PROTAC‑Mediated HMGCR Depletion Reprograms Lipid Metabolism in Breast Cancer to Potentiate Photoimmunotherapy via Ferroptosis

ABSTRACT

Aberrant lipid metabolism characterizes the progression of breast cancer. Statins, the canonical agents for modulating this pathway, have been associated with improved overall survival in patients with triple‐negative breast cancer (TNBC). However, their clinical benefit remains limited because the reversible inhibition of 3‐hydroxy‐3‐methylglutaryl‐CoA reductase (HMGCR) elicits a rebound in the mevalonate pathway and enables evasion of ferroptosis. Therefore, we developed a 170 nm self‐assembled nanomedicine (PRO‐P) that integrates an HMGCR‐targeting PROTAC (PRO) with a disulfide‐linked Pyropheophorbide‐a (Ppa) photosensitizer, enabling laser‐gated protein HMGCR degradation and photodynamic stress within one formulation. Under laser irradiation, PRO‐P catalytically depletes HMGCR while generating reactive oxygen species (ROS), collapsing the mevalonate/CoQ10‐GPX4 axis and redirecting lipids into ferroptosis. In 4T1 cells, PRO‐P enhanced cellular uptake by 1.34‐fold and elevated ROS by 9.5‐fold. Following intravenous administration in TNBC xenografts, PRO‐P achieved 92.5% tumor regression, eradicated pulmonary metastases, and elicited no systemic toxicity after single laser exposure. Immune profiling revealed remodeling of the microenvironment, with 2.6‐fold more CD8⁺ Granzyme‐B⁺ T cells, 4.3‐fold more mature dendritic cells, and fewer Tregs, thereby establishing durable memory. PRO‐P exploits multi‐omics–guided HMGCR targeting to convert lipid addiction into a redox–immunologic vulnerability, yielding a low‐toxicity therapy for TNBC and other lipid‐driven cancers.

This work introduces a study that identifies HMGCR as a novel target in TNBC and develops a light‐gated PROTAC nanomedicine. Upon irradiation, this agent selectively degrades HMGCR, reprogramming lipid metabolism to induce ferroptosis and potent antitumor immunity, thereby significantly enhancing photoimmunotherapy efficacy.

1. Introduction

Aberrant lipid metabolism is a critical hallmark of breast cancer progression [1]. In triple‐negative breast cancer (TNBC), dyslipidemia is particularly pronounced, especially post‐menopause [2]. Notably, the clinical use of statins, as classical modulators of lipid metabolism, has been shown to improve TNBC‐specific and overall survival by 58% and 30%, respectively [3]. The central component of this statin‐regulated lipid metabolic pathway is 3‐hydroxy‐3‐methyl‐glutaryl‐CoA reductase (HMGCR), the rate‐limiting enzyme of the mevalonate (MVA) pathway [4]. In addition to cholesterol and isoprenoids regulation [5], sustained HMGCR flux maintains coenzyme Q10‐glutathione peroxidase 4 (CoQ10‐GPX4) antioxidant capacity that detoxifies lipid peroxides and suppresses ferroptosis [6, 7]. Within the tumor microenvironment, GPX4‐mediated detoxification of lipid peroxides sustains redox homeostasis, thereby protecting cancer cells from oxidative stress–based modalities such as ferroptosis and photodynamic therapy [8, 9]. Therefore, pharmacological or genetic interference with HMGCR activity presents a twofold opportunity: it disrupts tumor lipid homeostasis, thereby impairing energy supply, and dismantles a critical barrier to oxidative stress–based therapy resistance, thereby promoting apoptosis of tumor cells. Early studies in 2003 identified HMGCR as a metabolic oncogene via MVA pathway dysregulation in cancers [10]. Statin‐mediated HMGCR inhibition depletes mevalonate and downstream metabolites essential for membrane integrity, signal transduction, protein biosynthesis, and cell cycle control. This established the molecular rationale for repurposing statins as anticancer agents. Subsequent work links chemoresistance to aberrant protein glycosylation/cholesterol metabolism‐pathways targeted by statins. In a phase II trial, Gao et al. showed that the combination of atorvastatin and chemotherapy improved responses in advanced pancreatic adenocarcinoma (70.3% patients achieved ≥20% tumor biomarker reduction) [11]. Concurrently, Yao's team developed zwitterionic polymer‐coated magnetic beads (Fe₃O₄@PCBMA‐SIM) for simvastatin delivery, overcoming triple‐negative breast cancer resistance through MVA pathway inhibition and combined ferroptosis induction [4]. Complementing these therapeutic approaches, Liming Gui et al. further demonstrated that targeting the mevalonate pathway potentiates immunogenic cell death and augments antitumor immunity [12].

Extensive research has been conducted on therapeutic strategies targeting HMGCR‐mediated regulation of lipid metabolism in breast cancer, primarily involving monotherapy using statins and combination therapies integrating statins with chemotherapeutic agents or immunoadjuvants. However, emerging research on the expanding roles of HMGCR and the MVA pathway in cancer metabolic reprogramming has revealed a critical limitation: the reversible inhibition of HMGCR in statin‐based therapies induces compensatory HMGCR gene upregulation [13] that reinstates the MVA pathway, enhances GPX4‐mediated reactive oxygen species (ROS) tolerance [12], and solidifies immune suppression. Furthermore, the elevated doses necessary to counteract this rebound are constrained by dose‐dependent myopathy and hepatotoxicity, which collectively inhibit metabolic remodeling and anticancer immunity [14]. Thus, an optimal technique would (i) irreversibly eliminate HMGCR to prevent compensatory rebound and (ii) augment the combined effects of lipid metabolism–driven energy supply impairment and oxidative stress–based therapies through a controllable approach. Proteolysis‐targeting chimeras (PROTAC) facilitate event‐driven degradation of HMGCR through E3‐ligase recruitment, overcoming the limitations of reversible blockade [15, 16]. Notably, PROTACs have recently been found to possess great potential for reprogramming the immune microenvironment via sustained target protein degradation [17, 18]. PROTAC facilitates catalytic poly‐ubiquitination and proteasomal degradation by co‐recruiting an E3 ligase and HMGCR, resulting in event‐driven, irreversible ablation of HMGCR [19, 20]. Furthermore, to amplify the therapeutic impact of statin‐induced energy supply disruption and oxidative stress–based treatments, while avoiding the systemic toxicities inherent to the traditional chemotherapy or radiotherapy, photodynamic therapy (PDT)‐augmented immunotherapy (PIT) offers a compelling alternative. In this process, near‐infrared laser activates tumor‐bound photosensitizers to produce ROS and induce immunogenic cell death (ICD) exclusively within the illuminated lesion. PIT enables spatially precise tumor eradication, potent local antigen release, and minimal collateral tissue injury [21]. However, to our knowledge, the integration of PROTAC‐mediated HMGCR degradation and PIT represents a promising yet unexplored therapeutic strategy for combating TNBC.

Thus, this work developed a 170 nm self‐assembled nanomedicine (PRO‐P) that integrates an HMGCR‐targeting PROTAC (PRO) with a disulfide‐linked Pyropheophorbide‐a (Ppa) photosensitizer, enabling laser‐gated HMGCR protein degradation and photodynamic stress within one formulation. Integrative scRNA‐seq and Bulk‐transcriptomic database analyses map the HMGCR‐dependent signaling circuitry and immune milieu in breast cancer, sharpening the scientific premise and guiding subsequent development of the PRO‐P (Figure 1). In addition, this nanoplatform was anticipated to obstruct the MVA route irreversibly, diminishing GPX4/CoQ10 defenses, while photodynamically supplying reactive ROS to drive lipid peroxides (LPO) accumulation‐thereby sensitizing tumors to ferroptosis and eliciting ICD within the tumor microenvironment. Consistent with this design, PRO‐P increased cellular uptake 1.34‐fold and showed potent cytotoxicity in 4T1 cells; in vivo, it preferentially accumulated and persisted in tumors, achieved 92.5% tumor inhibition, prevented distant metastasis, and caused minimal systemic toxicity. Immune remodeling was demonstrated by flow cytometry and multiplex immunohistochemistry, with increased CD8⁺ T and natural killer (NK) cells, dendritic‐cell (DCs) maturation, reduced regulatory T (Treg) cells, and emergence of central memory T cells; in addition, the quantitative proteomics supported the intratumoral ferroptosis and ICD events. Moreover, Positron Emission Tomography/Magnetic Resonance Imaging (PET/MRI) demonstrated significant reductions in total lesion glycolysis (TLG) and metabolic tumor volume (MTV) at both primary and pulmonary metastatic sites, indicating dual suppression of tumor metabolism and burden. Collectively, this work (i) establishes HMGCR as a therapeutically actionable driver in TNBC, (ii) constructs and characterizes a self‐delivering PROTAC–Ppa nanoparticle, and (iii) demonstrates laser‐gated HMGCR degradation that reprograms lipid metabolism toward ferroptosis and augments photoimmunotherapy with favorable safety.

FIGURE 1.

Schematic overview of the study workflow and therapeutic cascade for PROTAC‐mediated HMGCR depletion, which reprograms TNBC lipid metabolism and to potentiate photoimmunotherapy via ferroptosis.

2. Experimental Section

2.1. BulkRNA‐seq and scRNA‐seq Analysis

The single‐cell RNA sequencing (scRNA‐seq) data were sourced from dataset GSE176078 (platform GPL18573) in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), comprising 24,277 genes from 9 TNBC patients and data from 32,235 cells. Simultaneously, another dataset, GSE65194 (platform GPL570) from the GEO database, was analyzed using clinical information to get bulk transcriptome (bulkRNA‐seq) data from TNBC patients. The TCGA‐BRCA dataset was obtained from the Cancer Genome Atlas (TCGA) database. Samples lacking survival time, incomplete sample information, or positive for estrogen receptor (ER), human epidermal growth factor receptor 2 (HER2), or progesterone receptor (PR) were excluded. This process yielded transcriptome data for 105 TNBC samples and 85 control samples (https://portal.gdc.cancer.gov/). Both GEO and TCGA were open‐access repositories, with all studies adhering to ethical standards and data being publicly accessible.

