collapse. This marked morphological alteration indicates that GSH-induced destabilization of the nanoscale architecture facilitates the efficient liberation of the encapsulated payload. These observations provide direct morphological evidence supporting the subsequent quantitative investigations of the GSH-responsive release kinetics. The activation of RGD-PEG-MZ1 in the presence of GSH was investigated using S2 as the model compound (Fig. 1F). The retention times of S2 was determined to be 11.6 min in high-performance liquid chromatography. After incubating S2 with excessive GSH for 4 h, the peak corresponding to the S2 disappeared, and the peak corresponding to MZ1 ([M + H]⁺ = 1016.4) at 11.2 min appeared. To more precisely elucidate the GSH-mediated cleavage of S2 and the subsequent release of MZ1, we performed a systematic and rigorous quantitative analysis of its release kinetics. The results revealed that under a pathologically relevant GSH concentration (2.5 mM), S2 exhibited a sustained and markedly time-dependent liberation of MZ1. Notably, by 180 min, MZ1 release approached completion, with a cumulative release exceeding 85%, indicating that the system can undergo rapid and highly efficient activation within a reductive pathological microenvironment (Fig. S7B). Furthermore, we examined the modulatory influence of varying GSH concentrations on the release behavior of S2. The data demonstrated a clear GSH concentration–dependent release profile, wherein increasing GSH levels progressively enhanced the proportion of MZ1 released (Fig. S7B). This well-defined trend strongly supports the microenvironment-responsive prodrug mechanism underlying our system and indicates that the nanoplatform remains structurally stable under physiological GSH conditions, but undergoes rapid depolymerization and payload liberation in the highly reductive milieu characteristic of EndMT-associated endothelial cells. Taken together, these quantitative release profiles unequivocally demonstrate that MZ1 liberation from RGD-PEG-MZ1 is jointly governed by temporal kinetics and the local GSH concentration. Such dual responsiveness is entirely consistent with our rationally designed pathological microenvironment–triggered prodrug strategy and provides robust experimental evidence supporting the targeted activation mechanism of this nanodelivery system.

RGD-PEG-MZ1 targets pathological endothelial cells via platelet-mediated delivery

The activated platelets within the bloodstream exhibit a propensity to adhere to damaged vascular locales via their reciprocal interactions with impaired endothelial cells (ECs) in the context of cardiac fibrosis [65–67]. RGD-recognizing integrins on activated platelets [58, 68, 69]. To investigate the ability of RGD-PEG-MZ1 to target ECs through platelet-mediated delivery, we conducted an assessment of RGD-PEG-MZ1’s homing to injured vascular locales and its specificity for discrete subsets of ECs at the site of injury, employing both in vitro and in vivo methodologies. Initially, we discerned that RGD-PEG-MZ1 was avidly internalized by activated platelets in vitro, a finding that underscored its robust affinity for platelets in their activated state (Fig. 2A). This observation established a foundational premise for the subsequent in vivo targeting of ECs within cardiac lesions by RGD-PEG-MZ1, harnessing the tropism of platelets. Prior to assessing the in vivo targeting efficacy of RGD-PEG-MZ1 on ECs, we employed isoproterenol (ISO) to generate a murine model that recapitulates the pathophysiological features of cardiac fibrosis. The dosing regimen and administration protocol for ISO were aligned with those delineated in preceding scientific inquiries [60]. Following the successful establishment of the cardiac fibrosis model, 200 µL of a DiD fluorescent dye was utilized to conjugate with RGD-PEG-MZ1 for fluorescent labeling. Utilizing a small animal live imaging apparatus, we monitored the progressive accumulation of RGD-PEG-MZ1 within the murine cardiac tissue. Relative to control mice, the enrichment of RGD-PEG-MZ1 was markedly augmented in the cardiac fibrosis cohort, attesting to the enhanced targeting efficacy of RGD-PEG-MZ1 towards the myocardial lesion sites (Fig. 2B and Fig. S8). To elucidate the capacity of RGD-PEG-MZ1 to home to ECs via platelet bridging, we conducted immunofluorescence assays on murine cardiac tissues. The immunofluorescence data revealed a significant degree of colocalization between RGD-PEG-MZ1 and the endothelial marker CD31, thereby demonstrating that RGD-PEG-MZ1 is capable of being directed to ECs through platelet-mediated interactions in vivo (Fig. 2C). Additionally, our research revealed that RGD-PEG-MZ1 forfeited its capacity to induce BRD4 degradation in non-stressed HUVECs; however, within the context of angiotensin II (Ang-II)-induced ECs, RGD-PEG-MZ1 regained its potency to orchestrate BRD4 proteolysis. The selective proteolytic efficacy of RGD-PEG-MZ1 targeting BRD4 within ECs is underscored by the differential DC₅₀ values; specifically, MZ1 and RGD-PEG-MZ1 exhibit DC₅₀ values of 312 nM and > 500 nM, respectively, for BRD4 protein degradation in untreated HUVECs. In contrast, HUVECs subjected to Ang-II stimulation display a DC₅₀ value of 98 nM for BRD4 degradation by RGD-PEG-MZ1 (Fig. 2D and Fig. S9). In conclusion, our empirical data substantiate that RGD-PEG-MZ1 possesses the ability to selectively home in on ECs and is efficaciously activated within the context of endothelial pathology. This approach differs from other PROTAC-based targeted delivery strategies, such as bioresponsive PROTACs that facilitate targeted protein degradation at tumor sites [6], ROS-triggered PROTACs that promote protein degradation in the liver to ameliorate liver fibrosis [40], or PROTAC drug delivery systems loaded with conventional biological materials [70]. Our research underscores the potential of PROTACs for the targeted intervention of specific cell subpopulations. This innovative concept not only advances the application of PROTACs in precision medicine but also offers a novel perspective for a more comprehensive understanding of the biological functions of relevant targets within distinct cell subpopulations.