2.2. Materials

All chemicals and reagents utilized were of analytical grade purity. All investigations utilized deionized water with a resistivity of 18 MΩ‐cm (Merck Millipore, Germany). Lovastatin, Simvastatin, and the CCK‐8 cell counting kit were acquired from TargetMol (USA). 4T1 cells, DMEM medium, heat‐inactivated fetal bovine serum (FBS), and phosphate‐buffered saline (PBS) were acquired from Haixing Biosciences (China). The Glutathione (GSH) assay kit and ATP assay kit were acquired from Biyuntian Biotechnology Institute (China). The Calcein AM/PI Cell Viability and Cytotoxicity Detection Kit and the Annexin V‐FITC/PI Apoptosis Detection Kit were acquired from Jiangsu Kaiji Biotech Co. (China). 2′,7′‐dichlorodihydrofluorescein diacetate (DCFH‐DA) was acquired from MedChemExpress (USA). The BODIPY 581/591 C11 probe was acquired from Thermo Fisher Scientific (USA). The HMGCR antibody and the GPX4 antibody were acquired from Abcam (UK). The GAPDH antibody, β‐Actin antibody, HMGB1 antibody, and CRT antibody were acquired from Proteintech (China). Sodium deoxycholate (SDC), urea, triethylammonium bicarbonate (TEAB), dithiothreitol (DTT), iodoacetamide (IAA), and ammonium bicarbonate (ABC) were purchased from Sigma. Sequencing grade modified trypsin was purchased from Promega.

Transmission electron microscopy (TEM) was conducted using a Hitachi HT7700 microscope (Japan) at an acceleration voltage of 100 kV, while HRTEM was performed using a FEI Tecnai G2 F30 (USA). Ultraviolet‐visible‐near infrared (UV‐Vis‐NIR) absorption spectra were acquired utilizing a Shimadzu UV‐3600i Plus spectrometer (Japan). In vivo fluorescent imaging of small animals was conducted using a PerkinElmer IVIS Lumina XRMS series imaging system (USA).

2.3. Synthesis of PROTAC

Lovastatin (15.0 g, 37.08 mmol) was measured into a 500 mL round‐bottom flask, dissolved by the addition of 100 mL of ethanol and 300 mL of H₂O, and agitated at ambient temperature. Subsequently, KOH (20.8 g, 370.8 mmol) was added in increments, and the mixture was heated to 95°C with continuous stirring for 24 h. The reaction was examined by TLC (Expanded Coagent). The reaction was identified using TLC (DCM: MeOH = 15:1). Upon completion of the response involving raw materials, the reaction solvent was evaporated under vacuum. Subsequently, the residue was dissolved in 100 mL of DCM and 500 mL of H₂O, with the pH adjusted to 2 using concentrated hydrochloric acid (12 m) and stirred overnight. The organic phase was then collected, dried with anhydrous sodium sulfate, and subjected to silica gel column chromatography, yielding a white solid (11.8 g, 27.14 mmol) with a yield of 73.2%.

Compound 1 (5.0 g, 15.6 mmol) was measured into a 250 mL round‐bottom flask, followed by the addition of 100 mL of DCM to achieve complete dissolution, which was stirred at room temperature for 20 min. Subsequently, TFA (1.89 g, 18.7 mmol) was introduced dropwise while maintaining an ice bath, and stirring continued for 30 min under identical conditions, after which TBDMSCL was added incrementally. TBDMSCL was subsequently introduced in increments, the ice bath was discontinued, and the temperature was elevated to 35°C while stirring. The reaction was monitored using TLC (solvent system DCM: MeOH = 20:1). Upon completion of the reaction, the solvent was evaporated under vacuum, followed by washing with saturated sodium bicarbonate solution. The product was then purified via silica gel column chromatography, yielding a white solid (3.8 g, 8.70 mmol) with a 55.9% yield.

Compound 3 (4.0 g, 9.2 mmol) and compound 4 (1.2 g, 11.0 mmol) were measured into a 150 mL round‐bottom flask. EDCI (2.7 g, 14.3 mmol) and DMAP (1.9 g, 15.4 mmol) were introduced and agitated at ambient temperature for 20 min, then heated to 35°C and stirred overnight. The reaction process was monitored using TLC with a spreading agent ratio of PE: EA = 1:1. Upon completion of the reaction involving the raw materials, the solvent was evaporated under reduced pressure, yielding a yellow oil. This oil was subsequently dissolved in 200 mL of DCM, washed with saturated saline, and the organic phase was collected and dried using anhydrous sodium sulfate. The product was then purified via silica gel column chromatography (PE: EA = 10:1‐4:1), resulting in a white solid (2.2 g, 4.19 mmol) with a yield of 45.5%.

Compound 5 (1.0 g, 1.90 mmol) and compound 6 (741.2 mg, 2.08 mmol) were measured in a 150 mL round‐bottom flask, to which 50 mL of methanol was added for dissolution, followed by the addition of CuSO₄·5H₂O (10.0 mg, 0.04 mmol) and ascorbic acid (14.1 mg, 0.08 mmol). The reaction was monitored using TLC (solvent system DCM: MeOH = 15:1). Following the complete reaction of the starting materials, the solvent was evaporated under reduced pressure to yield a laser green solid. Subsequently, 200 mL of DCM was added for dissolution. The mixture was subjected to silica gel column chromatography (solvent gradient DCM: MeOH = 100:1 to 30:1), resulting in the isolation of a white solid (595.3 mg, 0.67 mmol) with a final yield of 35.3 mg, 0.67 mmol, after solvent removal—35.4% yield (mmol).

Compound 7 (500.0 mg, 0.56 mmol) was measured into a 100 mL round‐bottom flask, to which 50 mL of methanol was added for dissolution. Subsequently, TsOH (389.1 mg, 2.26 mmol) was incorporated, and the mixture was heated to 35°C and agitated for 1 h. The reaction's progress was monitored by TLC (solvent system DCM: MeOH = 15:1), confirming the completion of the raw material's response. Upon the complete reaction of the raw materials, the solvent was evaporated under reduced pressure to yield a yellow solid. Subsequently, 200 mL of DCM was introduced for dissolution, followed by sand addition. A white solid (178.3 mg, 0.23 mmol) was isolated via silica gel column chromatography (DCM: MeOH = 100:1‐20:1), achieving a yield of 41.3%.

2.4. Synthesis of Ppa‐SS‐AA

This two‐step reaction was completed by esterification. Arachidonic acid (2.0 g, 6.57 mmol) was charged into a 100 mL round‐bottom flask, followed by dissolution in 30 mL of DCM under stirring at the same temperature for 5 min. EDCI (1.6 g, 8.53 mmol) and DMAP (1.0 g, 8.53 mmol) were weighed and added to the flask. The mixture was stirred at this temperature for an additional 10 min. Subsequently, Compound 1 (1.3 g, 8.54 mmol) was dissolved in 10 mL of THF and added dropwise to the reaction mixture at the same temperature. After complete addition, the reaction was stirred at 30°C for 4 h. The progress was monitored by TLC (eluent: PE/EA = 1:1). Upon complete consumption of the starting material, the solvent was removed under reduced pressure to afford a yellow oil. The crude product was dissolved in 200 mL of DCM, washed with saturated brine, and the organic layer was collected and dried over anhydrous Na₂SO₄. Purification via silica gel column chromatography (PE:EA = 8:1 to 4:1) yielded a pale yellow solid (2.1 g, 4.94 mmol) with a yield of 75.2%. Ppa (0.5 g, 0.94 mmol) was dissolved in 30 mL of DCM and added dropwise to the compound 2 solution in the presence of EDCI/DMAP. The product (Ppa‐SS‐AA) was separated and purified by silica gel column chromatography.

2.5. Synthesis of PRO‐P

Self‐assembled nanoparticles were fabricated via the nanoprecipitation method, wherein the synthesized AA‐SS‐Ppa and PROTAC molecules were dissolved in a minimal quantity of organic solvent. A small amount of DSPE‐PEG served as a stabilizer and was gradually introduced dropwise into deionized water under vigorous stirring, facilitating the evaporation of the organic solvent.

2.6. Physicochemical Characterization of Self‐Assembled Nanoparticles PRO‐P, PROTAC, and Ppa

The hydrodynamic size, polydispersity index (PDI), and zeta potential of the self‐assembled nanocomplexes PRO‐P, PROTAC, and photosensitizer Ppa were assessed using a Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK). Their particle shape in ultrapure water was analyzed using transmission electron microscopy (TEM, CM‐200, Philips, USA). The elemental composition of PRO‐P was examined using energy‐dispersive X‐ray spectroscopy (EDS). The distinctive emission peaks of PRO‐P, PROTAC, and Ppa were assessed using a UV spectrophotometer. The absorbance values of various doses of PRO‐P, following co‐incubation with mouse blood and 4T1 cells, were evaluated using the SPARK Multifunctional Enzyme Labeling Instrument, and the hemolysis rate and half inhibitory concentration of the medication were determined.

2.7. Cellular Uptake

4T1 cells were seeded in six‐well plates at a density of 1 × 10⁵ cells and incubated at 37°C with 5% CO₂ for 24 h. After the adherence of the cells to the wall, drugs were administered at various time intervals (2, 4, 8, 12, 24, and 36 h). The cells underwent two washes with PBS and were subsequently fixed in 4% paraformaldehyde. The self‐assembled nanoparticles of PRO‐P were visualized using a confocal microscope and flow cytometer, assessing the cellular uptake and cytotoxicity of PROTAC and the Ppa in 4T1 cells.