Fig. 2.

RGD-PEG-MZ1 manifests a distinct predilection for platelets and is amenable to targeted activation. (A) Fluorescence micrograph illustrating the co-localization of activated platelets and the RGD-PEG-MZ1 conjugate, with RGD-PEG-MZ1 loaded with DID (red fluorescence) and activated platelets labeled with FITC-CD62P (green fluorescence). Scale bar = 100 μm. (B) Fluorescence images of the biodistribution of RGD-PEG-MZ1 in vivo. (C) Immunofluorescence analysis delineating the distribution of CD31 (green) and the RGD-PEG-MZ1 (Red) complex within murine cardiac tissue. Scale bar = 100 μm. (D) The assessment of BRD4 expression across various treatment modalities was ascertained through Western blot analysis

RGD-PEG-MZ1 exhibits favorable biocompatibility and negligible toxicological profile within the therapeutic dosage range

PROTACs, heralded as the cutting-edge modality in oncotherapeutic drug discovery and development, have become the subject of intense scrutiny within the pharmaceutical R&D arena [52, 71–73]. However, similar to traditional small molecular inhibitors, the off-target toxicity of PROTACs has severely hindered its clinical application [50]. Our preceding investigations have identified that the RGD-PEG-MZ1 is capable of selectively inducing the proteasomal degradation of BRD4 within ECs afflicted by pathology (Fig. 2). In theoretical construct, the precise targeting efficacy of RGD-PEG-MZ1 could potentially mitigate the off-target toxicities associated with PROTACs to non-diseased organ systems. To augment the scrutiny of the in vivo biosafety profile of RGD-PEG-MZ1, a 14 days intraperitoneal injection regimen was meticulously implemented in C57BL/6J mice of female and male. The dosing parameters, including the therapeutic dosage, modality, and temporal injection schedule of RGD-PEG-MZ1, were congruent with those employed in the ISO-induced cardiac fibrosis model.

Upon the culmination of a 14 days therapeutic treatment, we procured tissue specimens from the mice. Our initial inquiry focused on the organometric indices of the cardinal organs within these mice. The empirical data elucidated that the RGD-PEG-MZ1 intervention did not engender any marked deviations in the organometric indices of the test subjects (Fig. S10). Concurrently, H&E staining corroborated that the organoarchitectural integrity of the animals remained largely unaltered subsequent to the administration of RGD-PEG-MZ1 (Fig. S11). Furthermore, the haematological profiling of the subjects demonstrated that the RBC, Hb, HCT, WBC, and PLT remained within the physiological parameters without any aberrant deviations subsequent to the administration of RGD-PEG-MZ1 (Fig. S12). Additionally, serological assessments revealed that RGD-PEG-MZ1 exerted no detrimental effects on hepatic and renal functions in murine models, as evidenced by the maintenance of AST, ALT, BUN, UA, CRE, and Glu within homeostatic ranges following RGD-PEG-MZ1 administration (Fig. S13). In conclusion, the aforementioned findings collectively underscore the favorable biosafety profile of RGD-PEG-MZ1 in preclinical models, thereby providing a solid foundation for its potential therapeutic application in the treatment of cardiac fibrosis in vivo.