2.8. Intracellular ROS Production

Intracellular reactive oxygen species (ROS) production was assessed via confocal laser scanning microscopy and flow cytometry utilizing DCFH‐DA at a concentration of 20 µm as a ROS indicator. The 4T1 cells were seeded in six‐well plates at a density of 1 × 10⁵ cells and incubated at 37°C with 5% CO₂ for 24 h. Following the adherence of the cells to the wall, drugs were administered, and the cells were co‐incubated with the pharmaceuticals for 24 h. The cells were exposed to laser irradiation (with a laser of 660 nm, 100 mW/cm², 1 min) following the addition of DCFH‐DA, and the generation of reactive oxygen species was assessed using confocal microscopy and flow cytometry after 30 min.

2.9. Evaluation of the Effect of Tumor Immunogenic Death Produced

The investigation of immunogenic death involved alterations in mitochondrial structure and the production of damage‐associated molecular patterns (DAMPs) in 4T1 cells. To assess alterations in mitochondrial structure, 4T1 cells were seeded in six‐well plates at a density of 1 × 10⁵ cells and incubated at 37°C with 5% CO₂ for 24 h. The cells were subsequently exposed to various treatments. After 24 h, the medium was substituted, and the cells were treated with JC‐1 (5 µg/mL) at 37°C for 20 min. Fluorescence signals were subsequently examined using CLSM following three washes with PBS. The expression of DAMPs in 4T1 cells primarily involved the detection of calreticulin (CRT) and high mobility group protein B1 (HMGB1). The 4T1 cells were cultured in glass dishes at a density of 1 × 10⁵ cells per dish for 24 h. The cells underwent distinct treatments, and after 20 h, they were rinsed with PBS and fixed with 4% paraformaldehyde for 15 min. The levels of CRT and HMBG1 were subsequently measured following the assay kit instructions.

2.10. Cytotoxicity Assay

2.10.1. CCK8 Assay

4T1 cells were seeded in 96‐well plates at a density of 1 × 10³ cells and incubated for 24 h at 37°C with 5% CO₂. Following cell attachment, the experimental group received drug‐containing medium, while the control group was supplemented with non‐drug‐containing medium, and incubation continued for an additional 24 h. Following 1 min of laser illumination, 10 µL of CCK8 reagent and 90 µL of culture media were added to each well. The samples were cultured for 30 min to 2 h, after which the absorbance at 450 nm was measured using an enzyme marker to assess cell viability.

2.10.2. Flow Apoptosis Assay

4T1 cells were seeded in six‐well plates at a density of 1 × 10⁵ cells. The cells were incubated with the drug for 24 h, followed by laser illumination for 1 min. The medium was subsequently removed, and the cells were washed once with PBS. Trypsin (500 µL per well) was used to digest the cells for 2 min, and the digestion was halted by the addition of serum to the complete medium (1 mL per well). The products underwent centrifugation, the supernatant was removed (1500 rpm, 3 min), this process was repeated once, apoptotic dye (Annexin V‐PI double labeling) was introduced, and the cells were subsequently examined using a CytoflexS flow analyzer after 15 min of laser exposure in tinfoil.

2.10.3. Cell Cycle Assay

4T1 cells were seeded in six‐well plates at a density of 1 × 10⁵ cells. The cells were incubated with drugs for 24 h, followed by laser illumination for 1 min. The medium was then removed, and the cell suspension was washed once with PBS. The cell suspension was collected, and the supernatant was discarded (1,200 rpm, 3 min), followed by two washes with PBS. The cells were fixed in 75% ethanol overnight, after which the ethanol was discarded (1,200 rpm, 3 min), and the cells were washed once with PBS. A staining solution (RNase: PI = 1:9) was added for 1 h, and the samples were analyzed using a CytoflexS flow analyzer.

2.11. Western Blot Experiment

The 4T1 cells were distributed in 6‐well plates (1 × 10⁷ cells per well), incubated with drugs for 12 h, subjected to laser illumination for 1 min, and thereafter incubated for a further 12 h. The medium was discarded, and the plates were rinsed once with PBS. Protease, phosphatase, and protein cleavage solution were proportionally added to the 6‐well plates, followed by centrifugation of the products, and extraction of the supernatant (13,300 rpm for 20 min) to isolate the proteins. The BCA assay was conducted, and the protein sample concentration was determined. Prepare PAGE gel, sequentially add Maker and protein, perform electrophoresis (40 V for 30 min; 80 V for 1 h), conduct membrane transfer (250 mA for 80 min), incubate in 5% skimmed milk powder or BSA solution at room temperature or 37°C with slow shaking for 1–2 h, incubate with primary antibody overnight, after incubation, remove the primary antibody, wash with TBST on a shaking table for 10 min, and repeat the rinse three times with shaking incubation. The secondary antibody was incubated for 1 to 2 h. Following incubation, the secondary antibody was aspirated, TBST was agitated on a shaking table for 10 min, and rinsed three times before the bands were exposed.

2.12. Immunofluorescence Staining

Cells were inoculated onto coverslips, and when reaching 70%–90% confluence, they were treated with drugs for 12 h. Following incubation, the cells were subjected to laser illumination for 1 min and subsequently incubated for a further 12 h. After medium removal, the cells were fixed with 4% paraformaldehyde for 15 min, rinsed with PBS, permeabilized with 0.1%–0.5% Triton X‐100 for 10 min, and washed with PBS three times. Thereafter, the cells were incubated with 3% BSA or 5% serum for 1 h to block. The diluted primary antibody (1:100–1:500) was applied and incubated overnight at 4°C or room temperature for 1–2 h. Following PBS washing, a fluorescent secondary antibody (1:200–1:1000) was incubated for 1 h in the absence of laser, and the nuclei were further stained with DAPI (1 µg/mL) for 15 min. The slices were ultimately sealed with a drop of anti‐quenching sealer, maintained in a laser‐protected environment, and examined using a fluorescence microscope promptly.

2.13. Scratch Experiment

Disperse 4T1 cells in a 6‐well plate and allow them to attain a confluence of 90% or greater to establish a monolayer. Utilize a sterile pipette tip or cell spatula to perform a vertical scratch, ensuring consistent pressure to create a linear gap. Rinse with PBS to eliminate the exfoliated cells. Substitute the serum‐free media containing the medication under investigation, while the control group utilizes solely serum‐free medium. Photographs were captured from a stationary posture under a microscope at 0 and 36 h to document the alteration in the width of the scratches. The scratch area was quantified utilizing software like ImageJ, and the mobility was computed.

2.14. In Vivo Tissue Distribution Studies of Co‐Assembled Nanoparticles

Following the preparation of co‐assembled nanoparticles of Ppa (5 mg/kg) fluorescent dye, they were administered to tumor‐bearing mice via the tail vein and imaged at various time intervals (2, 6, 12, 24, and 36 h) for in vivo visualization, with the distribution of the nanoparticles in different tissues quantified by fluorescence intensity.

2.15. Co‐Culture Experiment of Tumor and Immune Cells

4T1 cells were seeded in Transwell and co‐incubated with immune cells (lymphocytes isolated from mouse spleen). The cells were cultured in RPMI‐1640 medium containing 10% FBS, collected after 24 h, and the immune effect was evaluated by flow cytometry using immune cell activation markers (DC cells: anti‐CD11c‐BV421 antibody, anti‐CD80‐APC antibody, anti‐CD86‐PE‐Cy7 antibody; cytotoxic T cells: anti‐CD45‐FITC antibody, anti‐CD3‐PC5 antibody, anti‐CD8‐PE antibody; Treg: anti‐CD3‐BV510 antibody, anti‐CD4‐APC antibody, anti‐CD25‐PE‐Cy7 antibody).

2.16. Pharmacodynamic Study of Combined Ferroptosis‐PDT in Animals

Following the establishment of a subcutaneous tumor model for mouse mammary cancer, antitumor therapy was administered via tail vein injection of co‐assembled nanoparticles combined with laser irradiation (660 nm, 100 mW/cm², 5 min; spot diameter: 1 cm; single‐session dose: 30 J/cm²). Tumor dimensions were assessed every three days, and the body weight of the mice was documented until 18 days post‐administration of the nanoparticles. All animal experiments were conducted in accordance with relevant guidelines and regulations and were approved by the Animal Ethics Committee of Nanjing Hospital, Affiliated to Nanjing Medical University (Approval No. DWSY‐24067523). Isoflurane was used for anesthesia during surgical and imaging procedures. At the experimental endpoint, euthanasia was performed by cervical dislocation following deep anesthesia with an overdose of sodium pentobarbital. Mice were euthanized when the tumor volume reached 1500 mm³ or if they lost over 20% of their body weight. Subsequently, the harvested tumor tissues were processed for H&E staining to examine and compare the extent of cancer cell death across different treatment groups.

2.17. Analysis of Antitumor Immune Response

Tumor tissue and spleen were digested to prepare a single‐cell suspension, which was filtered through a 200‐mesh nylon filter to remove clumps, and the filtrate was collected. The cells were incubated in the dark with specific fluorescently labeled antibodies (4°C, one h) and then analyzed by flow cytometry to detect the expression profiles of surface markers of immune cell subsets (T cells: anti‐CD45‐FITC/APC antibody, anti‐CD3‐BV510 antibody, anti‐CD8‐PE antibody, anti‐CD4‐APC antibody, anti‐Granzyme B‐FITC antibody, anti‐CD25‐PC7 antibody; NK cells: anti‐CD45‐APC antibody, anti‐CD3‐BV510 antibody, anti‐CD49b‐FITC antibody; DCs: anti‐CD11c‐BV421 antibody, anti‐CD80‐APC antibody, anti‐CD86‐PE‐Cy7 antibody; Memory cells: anti‐CD45‐FITC antibody, anti‐CD3‐BV510 antibody, anti‐CD44‐APC antibody, anti‐CD62L‐eFluor 450 antibody).

2.18. Construction of Mouse Metastasis Models

2.18.1. Distant Tumor Model

First, 4T1 breast cancer cells were implanted in the left inguinal region of BALB/c mice to establish a primary tumor model. After treatment of the primary tumor was completed, 4T1 cells were implanted in the right inguinal region of the same group of animals to simulate untreated distant secondary tumors. Following distant tumor implantation, tumor growth was continuously monitored for 17 days.