RGD-PEG-MZ1 exerted a inhibitory effect on the endothelial mesenchymal transition

EndMT represents a critical biological process wherein ECs undergo phenotypic reprogramming to acquire mesenchymal characteristics, playing a significant role in the pathogenesis of fibrosis [26]. The primary objective of developing RGD-PEG-MZ1 was to target and inhibit EndMT as a therapeutic approach to prevent cardiac fibrosis. This novel strategy aims to reverse the transition of ECs into fibroblasts or myofibroblasts, thereby mitigating the fibrotic response and preserving cardiac tissue integrity. To evaluate the therapeutic efficacy of RGD-PEG-MZ1 in modulating EndMT, we investigated its inhibitory effects on EndMT in HUVEC and HMEC-1 cells. Ang-II was employed to induce EndMT in these cell models, effectively mimicking the pathological process. Our results demonstrated that RGD-PEG-MZ1 suppressed EndMT, as evidenced by a marked downregulation of mesenchymal markers (Collagen-I, Fibronectin, α-SMA and FSP1) and a concomitant upregulation of the endothelial marker VE-cadherin (Fig. 3. A, B, C and D). These findings underscore the capacity of RGD-PEG-MZ1 to effectively inhibit EndMT in vitro, highlighting its potential as a therapeutic agent for fibrosis-related pathologies. Moreover, the results of our immunofluorescence assays provide robust evidence that the RGD-PEG-MZ1 conjugate exerts a significant inhibitory effect on the initiation of EndMT (Fig. 3. E). In summary, the empirical findings from our study strongly demonstrate that the RGD-PEG-MZ1 effectively suppresses EndMT in vivo, thereby highlighting its potential as a promising pharmacological agent for the therapeutic intervention of EndMT-associated pathologies.

Fig. 3.

RGD-PEG-MZ1 exerted a profound inhibitory effect on the EndMT in vitro. (A-D) The expression of Collagen-I, Fibronectin, VE-cadherin, α-SMA and FSP1 was assessed using western blotting (WB) in HUVECs and HMEC-1. (E) Immunofluorescence staining for α-SMA (green) and VE-cadherin (green) in HUVECs. Scale bar, 100 μm. The data are presented as the Mean ± SD of three independent experiments. ^#p < 0.05 and ^##p < 0.01 vs. NC group; ns, p > 0.05 vs. control group; *p < 0.05 and **p < 0.01 vs. control group (one-way ANOVA followed by Bonferroni post hoc test)