2.18.2. Lung Metastasis Model

First, 4T1 breast cancer cells were orthotopically implanted into the left inguinal region of BALB/c mice to establish a primary tumor model. After the treatment intervention for the primary tumor was completed, 4T1 breast cancer cells were selected to prepare a log‐phase single‐cell suspension. A 100 µL cell suspension (containing 1 × 10⁶ cells) was injected into the tail vein of BALB/c mice to simulate the process of hematogenous metastasis. After 17 days, lung tissue was collected for ex vivo imaging, H&E staining, and metastasis nodule counting to assess the lung metastasis burden quantitatively.

2.19. PET/MR Imaging and Quantitative Analysis

Mice bearing TNBC tumors were intravenously injected with 100 µL of 18F‐FDG via the tail vein. For baseline imaging, an activity of 3.7 MBq was administered, while post‐treatment studies employed 7.4 MBq to enhance detection sensitivity for small lesions. After a 20–40 min uptake period under physiologically stable conditions, animals were anesthetized with (specify an anesthetic agent and concentration). Imaging was performed on a 3.0 T hybrid PET/MR scanner (uPMR 790; United Imaging Healthcare, Shanghai, China). PET data were acquired over a 20‐min duration with a reconstructed voxel size of 1.17 × 1.17 × 1.2 mm³. Concurrent MR imaging was performed using two protocols: whole‐body T₂‐weighted imaging (TR/TE: 5048/81.2 ms, slice thickness: 1.0 mm, slice spacing 0.1 mm, field of view 50 × 100 mm², matrix 112 × 224) and localized high‐resolution T₂‐weighted imaging (TR/TE: 5048/84.2 ms, slice thickness: 1.0 mm, slice spacing 0.1 mm, field of view 30 × 30 mm², matrix 144 × 144), focusing on tumor and lung regions.

Standardized uptake values normalized to body weight (SUVbw) were calculated as:

$$\mathrm{SUVbw}=\frac{\mathrm{Activity}\mathrm{concentration}\left(\mathrm{kBq}/\mathrm{ml}\right)}{\mathrm{Injected}\mathrm{dose}\left(\mathrm{MBq}\right)\times \mathrm{Decay}\mathrm{factor}/\mathrm{Body}\mathrm{weight}\left(\mathrm{Kg}\right)}$$

where Decay factor = exp(−0.693 × wait time/radionuclide half‐life).

Two nuclear medicine specialists delineated tumor boundaries on T₂‐weighted MR images to define regions of interest (ROIs), ensuring precise anatomical localization. The metabolic tumor volume (MTV) was derived from the volumetric measurement of these ROIs. The defined ROIs were then mapped onto the corresponding PET images to obtain the mean SUV_bw within each region. Total lesion glycolysis (TLG) was calculated as the product of the mean SUV_bw and MTV.

2.20. Quantitative Proteomics

2.20.1. Sample Preparation

Mouse tumor tissues were collected and washed with PBS. The tissues were homogenized using 8 m urea in SDC buffer (1% SDC in 0.1 m TEAB), followed by sonication for 30s. Next, protein concentration was measured by the Pierce BCA Protein Assay Kit.

The extracted proteins were reduced by eight mm DTT at 56°C for 30 min, alkylated by 16 mm IAA at room temperature for 30 min in the dark, and then diluted by 100 mm ABC buffer to 1.6 m urea. The proteins were digested by trypsin at 37°C (1/50, w/w) for 20 h. Peptides were desalted using C18 tips. Briefly, the C18 tips were conditioned using methanol, washed with 80% ACN (v/v) in 0.1% formic acid (FA), equilibrated with 0.1% FA, followed by loading peptides in 0.1% FA. The digested peptides were desalted using 0.1% FA and finally eluted by 80% ACN in 0.1% FA, which were concentrated using SpeedVac for subsequent MS analysis.

2.20.2. Mass Spectrometric Analysis and Data Processing

Peptides were redissolved into 0.1% FA in 2% ACN and measured by a nanoLC system coupled to an Orbitrap Astral mass spectrometer. Solvent A was 0.1% FA and solvent B was 0.1% FA in 80% ACN, respectively. Peptides were separated using a home‐made C18 nano‐capillary analytical column. Starting with 1% solvent B by nanoLC system, the peptides were mainly separated from 9% solvent B at 1.6 min, to 25% at 14.6 min, and then to 40% at 20 min. Mass spectrometric data were acquired using data‐independent acquisition (DIA) mode. The MS1 parameters were set as follows: Orbitrap resolution at 240,000 FWHM, AGC target at 5 × 10⁶, and m/z scan range at 380–980. MS2 parameters were set as follows: isolation window of 2, higher‐energy collisional dissociation (HCD) at 25%, AGC target at 5 × 10⁴.

The mass spectrometric data were analyzed by DIA‐NN by precursor FDR at 1% using the reviewed Mus musculus proteome database downloaded from Uniprot [22]. The quantitative proteomics was performed by Perseus software to profile the quantitative proteomics [23]. Gene ontology (GO) enrichment analyses were performed by clusterProfiler [24].

2.21. Statistical Analysis

For the semiquantitative analysis of CLSM fluorescence intensity, all images were processed using ImageJ software, with a minimum of three samples evaluated per group. Each experiment was independently replicated at least three times. Data were presented as mean ± standard deviation (SD). Comparisons between two groups were conducted using a two‐tailed Student's t‐test, while comparisons across multiple groups were carried out by one‐way analysis of variance, both performed in GraphPad Prism 9 (GraphPad Software, Inc., San Diego, CA). A p‐value < 0.05 (*p < 0.05) was defined as statistically significant. Differences denoted by **p < 0.01, ***p < 0.001 and ****p < 0.0001 were considered highly statistically significant.

3. Results and Discussion

3.1. Bulk RNA‐seq and scRNA‐seq Analysis Revealed that HMGCR Promotes Tumor Growth

Statins, the standard pharmacological agents for dyslipidemia, primarily operate by inhibiting HMGCR, thereby modulating the hyperactive MVA metabolic pathway [25]. The biological role of the HMGCR in TNBC tumor cells and its effects in the tumor microenvironment still require further investigation. Thus, we acquired transcriptome data from 105 TNBC tissues and 85 healthy controls from adjacent tissues to investigate the biological role of the HMGCR. Leveraging the GeneMANIA database, we conducted a systematic analysis of functional interactors associated with HMGCR, identifying a network of 20 coregulated genes (including INSIG1, SREBF2, ACAT2, FDPS, MSMO1, and NPC1). This integrated bioinformatic profiling reveals that HMGCR and its co‐expression partners constitute a cohesive functional module governing sterol metabolic homeostasis. (Figure 2a). Gene set enrichment analysis (GSEA) related to HMGCR identified 20 pathways with the highest enrichment for visual display, and the results revealed significant enrichment of the gene in multiple breast cancer‐related signaling pathways (Figure 2b), highlighting subtype‐specific transcriptional programs: SMID Breast Cancer Basal UP (cholesterol‐driven oncogenesis) and Basal DN (tumor suppressor inactivation). Further analysis revealed that the HMGCR gene exhibited significant differential expression in TNBC samples, and its expression level was notably elevated (Figure 2c), providing strong evidence for the potential association between candidate genes and TNBC treatment. Subsequently, the expression levels of HMGCR were examined between TNBC cancer tissue and surrounding tissues, demonstrating a tendency toward overexpression in both (Figure 2d,e). Furthermore, the predictive value of HMGCR expression for patient survival outcomes was assessed using time‐dependent TimeROC curves. The AUC values increased over time (0.55, 0.62, and 0.65 for 3‐, 5‐, and 8‐year survival, respectively), suggesting that HMGCR expression may have a slightly better association with long‐term survival outcomes (Figure 2f). These findings suggest that HMGCR exhibits increasing predictive value for long‐term patient survival outcomes. Given that immunotherapy is presently a significant advantage for TNBC patients, the study further examined the Immunophenotypic Score (IPS) between groups with high and low HMGCR expression. It indicated that individuals in the low expression cohort derive greater benefits from immunotherapy than those in the HMGCR high expression cohort (Figure 2g,h).

FIGURE 2.

Bulk RNA‐seq and scRNA‐seq analysis revealed that HMGCR promotes tumor growth by establishing an immunosuppressive microenvironment. (a) GeneMANIA analysis related to HMGCR. (b) GSEA Ridge Map related to HMGCR. (c) Volcano plot of differentially expressed genes between TNBC and normal sample groups. (d,e) Expression of HMGCR in TNBC cancer tissue and surrounding tissues. (f) The precision of HMGCR expression in patient diagnosis and prognostic outcome prediction was assessed using the TimeROC curve. (g,h) IPS score of high and low HMGCR expression groups. (i) UMAP diagram for cell annotation. (j) UMAP clustering diagram of cells across various TNBC groups. (k) Various categories consist of subsets of TNBC cells. (l) The expression differences of the HMGCR gene in TNBC among various cells. (m,n) Diagram of dimensionality reduction and grouping of T cell clusters. (o) Cytotoxicity functional score.