The MAPK signaling pathway was interfered with the RGD-PEG-MZ1 during EndMT

Our previous experimental findings demonstrate that RGD-PEG-MZ1 suppress the progression of EndMT both in vitro and in vivo (Fig. 3 and Fig. 7). However, the mechanisms by which RGD-PEG-MZ1 regulates EndMT remain unclear. To elucidate the mechanisms by which RGD-PEG-MZ1 regulates EndMT, we performed comprehensive RNA sequencing (RNA-seq) analysis on three independent experimental cohorts: the control group, consisting of untreated HUVECs; the model group, comprising HUVECs subjected to EndMT induction via Ang-II; and the experimental group, in which HUVECs were induced with Ang II and subsequently treated with RGD-PEG-MZ1. In the RNA-seq analysis, scatter plots were initially employed to visualize the global landscape of differentially expressed genes. The analysis revealed substantial gene expression shifts following treatment with RGD-PEG-MZ1, indicating that the compound exerts a pronounced regulatory effect on gene expression profiles (Fig. S14A). To elucidate the role of RGD-PEG-MZ1 in ECs during EndMT, KEGG pathway enrichment analysis was conducted on the differentially expressed genes (DEGs). The findings demonstrated that RGD-PEG-MZ1 intervention elicited pronounced alterations in the MAPK signaling pathway in HUVEC cells, thereby proposing a mechanistic paradigm through which RGD-PEG-MZ1 regulates EndMT via modulation of this critical signaling axis, as supported by an extensive Gene Set Enrichment Analysis (Fig. 4A and B, Fig. S14B). The MAPK signaling pathway, a central orchestrator of diverse cellular functions, has emerged as a pivotal mediator of the EndMT process, characterized by the phenotypic shift of ECs toward a mesenchymal state [74]. This phenotypic transition is evidenced by the downregulation of endothelial markers and concomitant upregulation of mesenchymal markers. To further elucidate the role of MAPK signaling in EndMT, sorafenib, a selective inhibitor of this pathway, was employed. The inhibition of MAPK signaling notably restored endothelial identity, as indicated by the enhanced expression of the endothelial marker CD31 and VE-cadherin, coupled with the attenuated expression of mesenchymal markers including α-SMA following sorafenib treatment (Fig. 4C-E). These results underscore the MAPK pathway’s critical contribution to the EndMT process and its potential as a therapeutic target in modulating this cellular transition.

Fig. 7.

RGD-PEG-MZ1 effectively alleviates EndMT by inhibiting MAPK pathway in vivo. (A) Costaining for the endothelial marker CD31 (red) and the mesenchymal markers α-SMA (green) in myocardial sections. Scale bar, 100 μm. (B) Immunofluorescence staining for RLK1 (red), p-RLK1 (red), ERK1 + ERK2 (green) and p- ERK1 + ERK2 (green) in myocardial sections (n = 4 mice/group). Scale bar, 100 μm

Fig. 4.

The MAPK signaling pathway was modulated by the RGD-PEG-MZ1 during EndMT. (A, B) KEGG analysis of ECs. (C) Immunofluorescence staining for CD31 (red) and VE-cadherin (green) in HUVECs. Scale bar, 100 μm. (D, E) The expression of VE-cadherin and α-SMA was assessed using western blotting (WB) in HUVECs. (F) The expression of ERK1-ERK2, p-ERK1-ERK2, ELK1 and p-ELK1 was assessed using WB in HUVECs. The data are presented as the Mean ± SD of three independent experiments. ^#p < 0.05 and ^##p < 0.01 vs. NC group; ns, p > 0.05 vs. control group; *p < 0.05 and **p < 0.01 vs. control group (one-way ANOVA followed by Bonferroni post hoc test).

Utilizing RGD-PEG-MZ1, we achieved the selective degradation of the BRD4 protein and subsequently evaluated the phosphorylation status of key proteins within the MAPK signaling cascade. Our results revealed a significant attenuation of MAPK signaling activity following BRD4 degradation, as demonstrated by pronounced reductions in the phosphorylation levels of ERK1/2 and ELK1 (Fig. 5F). These pharmacological investigations provide evidence that BRD4 play pivotal roles in the MAPK signaling pathway. However, the precise molecular mechanisms through which BRD4 regulates MAPK pathway activity remain unclear. Taken together, these findings highlight the indispensable role of the MAPK signaling pathway in governing the EndMT.

Fig. 5.

BRD4 enhances MAPK signaling.(A, B) Western blot analysis shows that overexpression of BRD4 in HUVEC cells largely promotes Ang II indiced MAPK. (C) Co-immunoprecipitation detects that BRD4 can directly interact with Raf1. (D) RMSD Analysis of BRD4 and Raf1 Protein Docking from GROMACS Simulation. (E) The 3D diagram of BRD4 binding to Raf1. (F) 2D visualization of the interactions at the interface between BRD4 and Raf1. The data are presented as the Mean ± SD of three independent experiments. ^#p < 0.05 and ^##p < 0.01 vs. NC group; ns, p > 0.05 vs. control group; *p < 0.05 and **p < 0.01 vs. control group (one-way ANOVA followed by Bonferroni post hoc test)

BRD4 directly interacts with RAF1 to enhance MAPK signaling

Recent studies have highlighted the pivotal role of non-histone protein acetylation in regulating a wide array of cellular processes, including signal transduction, gene transcription, protein folding, metabolic homeostasis, and autophagy [75, 76]. BRD4, a well-characterized acetylation reader, exerts its functional influence by specifically recognizing and binding to acetyl-lysine residues on non-histone proteins, thereby amplifying downstream signaling networks [32, 77]. This unique mechanism underscores BRD4’s potential as an attractive therapeutic target for the treatment of a variety of pathologies, such as liver fibrosis, cancer, and heart failure [33, 41, 78, 79]. Pharmacological studies have demonstrated that RGD-PEG-MZ1 significantly attenuates the activity of the MAPK signaling cascade by targeting BRD4 for degradation (Fig. 4). The MAPK signaling cascade, a central pathway governing cellular proliferation, differentiation, and survival, plays a pivotal role in the EndMT [59, 80, 81]. However, the precise molecular mechanism through which BRD4 regulates MAPK signaling remains incompletely understood.

As a direct effector of the RAS protein, RAF1 is indispensable for the full activation of the MAPK signaling pathway. Through its kinase activity, RAF1 phosphorylates and activates MEK, which subsequently activates ERK, ultimately leading to the phosphorylation and activation of numerous nuclear transcription factors, thereby influencing cellular processes such as proliferation and differentiation [82]. Considering the essential role of RAF1 in MAPK pathway activation, we hypothesize that BRD4 may potentiate the regulatory capacity of RAF1 on the MAPK signaling cascade through direct interaction. To test this hypothesis, we overexpressed BRD4 in cells via plasmid transfection. Our experimental results revealed that BRD4 overexpression markedly enhanced the activity of the MAPK signaling cascade (Fig. 5A and B). To further elucidate the relationship between BRD4 and MAPK signaling, we employed co-immunoprecipitation (Co-IP) to investigate the interaction between BRD4 and RAF1 at the protein level. The Co-IP results demonstrated a robust interaction between BRD4 and RAF1, providing direct evidence of their binding (Fig. 5C). Furthermore, structural predictions using AlphaFold3 revealed the binding interface between BRD4 and RAF1, identifying key residues involved in their interaction (Fig. 5D, E and F). These predictions were consistent with Co-IP experimental data, reinforcing the evidence of a direct molecular interaction between BRD4 and RAF1. Collectively, these findings clearly delineate the direct interaction between BRD4 and RAF1, demonstrating that BRD4 enhances MAPK signaling by binding to RAF1. This interaction drives the progression of EndMT and provides critical new insights into the molecular mechanisms underlying EndMT, offering a theoretical foundation for future therapeutic interventions targeting this pathway. However, although our experiments have demonstrated that BRD4 protein degradation attenuates EndMT by inhibiting RAF1-mediated activation of the MAPK signaling pathway, the potential involvement of other regulatory proteins in this process remains unclear. To address this, future studies will employ a combination of CO-IP and MS technologies to systematically elucidate the protein interaction network between BRD4 and the MAPK signaling cascade, thereby providing a clearer understanding of the biological mechanisms underlying BRD4-mediated regulation of EndMT.