Then, the bulk RNA‐seq analysis indicated that elevated expression of HMGCR may promote the development of an immunosuppressive microenvironment in malignancies. Inversely, its downregulation may facilitate the release of immunological factors and the recruitment of chemotactic immune cells to infiltrate tumor tissues, thus promoting the response to immunotherapy. To gain further insight into the role of HMGCR at the cellular level within the tumor microenvironment (TME), we performed single‐cell RNA sequencing to perform high‐resolution examination of 24,277 genes and 32,235 cell samples from 9 patients with TNBC, by dimensionality reduction and clustering analysis techniques, alongside conventional cell markers and the CellMarker 2.0 database for cell annotation, identifying seven principal cell types: B cells, fibroblasts, macrophages, T cells, endothelial cells, vascular smooth muscle cells, and epithelial cells (Figure 2i). Various cell types were examined for differential gene expression utilizing the FindAllMarkers tool, and the findings were illustrated through heatmaps (Figure 2j). Subsequently, stacked plots illustrated the proportion of various cell types and their distinctions in each sample (Figure 2k). The detailed analysis of the differential expression of the HMGCR gene across several cell types in TNBC revealed that HMGCR exhibited high expression levels in epithelial cells and T cells (Figure 2l). To thoroughly examine the expression pattern of HMGCR in T cell and its impact on T cell functionality, we performed independent downscaling and clustering analyses based on T cell clusters and categorized the resultant T cell subpopulations as follows: Treg, CD8⁺ T cells, NK cells, central memory T cells, cytotoxic T cells, RPSs⁺ T cells, and precursor T cells (Pre T cells) (Figure 2m,n). Subsequently, 12 pivotal cytotoxicity‐related genes (GZMA, GZMB, GZMH, GZMK, GZMM, GNLY, PRF1, FASLG, IFNG, TNF, IL2RA, and IL2) were evaluated using the AddModuleScore function to determine the impact of HMGCR expression on T cell toxicity, showing that cytotoxicity scores in the HMGCR‐negative group were significantly elevated compared to the positive group (Figure 2o). These findings suggest that pharmacological or genetic interference with HMGCR activity may promote the production of a robust and enduring immune response, potentially serving as a novel target to improve the effectiveness of TNBC immunotherapy.

Dual‐omics integration analysis demonstrates that HMGCR is highly elevated in TNBC and possesses predictive value. HMGCR inhibits T cell cytotoxicity, and its reduced expression may improve the efficacy of immunotherapy. The investigation of clinical samples indicates that elevated HMGCR expression facilitates the development of a tumor immunosuppressive microenvironment, hence expediting the evolution of TNBC. Conversely, diminished expression of HMGCR can elicit a systemic immune regulatory response. The data indicate that HMGCR may serve as a significant target for the therapy of TNBC. Nonetheless, while conventional statins suppress HMGCR, their high‐dose administration can provoke compensatory upregulation of HMGCR. Therefore, current approaches to HMGCR modulation remain suboptimal and require strategic refinement.

3.2. Synthesis and Physicochemical Characterization of PRO‐P

An ideal solution would irreversibly suppress HMGCR to avert mevalonate‐pathway rebound while converting the resulting lipid‐peroxide surge into a controllable, tumor‐selective cytotoxic cue. Thus, this study presents the design and synthesis of an innovative PROTAC featuring a dual functional architecture, wherein one terminus functions as a ligand for HMGCR with a statin molecule as the core structure (Figure 3a). Importantly, the opposing terminus acts as a ligand for the E3 ubiquitin ligase CRBN, linked by a short‐chain alkyl connector. This PROTAC is engineered to selectively bind to HMGCR, facilitating the recruitment of E3 ubiquitin ligase to mediate the ubiquitination of target proteins, ultimately leading to the degradation of HMGCR via the ubiquitin‐proteasome system (UPS). To optimize the delivery system, the amphiphilic molecule AA‐SS‐Ppa was initially synthesized by conjugating the unsaturated fatty acid with the photosensitizer Ppa through a disulfide linkage. This molecule was then co‐assembled with PROTAC via nano‐precipitation to create self‐assembled nanoparticles encapsulating PROTAC, designated as PRO‐P (Figure 3a; Figures S1 and S2). The chemical structure of the intermediate was validated by ¹H‐NMR (Figure S3). The assembly process is attributed to the intermolecular π–π stacking resulting from the benzene rings and double bonds present in the molecules of PROTAC and AA‐SS‐Ppa. Also, by utilizing hydrophobic interactions, the prodrug coupling compound autonomously generated nanoparticles in an aqueous medium, which not only retained the benefits of macromolecular nanocarriers but also achieved a high drug loading capacity through the self‐assembly form, thereby significantly diminishing drug dosage and potential toxicity [26]. To furtherly improve the colloidal stability in a physiological setting, phospholipid‐polyethylene glycol (DSPE‐PEG) was added to the nanoparticle surface.

FIGURE 3.

Synthesis and Physicochemical Characterization of PRO‐P. (a) Molecular structural formulas of PROTAC and AA‐SS‐Ppa. (b–d) Diameter dimensions and transmission electron microscopy images of PRO‐P, PRO, and Ppa, scale bar: 200 µm. (e) Energy‐dispersive X‐ray spectroscopy elemental mapping images of PRO‐P, scale bar: 100 µm. (f) Atomic Force Microscopy of PRO‐P. (g) Polymer dispersibility index of PRO‐P in water. (h) Zeta potential of PRO‐P. (i) UV–vis absorption of PRO‐P, PRO, and Ppa in water. (j,k) Hemolysis experiments of different drugs (10 µg/mL) and various concentrations of PRO‐P.

Next, the physicochemical properties of PROTAC, Ppa, and the nano‐assembly PRO‐P were extensively studied. The integrated transmission electron microscopy (TEM), dynamic laser scattering (DLS), and Atomic Force Microscope (AFM) examination revealed that PRO‐P displays an average particle size under 200 nm (Figure 3b–d,f). Compared with Ppa (∼130 nm), PRO‐P exhibited a slightly larger average particle size (∼170 nm) due to hydrophobic linkages in the composite structure that promoted efficient assembly, yet remained below 200 nm. This nanoscale characteristic facilitates passive targeting of tumor tissues by leveraging the enhanced permeability and retention effect (EPR) of tumors, thereby improving medication accumulation and retention at the tumor location. Next, the Energy‐dispersive X‐ray spectroscopy (EDS) analyses (Figure 3e; Figure S4) further confirmed that PRO‐P comprises the elements C, N, O, and S (PROTAC consists of C, N, and O), consistent with its molecular structure, and exhibits the elemental attributes of both PROTAC and AA‐SS‐Ppa. Within 72 h, the particle size distribution index (PDI) of PRO‐P remained approximately 0.2 (Figure 3g), exhibiting favorable aqueous dispersity. The seven‐day stability study of PRO‐P in PBS and FBS demonstrated that its particle size increased to approximately 180 nm and 200 nm, respectively, while the PDI consistently remained within the low range of 0.35–0.40 (Figure S5). Compared to the free PROTAC, the self‐assembled PRO‐P nanoparticles demonstrated superior dispersity and storage stability, as evidenced by a low polydispersity index and highly consistent size distribution over time (Figure S6). These results indicate that the formulation maintains favorable colloidal dispersity across physiological simulation media of varying complexity, confirming its excellent environmental adaptability. In addition, the zeta potential analysis indicated that PROTAC, Ppa, and PRO‐P exhibited stable electronegative surfaces (Figure 3h; Figure S7), which is advantageous for prolonged in vivo delivery. The UV–vis absorption spectra indicated that PRO‐P exhibited enhanced characteristic absorption peaks at 420, 560, and 670 nm, aligning with the respective spectral characteristic absorption peaks of PROTAC and Ppa (Figure 3i), indicating their successful incorporation into nanoparticles. Finally, the responsive release of PRO‐P upon GSH exposure in a simulated tumor microenvironment was visualized by TEM and quantified by high performance liquid chromatography (HPLC) (Figure S8). Based on fluorescence emission spectrum of the free photosensitizer Ppa and its nano‐formulation PRO‐P, this study confirms the successful construction and functionality of PRO‐P from both structural and functional perspectives. Structurally, PRO‐P exhibited approximately 20% synchronous quenching in both excitation and emission intensity compared to Ppa, while the characteristic absorption and emission peaks remained unchanged. This indicates effective encapsulation of Ppa within the nanocarrier, with its chemical structure preserved despite fluorescence quenching due to microenvironmental changes. Functionally, methylene blue photodegradation assays demonstrated that PRO‐P generates reactive oxygen species (ROS) in a light‐ and concentration‐dependent manner, confirming retained photodynamic activity. In summary, PRO‐P achieves efficient loading and structural protection of the photosensitizer while maintaining its ROS‐generating capability, laying an experimental foundation for further investigation into its potential in photodynamic therapy and its synergistic mechanisms with other regulatory pathways such as PROTAC and ferroptosis. (Figure S9).

In vitro evaluation of blood compatibility revealed that all drugs, including various concentrations of PRO‐P nanoparticles, did not induce substantial hemolysis after co‐incubation with blood (hemolysis rate <5%) (Figure 3j,k). This suggests that PRO‐P exhibits favorable blood compatibility in a simulated physiological setting. In conclusion, PRO‐P was successfully developed as a self‐assembled PROTAC–photosensitizer nanoparticle with favorable physicochemical stability, high drug loading capacity, and excellent biocompatibility, enabling targeted HMGCR degradation and enhanced tumor‐selective delivery.

3.3. PRO‐P Enhances Cellular Uptake and Exhibits Potent Tumor Cell Inhibition After Laser Activation

Then, the cellular behaviors of PRO‐P nanoparticles, including the uptake efficiency of PRO‐P self‐assembled nanoparticles in 4T1 cells, were assessed by observing the autofluorescence of the photosensitizer Ppa. The fluorescent signals of PRO‐P exhibited a time‐dependent increase in intracellular accumulation, with faint signals observable at 8 h and stabilizing after 24 h (Figure 4a). Flow cytometry analysis demonstrated that the cellular uptake rate in the PRO‐P group attained 55.7% after 24 h of uptake, significantly surpassing the 41.4% observed in the Ppa group, indicating a 1.34‐fold enhancement (Figure 4b) and suggesting that PRO‐P nanoparticles exhibit superior cellular uptake characteristics. Building on the superior cellular uptake of PRO‐P, we subsequently employed confocal colocalization analysis with the lysosomal marker to evaluate its intracellular targeting and lysosomal escape capability. The imaging results revealed a dynamic process: in the initial stage (0‐8 h post‐treatment), PRO‐P (green) partially overlapped with lysosomes (red), indicating that a portion of nanoparticles were engulfed by lysosomes. From 12 h onward, a clear separation of the green signal from the red lysosomal compartments was observed. By 24–48 h, the green fluorescence became diffusely distributed throughout the cytoplasm, demonstrating successful escape from lysosomes. Quantitative colocalization analysis further supported this, showing a significant decrease in Pearson's coefficient from >0.8 at 8 h to <0.35 at 48 h. These findings confirm that PRO‐P can effectively escape lysosomal entrapment and deliver its payload to the cytoplasmic compartment (Figure S10).