RGD-PEG-MZ1 can effectively prevents ISO-Induced cardiac fibrosis

To investigate the therapeutic efficacy of RGD-PEG-MZ1 in mitigating cardiac fibrosis, an in vivo study was conducted using an ISO-induced cardiac fibrosis model [83]. Over a continuous 14-day intervention period, JQ1, MZ1, and RGD-PEG-MZ1 were administered, followed by a comprehensive evaluation of cardiac alterations in the murine subjects. We continuously monitored the body weight of mice throughout the entire treatment period, and the results revealed no significant fluctuations or aberrant weight loss in any treatment group (Fig. S15). These findings indicate that the therapeutic regimen employed in this study exhibits favorable tolerability and a satisfactory safety profile. A suite of experimental methodologies was meticulously applied to characterize the cardiac structural and functional changes resulting from the pharmacological treatments. In addition, we conducted a quantitative assessment of MZ1 levels in mouse cardiac tissue following intervention with either free MZ1 or RGD-PEG-MZ1 using HPLC analysis. The results demonstrated that RGD-PEG-MZ1 exhibited approximately a 2.8-fold increase in cardiac accumulation compared with the free small-molecule MZ1 (Fig. S16). This notable enhancement in tissue deposition further substantiates the cardiac-targeting capability of RGD-PEG-MZ1 and provides compelling evidence supporting its superior tissue enrichment properties. Echocardiography was initially employed to evaluate the cardiotropic effects of JQ1, MZ1, and RGD-PEG-MZ1 on murine cardiac function. The results demonstrated that RGD-PEG-MZ1 significantly improved the ejection fraction (EF%) and fractional shortening (FS%), indicating a marked enhancement in the systolic performance of the murine heart (Fig. S17). RGD-PEG-MZ1 effectively attenuated the pathological morphological and structural abnormalities in myocardial tissue induced by ISO, as demonstrated by a marked reduction in inflammatory cell infiltration (Fig. 6A). Moreover, histological analyses, including Masson’s Trichrome staining, Sirius Red staining, and immunofluorescence, provided compelling evidence that RGD-PEG-MZ1 significantly suppressed collagen deposition, thereby mitigating the progression of cardiac fibrosis (Fig. 6A). To further investigate the in vivo anti-fibrotic potential of RGD-PEG-MZ1, we assessed the expression levels of fibrosis-related molecular markers via Western blot analysis (Fig. 6B-C and S18). The results provided compelling evidence supporting the robust in vivo anti-fibrotic efficacy of RGD-PEG-MZ1. In conclusion, our findings demonstrate that RGD-PEG-MZ1 exerts a therapeutic effect on ISO-induced cardiopathy, effectively mitigating the pathological hallmarks of cardiac fibrosis.

RGD-PEG-MZ1 suppresses EndMT and MAPK signaling in vivo to mitigate cardiac fibrosis

To further clarify the regulatory effect of RGD-PEG-MZ1 on EndMT in vivo, we performed immunofluorescence staining analysis on the heart tissue of mice treated with RGD-PEG-MZ1. The results showed that after treatment with JQ1, MZ1, or RGD-PEG-MZ1, EndMT during fibrosis in mice was significantly inhibited, manifested by a significant decrease in the expression of endothelial cell marker CD31 and mesenchymal cell marker α-SMA (Fig. 7A). In addition, we examined the degradation effect of RGD-PEG-MZ1 on BRD4 in mouse heart tissue and found that RGD-PEG-MZ1 treatment significantly reduced the level of BRD4 (Fig. S19). Meanwhile, immunofluorescence analysis was employed to assess the expression levels of pivotal proteins within the MAPK signaling cascade. The results demonstrated that RGD-PEG-MZ1 intervention markedly attenuated the activation of these key proteins, indicating a significant inhibitory effect on MAPK pathway activation (Fig. 7B). In addition, we assessed the phosphorylation status of pivotal proteins within the MAPK signaling cascade using Western blot analysis. The results revealed that in vivo administration of RGD-PEG-MZ1 markedly suppressed the phosphorylation of the key activating proteins ELK1 and ERK1/ERK2 (Fig. S20). These findings further corroborate the regulatory efficacy of RGD-PEG-MZ1 on the MAPK pathway and underscore its potential to modulate downstream signal transduction. In summary, these data further confirm the inhibitory potential of RGD-PEG-MZ1 on EndMT and MAPK signaling in vivo, providing strong experimental evidence for its treatment of cardiac fibrosis. Although the present study provides compelling evidence supporting the therapeutic efficacy of RGD-PEG-MZ1 in alleviating myocardial fibrosis through integrated in vitro and in vivo analyses, and offers initial insights into its pharmacological mode of action, several limitations inherent to the current experimental design should be carefully acknowledged. Notably, the targeted delivery of RGD-PEG-MZ1 is primarily mediated by a platelet-assisted endothelial targeting mechanism. Given the intrinsic heterogeneity in platelet activation status, adhesion dynamics, and endothelial–platelet interactions across individual subjects, this targeting strategy may introduce inter-individual variability, thereby potentially affecting targeting efficiency and therapeutic robustness. In addition, despite extensive physicochemical characterization, the full structural elucidation of RGD-PEG-MZ1 remains technically challenging. Owing to the limited sensitivity and resolution of current mass spectrometry methodologies for complex macromolecular conjugates, subtle structural heterogeneity and conformational features of RGD-PEG-MZ1 could not be comprehensively resolved. This limitation may partially constrain a deeper understanding of the relationship between nanoscale structural attributes and biological performance. Addressing these limitations, future studies should place greater emphasis on longitudinal and dynamic evaluation of cardiac function to more accurately delineate the temporal therapeutic profile of RGD-PEG-MZ1. Furthermore, comprehensive investigations into tissue biodistribution and pharmacokinetics are warranted to better define its in vivo fate, accumulation behavior, and clearance pathways. At the mechanistic level, more in-depth interrogation of the RAF1–MEK–ERK signaling axis is necessary to clarify its contribution to the observed anti-fibrotic effects. Collectively, these efforts will not only strengthen the mechanistic and materials-based understanding of RGD-PEG-MZ1, but also establish a more solid preclinical foundation for its translational development as a targeted therapeutic biomaterial.

Conclusion

In this study, we successfully engineered a single-component, self-assembling nano-PROTAC (RGD-PEG-MZ1) for endothelial cell delivery to prevent cardiac fibrosis. Initially, we assessed the endothelial cell-specific targeting properties of RGD-PEG-MZ1. Experimental results demonstrated that RGD-PEG-MZ1 exhibits high affinity for activated platelets and exploits their chemotactic influence on endothelial cells to facilitate precise targeted delivery, thereby enabling the selective degradation of BRD4 within endothelial cells. Subsequently, we conducted a comprehensive evaluation of the biocompatibility and preventive efficacy of RGD-PEG-MZ1 against cardiac fibrosis in both in vitro and in vivo models. The results revealed that RGD-PEG-MZ1 possesses excellent biocompatibility and exerts a pronounced antifibrotic effect, particularly by inhibiting EndMT, a pathological process during cardiac fibrosis. Furthermore, we elucidated the molecular mechanism underlying the EndMT-targeting anti-fibrotic activity of RGD-PEG-MZ1. Experimental data indicated that RGD-PEG-MZ1 exerts its inhibitory effects on EndMT and cardiac fibrosis primarily through modulation of the MAPK signaling pathway. Collectively, this study presents an innovative PROTAC-based targeted delivery strategy, offering a novel platform for the precise deployment of PROTACs in specific endothelial subpopulations. Moreover, it establishes a conceptual framework for leveraging PROTAC technology as a chemical biology tool to dissect the functional roles of target proteins in cellular subtypes.

Author contributions

**Tao Bi: ** Writing-original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. **Lei Chen: ** Investigation, Formal analysis, Data curation. **Ting Wang: ** Methodology, Investigation, Formal analysis, Data curation. **Wenjun Miao: ** Supervision, funding acquisition. **Silong Zhai: ** Investigation, Data curation. **Rui Huang: ** Supervision, funding acquisition. **Qin Sun: ** Supervision, Methodology. **Yihan Chen: ** Supervision, Project administration. **Hongna Su: ** Investigation, Data curation. **Jie Zhou: ** Investigation, Data curation. **Ruowen Li: ** Investigation, Data curation. **Weixue Huang: ** Supervision, Project administration. **Xiaojun Yao: ** Supervision, Project administration. **Pei Luo: ** Writing-review & editing, supervision, project administration, funding acquisition, conceptualization. **Zengjin Liu: ** Writing-review & editing, supervision, project administration, funding acquisition, conceptualization.

Funding

This work was supported by the Sichuan Science and Technology Program (No. 2025NSFSC2163), Sichuan Provincial Administration of Traditional Chinese Medicine (No. 25ZDIZX027), Science and Technology Strategic Cooperation Project of Luzhou Municipal People’s Government-Southwest Medical University (No. 2024LZXNYDJ060), the Integration of Traditional Chinese and Western Medicine Conducted at Southwest Medical University (No. 2025ZXYZX01), Guangdong-Hong Kong-Macao Universities Joint Laboratory for the Internationalization of Traditional Chinese Medicine (No. 2023LSYS002) and an open project of the State Key Laboratory of Quality Research in Chinese Medicine funded by the Macau Science and Technology Development Fund (Macau University of Science and Technology, 006/2023/SKL), Macau Special Administrative Region. The funders had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

Data availability

All data used in this paper are available from the corresponding author by reasonable requirements.

Declarations

Competing interests

The authors declare no competing interests.