FIGURE 4.

PRO‐P and statins related antitumor effects in vitro studies. (a) Confocal uptake images of 4T1 cells by PROTAC, Ppa, or PRO‐P, scale bar: 100 µm. (b) Flow cytometry uptake analysis of 4T1 cells by PROTAC, Ppa, or PRO‐P. (c) IC₅₀ of PRO‐P. (d) Flow cytometry analysis of the effects of different drug treatments on cell apoptosis. (e,f) Evaluation of drug cytotoxicity using Live‐Dead cell staining and (g) CCK‐8 assay in 4T1 cells, scale bar: 200 µm. (h,i) Scratch experiment analysis of the effect of different drug treatments in 4T1 cells, scale bar: 100 µm. Data are presented as mean ± SD (n = 3).

The in vitro cytotoxicity of PRO‐P against 4T1 cells was subsequently assessed in comparison with common statins (simvastatin, lovastatin) and Ppa. When combined with laser irradiation, the IC₅₀ of PRO‐P was measured at 0.2803 µg/mL (Figure 4c), demonstrating that its combination with photodynamic treatment (PDT) effectively eradicates tumor cells at minimal concentrations. Annexin V‐FITC/PI double‐staining flow cytometry analysis demonstrated (Figure 4d) that at an equivalent dosage, PROTAC exhibited cytotoxicity comparable to that of statins; the toxicity of the PRO‐P was marginally increased; and the apoptosis rate of the PROTAC‐Ppa‐Laser (PRO‐P‐L) group post‐laser irradiation reached 88.62%, which is 5 times greater than that of the PRO‐P group, thereby confirming the combined effect of laser. The outcomes of Live‐Dead cell staining and the CCK‐8 assay corroborated the flow cytometry findings, demonstrating that the PRO‐P‐L group markedly enhanced tumor cell mortality (Figure 4e–g). The scratch assay further indicated that PRO‐P‐L may effectively impede the migration of 4T1 cells (Figure 4h,i). In conclusion, PRO‐P nanoparticles, created from the self‐assembly of PROTAC with Ppa, markedly enhanced cellular uptake. Following combined laser irradiation, PRO‐P‐L demonstrated significant cytotoxicity and impeded tumor cell proliferation and migration.

3.4. PRO‐P Induces Enhanced Ferroptosis Through HMGCR Degradation, Mevalonate Pathway Blockade, and GPX4 Inhibition

Given the differential antitumor activities exhibited by the aforementioned experimental groups with various treatments, this study further investigates and confirms the inherent cytotoxic mechanisms of PRO‐P. Statins, as inhibitors of HMGCR, restrict cholesterol synthesis by obstructing the rate‐limiting enzyme of the mevalonate pathway, reducing the production of downstream metabolites, and resulting in the inactivation of GPX4 and CoQ10 [7]. Decreased GPX4 activity hinders lipid peroxide elimination, while inadequate CoQ10 diminishes antioxidant protection [27]. Accumulated lipid peroxides undergo oxidative damage via the Fenton reaction, leading to plasma membrane rupture and hallmark features of ferroptosis, including loss of mitochondrial cristae and increased membrane permeability [27]. Conventional statins, however, induce a compensatory upregulation of HMGCR, reducing efficacy and necessitating higher doses, thus increasing the risk of adverse effects [28]. Conversely, PROTAC molecules of PRO‐P facilitate the degradation of HMGCR via the UPS, thereby circumventing compensatory resistance (Figure 5a). As illustrated in Figure 5a, the prepared PRO‐P is anticipated to promote targeted protein degradation while circumventing compensatory feedback, thereby sabotaging the GPX4/CoQ10‐mediated antioxidant defense system via HMGCR disruption. This dual mechanism culminates in lethal lipid peroxidation and subsequent ferroptosis induction. To elucidate the tumoricidal mechanism of PRO‐P‐L, comprehensive biochemical and imaging analyses were conducted, revealing its capacity to degrade HMGCR, disrupt lipid peroxide detoxification, and trigger ferroptosis through GPX4 inhibition, GSH depletion, CoQ10 reduction, and mitochondrial impairment. The data of Western blotting confirmed that statins (lovastatin) induced a dose‐dependent upregulation of HMGCR (1.8‐fold increase at 5 µm). In comparison, statin‐based PROTAC facilitated its dose‐dependent degradation and time‐dependent degradation (>20% depletion at 5 µm). HMGCR expression was significantly increased in the statin group (lovastatin and simvastatin), with analogous levels in the other groups. PROTAC, PRO‐P, and PRO‐P‐L treatments reduced GPX4 expression (37%, 38%, and 70% reduction vs PBS, respectively), with PRO‐P‐L showing the most pronounced effect (Figure 5b–g; Figures S11 and S12). Confocal microscopy results corroborated GPX4 inhibition, exhibiting >90% depletion of fluorescence intensity in PRO‐P‐L group (Figure 5h,i).

FIGURE 5.

PRO‐P and statins targeting HMGCR to induce ferroptosis and antitumor effects in vitro studies. (a) Schematic illustration related to MVA and ferroptosis. (b–d) Expression of HMGCR with lovastatin and PROTAC, and expression of GPX4 with different drugs, as well as their (e–g) semi‐quantitative western blot analysis, respectively. (h,i) Confocal microscopy analysis of GPX4 protein expression after 4T1 treatment with different drugs. (j) GSH detection after treatment of 4T1 cells with different drugs. (k) CoQ10 synthesis in 4T1 cells treated with different drugs. (l,m) Confocal microscopy analysis of LPO accumulation after 4T1 treatment with different drugs. (n,o) Confocal microscopy analysis of mitochondrial membrane potential after 4T1 treatment with different drugs. (p,q) Confocal microscopy analysis of TFR‐1 after 4T1 treatment with different drugs. Data are presented as mean ± SD (n = 3), scale bar: 100 µm. Data are presented as mean ± SD (n = 3), scale bar: 100 µm.

In the tumor microenvironment, the disulfide bonds in PRO‐P were cleaved, resulting in the depletion of intracellular glutathione (GSH), which is similar to previously reported work [29]. GSH/GSSG assays revealed a 96% decrease in GSH levels in PRO‐P‐L‐treated cells (Figure 5j). ELISA demonstrated that both PROTAC and statins reduced CoQ10 production (30% and 28% reduction), with enhanced suppression (36% decrease) in laser‐irradiated PRO‐P‐L groups (Figure 5k). The link between GPX4 inhibition and ferroptosis was established by detecting LPO with a C11‐BODIPY probe, showing >90% elevated LPO fluorescence in PRO‐P‐L group that correlated with GPX4 suppression (Figure 5l,m). JC‐1 mitochondrial membrane potential assay showed that the PRO‐P‐L group led to a decrease in membrane potential, decreased red fluorescence, and increased green fluorescence, evidenced by >95% red/green fluorescence ratio reduction, indicating decreased mitochondrial function induced by the PRO‐P‐L group (Figure 5n,o). PRO‐P‐L treatment induced mitochondrial damage and ferroptosis, as evidenced by TEM observations of shrunken mitochondria with condensed membranes and reduced cristae, in contrast to the PBS group (Figure S13). The essential ferroptosis marker protein TFR‐1 showed significant fluorescence expression in the PRO‐P‐L group and was favorably linked with C11‐BODIPY signaling (Figure 5p,q), showing that enhanced iron intake facilitated the Fenton reaction and lipid peroxidation. The upregulation of TFR‐1 in tumor cells likely represents an adaptive response to lipid metabolic stress, aimed at increasing iron uptake to sustain essential enzymatic functions, mitochondrial activity, and proliferative capacity under conditions of disrupted lipid metabolic homeostasis.

In summary, to elucidate the mechanisms underlying the differences in antitumor activity among the various treatment groups, this study focused on clarifying the key pathway through which PRO‐P induces ferroptosis. As a PROTAC molecule, PRO‐P mediates the targeted degradation of HMGCR via the UPS, thereby effectively blocking the MVA pathway and overcoming the drug resistance associated with compensatory upregulation of HMGCR observed with conventional statins. Degradation of HMGCR leads to reduced synthesis of its downstream products, among which the activity of the antioxidant enzyme GPX4 is significantly impaired due to CoQ10 deficiency, severely compromising the cell's ability to clear lipid peroxides. Furthermore, when PRO‐P is delivered as the self‐assembled PRO‐P‐L and activated by laser irradiation, it exerts a complementary photodynamic effect. This photodynamic action synergizes with the PRO‐P‐induced metabolic blockade, accelerating lipid peroxidation and intensifying oxidative stress. Meanwhile, intracellular accumulation of lipid peroxides is further exacerbated, ultimately triggering hallmark morphological changes of ferroptosis, such as plasma membrane rupture and reduction of mitochondrial cristae. PRO‐P treatment first induces HMGCR degradation and lipid peroxide accumulation, with laser irradiation applied at the peak of peroxidation to amplify the ferroptotic cascade, followed by rapid loss of cell viability. This combined sequence aligns with the enhanced antitumor effects observed in vitro, supporting ferroptosis as the key mechanism through which PRO‐P, especially in its light‐activated form, exerts its therapeutic efficacy.

3.5. PRO‐P Induces Immunogenic Cell Death and Enhances Antitumor Immunity Through Photodynamic‐Ferroptosis Therapy and Lipid Metabolism Regulation

Beyond inducing ferroptosis, PRO‐P‐L also triggers robust immunogenic cell death (ICD), which is crucial for activating antitumor immunity. In addition, this work found that PRO‐P‐L, upon laser activation, functions as a dual‐action photoimmunotherapy (PIT) strategy by coupling PROTAC‐mediated HMGCR degradation with photodynamic therapy, thereby simultaneously inducing ICD and reprogramming lipid metabolism. As illustrated in Figure 6a, PRO‐P nanoparticles selectively accumulate in tumor cells, where laser activation triggers Ppa‐mediated ROS generation and PROTAC‐driven HMGCR degradation, converging on lipid peroxide accumulation and ICD induction. The release of DAMPs, including calreticulin exposure, ATP secretion, and HMGB1 release, subsequently activates dendritic cells, promotes cytotoxic T‐cell infiltration, and alleviates Treg‐mediated immunosuppression, culminating in a potent systemic antitumor immune response (Figure 6a). To experimentally validate these proposed mechanisms and sequential immunological events, a series of in vitro assays and clinical sample analyses were performed, including quantitative measurements of ROS generation, ATP depletion, CRT exposure, HMGB1 release, GO/KEGG pathway enrichment profiling, and immune co‐culture experiments.

FIGURE 6.

Laser‐activated PRO‐P‐associated antitumor immunotherapy in vitro studies. (a) Schematic representation of the cellular immunological response triggered by ICD. (b) Cell survival rate after treatment with PROTAC, Ppa, or PRO‐P with or without laser irradiation. (c) Flow cytometry analysis of ROS generated. (d,f) Confocal microscopy analysis of ROS generated. (e) Assessment of intracellular ATP levels. (g–j) Confocal microscopy analysis of CRT and HMGB1 was generated. (k,l) GO and KEGG enrichment analysis charts of the TCGA database related to TNBC. (m,n) Co‐culture of 4T1 cells with mouse spleen lymphocytes after treatment with different drugs, and population expression of mature DCs (CD11c⁺CD40⁺CD86⁺), (o,p) cytotoxic T cells (CD45⁺CD3⁺CD8⁺), and (q,r) Treg cells (CD3⁺CD4⁺CD25⁺). Data are presented as mean ± SD (n = 3), scale bar: 100 µm.

First, the in vitro cell killing studies revealed that the PRO‐P‐L group generated over 90% 4T1 cell mortality, representing a 3.2‐fold increase compared to the Ppa‐L group (Figure 6b), indicating a robust tumor cell inhibitory capacity. Following laser irradiation (with a laser of 660 nm, 100 mW/cm², 1 min), investigations by flow cytometry and laser confocal microscopy revealed that both Ppa‐L and PRO‐P‐L groups generated elevated amounts of ROS. The semi‐quantification using the DCFH‐DA probe indicated that their ROS levels were 7.26‐fold and 9.54‐fold greater than those of the control group, respectively (Figure 6c,d,f). However, no significant alteration was observed in the unirradiated group (‐L). It is worth noting that, similar to the conclusion of previous ICD‐related studies, ROS‐mediated cellular stress and apoptosis activate damage‐associated molecular patterns (DAMPs) [30, 31, 32]. A systematic study was then undertaken on the capacity of PRO‐P to cause ICD. The ATP assay revealed a notable decrease in intracellular ATP levels in the PRO‐P‐L group (Figure 6e). Confocal microscopy demonstrated that following laser irradiation, both Ppa and PRO‐P markedly enhanced the exposure of calreticulin (CRT) on the cell membrane surface (Figure 6g,h). Additionally, the laser‐treated groups, particularly the PRO‐P‐L group, exhibited a substantial release of high mobility group protein 1 (HMGB1) (Figure 6i,j). The data suggest that the ICD pathway initiated by PDT is a crucial mechanism for tumor eradication of the PRO‐P‐L. Given that ICD triggers systemic antitumor immunity, with the PRO‐P‐L regimen exhibiting superior ICD efficacy, we investigated HMGCR’s immunomodulatory role by integrating clinical dataset analyses with in vivo validation in murine models. In the process of clinical dataset analyses, GO/KEGG enrichment analyses (via R package “clusterProfiler”) revealed that HMGCR is critically involved in: (1) antigen presentation processes (MHC class II complex assembly, peptide antigen binding) and immune effector functions (B/T lymphocyte‐mediated immunity, immunoglobulin production); (2) immunological pathways (Th1/Th2 differentiation, NK cell‐mediated cytotoxicity, T‐cell receptor signaling) and lipid metabolic pathways (steroid biosynthesis, PPAR signaling). Collectively, HMGCR orchestrates immunometabolic crosstalk between cholesterol biosynthesis/PPAR signaling and immune activation (antigen presentation and T/NK cell engagement), defining a pivotal mechanism for PRO‐P‐L‐enhanced ICD potency (Figure 6k,l). Subsequent simulations at the cellular level delineated how HMGCR inhibition triggers ICD and enhances antitumor immunity. 4T1 cells were co‐cultured with splenic lymphocytes and analyzed via flow cytometry. Compared to the other groups, the PRO‐P‐L group exhibited a 4.3‐fold increase in the proportion of mature DCs (CD11c⁺ CD80⁺ CD86⁺) (Figures 6m,n); a 2.6‐fold rise in the proportion of infiltrating cytotoxic T cells (CD45⁺ CD3⁺ CD8⁺) (Figure 6o,p); while the proportion of immunosuppressive Treg cells (CD3⁺ CD4⁺ CD25⁺) diminished (Figure 6q,r). In conclusion, the PRO‐P‐L group markedly stimulated antitumor immune responses by inducing ICD in tumor cells, facilitating cytotoxic T‐cell infiltration and DC maturation, and mitigating the inhibitory effects of Treg cells, therefore augmenting the efficacy of immunotherapy.

The PRO‐P platform developed in this study features a core innovation: the integration of targeted protein degradation and photodynamic regulation into a self‐delivering, all‐in‐one nanosystem. Compared to most conventional PROTAC molecules or platform [19, 33], which primarily focus on achieving efficient intracellular target protein degradation but generally lack precise spatiotemporal control and systematic remodeling of downstream metabolic and immune pathways, PRO‐P enables light‐controlled activation. This not only allows “on‐demand” initiation of HMGCR degradation but, more critically, triggers the collapse of the mevalonate/CoQ10‐GPX4 axis, thereby redirecting the protein degradation event toward fundamental lipid metabolic reprogramming. Importantly, PRO‐P selects HMGCR as an intervention node not only because it is the rate‐limiting enzyme in cholesterol synthesis and thus central to metabolism, but also because it is an immunometabolic regulatory protein whose degradation can simultaneously affect the immunogenicity of tumor cells and the tumor microenvironment.

In the field of ferroptosis research, numerous strategies have been developed to directly induce or inhibit ferroptosis [34, 35]. The strategy of PRO‐P differs: rather than directly targeting classical ferroptosis pathway components, it leverages HMGCR—a hub protein with dual metabolic and immunoregulatory functions—as an actionable node. Through its degradation, PRO‐P simultaneously perturbs cellular lipid homeostasis and related immune signaling, thereby “indirectly” yet potently inducing lipid peroxidation and ferroptosis, while potentially synergistically enhancing antitumor immunity. This approach of “metabolic‐immune node intervention → downstream death mode output” provides a new perspective that combines precision and systematicity for inducing ferroptosis and modulating the immune microenvironment in tumors, particularly in metabolically active TNBC. Therefore, the mode of action of PRO‐P distinguishes it from either standalone PROTACs or ferroptosis inducers. It represents an exploration of a cascaded therapeutic paradigm of “protein degradation–metabolic/immune remodeling–cell fate control” for addressing complex biological processes.

3.6. PRO‐P Suppresses Tumor Growth In Vivo with High Safety and Prolonged Tumor Retention

The previously mentioned in vitro experiments definitively demonstrated that PRO‐P‐L reduced tumor cell resistance to ferroptosis by regulating lipid metabolism, while simultaneously improving the effectiveness of photodynamic therapy, resulting in a significant in vitro cytotoxic effect. To examine the in vivo antitumor effects of various drugs, a subcutaneous graft tumor model was established in the left lower abdomen of BALB/c mice, and treatment commenced when the tumor volume approximated 100 mm³ (Figure 7a). The in vivo antitumor efficacy and targeting characteristics of PRO‐P nanoparticles were comprehensively assessed. The spatial and temporal distribution of the therapeutic agent within tumor tissues was dynamically tracked using NIR fluorescence imaging. Following intravenous administration of an equivalent dosage (5 mg/kg of Ppa), the PRO‐P group demonstrated a considerable tumor‐specific accumulation within 36 h, with retention time significantly superior to that of the free Ppa group (Figure 7b). The in vitro NIR images demonstrated that the PRO‐P group exhibited prolonged circulation in vivo, with notable accumulation and substantial retention in the primary organs targeted by PRO‐P (liver > tumor > kidney) within 36 h of treatment, in contrast to the free Ppa group (Figure 7c; Figure S14). In summary, the observed in vivo and in vitro differences were primarily due to the passive targeting benefits of the PRO‐P, facilitated by the EPR effect [36], and the markedly improved colloidal stability.

FIGURE 7.

Laser‐activated PRO‐P exerts antitumor therapeutic effects in vivo studies. (a) Schematic diagram of treatment in 4T1 tumor‐bearing mice. (b) in vivo NIR fluorescence imaging of the breast cancer tumor model after intravenous administration of Ppa or PRO‐P. Data are presented as mean ± SD (n = 3). (c) Semi‐quantitative analysis of in vitro organs. (d) Tumor size curves of mice within 18 days of treatment. (e) Body weight curves of mice within 18 days of treatment. (f) Tumor image after 18 days of treatment. (g) Tumor weight of mice. (h) Tumor burden of mice. Data are presented as mean ± SD (n = 6). (i) H&E, Tunel, Ki67, and GPX4 staining of tumor tissues, scale bar: 100 µm.

Upon the tumor volume reaching approximately 100 mm³, intravenous injections (i.v.) (5 mg/kg) were followed by laser irradiation (100 mW/cm², 5 min) of the tumor site 24 h thereafter. Throughout the 18‐day observation period, the PRO‐P‐L group exhibited the most substantial tumor growth inhibition, characterized by a significant decrease in tumor growth rate and a tendency toward regression relative to the starting volume, with some tumors ultimately vanishing entirely. Conversely, the tumor volume in the PBS control group escalated roughly tenfold throughout the observation period. However, the PRO‐P‐L group consistently maintained the tumors at a reduced volume level (Figure 7d). Significantly, the mice across all experimental groups preserved normal body weight and exhibited no substantial toxic responses (Figure 7e). The mice were euthanized on day 18. Their tumors were excised for photography and weighing. The results above were corroborated by in vitro analysis of tumor tissues (Figure 7f–h), revealing that the average tumor volume and weight of the PRO‐P‐L group were significantly lower than those of the other administered groups. Collectively, these data indicate that PRO‐P significantly suppresses in situ tumor formation and development upon laser activation.

Histological examination of the tumor samples was then conducted (Figure 7i). The findings indicated that the PRO‐P‐L group exhibited pronounced apoptotic characteristics, with significantly compromised nuclear integrity and elevated TUNEL signal intensity compared to other groups. Concurrently, the expression of Ki67, a tumor proliferation marker, was diminished dramatically in this group, suggesting that its inhibitory effect on tumor cell proliferation was far more potent than that observed in the PRO‐P and Ppa‐L groups. The GPX4 protein level was markedly diminished in the PRO‐P‐L group relative to the control group, and the lack of this crucial antioxidant enzyme corroborated the particular activation of the ferroptosis pathway. Ultimately, the evaluation of drug safety indicated that blood biochemical analyses concerning liver and kidney function did not demonstrate significant adverse effects in all groups (Figure S15). Moreover, Hematoxylin‐eosin staining (H&E) of the principal organs (heart, liver, spleen, lungs, and kidneys) revealed no pathological alterations (Figure S16), suggesting favorable in vivo safety of the drugs at the experimental dosage. In conclusion, PDT in conjunction with PRO‐P markedly suppressed tumor development by simultaneously generating apoptosis and ferroptosis, offering an promising approach to enhance tumor treatment.

3.7. PRO‐P Reshapes the Tumor Microenvironment and Systemic Immunity by Enhancing Effector Cell Activity and Suppressing Immunosuppressive Populations

Immune cell infiltration, a key determinant of immune homeostasis, was evaluated following PRO‐P treatment. Building on prior evidence of its lipid metabolism regulation and antitumor activity, this study furtherly highlights the role of PRO‐P in modulating immune responses. Specifically, low HMGCR expression was associated with enhanced immunostimulatory effects, as validated through TCGA clinical analyses and in vivo BALB/c models. First, enrichment scores for 28 immune cell types were quantified by ssGSEA in the training cohort, and group differences were assessed by the Wilcoxon rank‐sum test. Differential analysis revealed nine immune cell types with significant changes between the groups as illustrated in Figure 8a, in which all were enriched in the low‐HMGCR cohort, including Activated CD8⁺ T cells, Central memory CD4⁺ T cells, Effector memory CD8 T cells, and Natural killer cells, suggesting that reduced HMGCR expression favors enhanced antitumor immunity. To further explore the relationship between HMGCR expression and these differentially infiltrating immune cells, Spearman correlation analysis was performed (Figure 8b). Using a significance threshold of p < 0.05, the infiltration levels of all 9 differential immune cells were significantly negatively correlated with HMGCR expression (p < 0.05). Among these, Activated CD8⁺ T cells showed the most significant negative correlation with HMGCR (Correlation coefficient = −0.286, p < 0.01). To validate these findings and evaluate the influence of HMGCR expression on the overall composition of the tumor microenvironment, this study utilized the ESTIMATE algorithm in the TCGA‐TNBC cohort to calculate the stromal score, immune score, and ESTIMATE score for each sample. Comparison between groups using the Wilcoxon rank‐sum test demonstrated (Figure 8c) that, compared to the HMGCR low‐expression group, patients in the high‐expression group exhibited significantly lower immune scores and ESTIMATE scores, alongside a considerably higher tumor purity score. This result suggests that high HMGCR expression is associated with reduced immune components and an increased proportion of tumor cells within the TME, potentially reflecting higher tumor malignancy. Finally, to elucidate the significance of HMGCR expression differences in the context of antitumor immune function, this study focused on the core framework of anticancer immune responses‐the Cancer‐Immunity Cycle. Leveraging the Tracking Tumor Immunophenotype (TIP) metaserver, we systematically evaluated the anticancer immune status across each of the seven steps of the Cancer‐Immunity Cycle based on RNA sequencing data. Comparison of TIP scores between the HMGCR high‐ and low‐expression groups (Wilcoxon rank‐sum test) revealed that the low‐expression group had significantly higher scores in the steps involving immune cell priming and activation (Step 3) and immune cell trafficking to tumors (Step 4) (Figure 8d). This advantage was particularly pronounced in immune cell subsets with direct cytotoxic functions, such as CD8⁺ T cells and NK cells.

FIGURE 8.

Laser‐activated PRO‐P demonstrates antitumor immune efficacy in vivo studies. (a) Variations in immune cell infiltration between groups with high and low expression of HMGCR. (b) Spearman analysis. (c) ESTIMATE analysis. (d) Variations in TIP scores across populations with high and low expression of HMGCR. (e) Flow cytometry detection of the effects of different drug treatments on cytotoxic T cells (CD8⁺ and CD4⁺) and (f) cytotoxic infiltrating lymphoid T cells (CD8⁺and Granzyme B⁺) generated in tumors. (g) Flow cytometry detection of the effects of different drug treatments on the production of Treg (CD25⁺) in tumors. (h) Flow cytometry detection of the effects of different drug treatments on NK cells (CD49b⁺ and CD3⁻) produced in tumors. (i) Flow cytometry detection of the effects of different drug treatments on cytotoxic T cells (CD8⁺) in the spleen. (j) Flow cytometry detection of the effects of different drug treatments on mature DCs (CD80⁺and CD86⁺) and (k) memory T cells (CD44⁺ and CD69L⁺) in the spleen. Data are presented as mean ± SD (n = 3).

Given the robust correlation between HMGCR expression levels and the immunosuppressive microenvironment identified above, along with prior experiments demonstrating that the laser‐activated PRO‐P‐L markedly inhibited tumor growth through HMGCR degradation, this study further employed flow cytometry to investigate its regulatory mechanisms within the TME and systemic immune organs. Analysis of tumor tissue revealed that the ratio of CD4⁺ CD8⁺ T cells was markedly elevated in the PRO‐P‐L group compared to both the PBS and conventional statin groups. In contrast, the PRO‐P and Ppa‐L groups exhibited a moderate rise (Figure 8e). Following the pattern of CD8⁺ T cell infiltration, the fraction of Granzyme B⁺ CD8⁺ T cells, indicative of the activating cytotoxic effect, was highest in the tumor of the PRO‐P‐L group compared to the other groups (Figure 8f). Simultaneously, the percentage of Treg cells exhibiting immunosuppressive function (CD25⁺) was diminished in the PRO‐P‐L group compared to the other groups (Figure 8g). The PRO‐P‐L group increased the percentage of NK cells (CD49b⁺ CD3⁻) in the tumor associated with natural immunity (Figure 8h). The systemic immune activation effect of PRO‐P‐L was further validated through flow analysis of spleen tissues: the proportion of splenic CD8⁺ T cells was markedly elevated compared to the control group (Figure 8i); the proportion of mature DCs (CD80⁺ CD86⁺), essential antigen‐presenting cells (APCs), was also significantly augmented in the spleens of the PRO‐P‐L group (Figure 8j), which is vital for initiating the adaptive immune response. Moreover, the PRO‐P‐L group significantly facilitated the proliferation of splenic central memory T cells (CD44⁺ CD62L⁺) (Figure 8k), thereby establishing a robust systemic antitumor immunological memory. PRO‐P‐L collectively elicited a strong antitumor immune response by simultaneously altering the cellular composition of the tumor microenvironment and systemic immune organs, enhancing effector T cell and NK cell cytotoxicity, inhibiting Treg cell function, facilitating DCs maturation, and inducing central memory T cell development.

3.8. PRO‐P Suppresses Primary and Metastatic Tumor Progression

Then, further flow cytometry data show that laser‐activated PRO‐P‐L restructured the tumor microenvironment, markedly stimulated cytotoxic T cells, and elicited enduring antigen‐specific immunological memory. This study developed a bilateral tumor model to examine the inhibitory effect of systemic antitumor immunity on distant tumors. The 4T1 primary tumor was established in the left groin of BALB/c mice, and the second inoculation on the right side simulated distal metastatic foci post‐treatment (Figure 9a). Continuous monitoring demonstrated substantial growth suppression of both primary and unirradiated distal foci in the PRO‐P‐L group, with unirradiated distal foci showing a 79% decrease compared to the control group (Figure 9b,c). This dual inhibition effect was significantly more substantial than that observed in either the radiation‐only or drug‐only groups (p < 0.001), underscoring the synergistic efficacy of the PRO‐P‐L combinational therapy. This dual inhibition effect was significantly more substantial than that observed in either the radiation‐only or drug‐only groups (p < 0.001), underscoring the synergistic efficacy of the PRO‐P‐L combinational therapy. Nevertheless, the mice in all groups maintained consistent body weight without exhibiting significant toxic effects (Figure 9d).

FIGURE 9.

Laser‐activated PRO‐P demonstrates anti‐metastatic ability in vivo studies. (a) Schematic diagram of distant tumor inoculation. (b) Distant tumor image after 17 days of treatment. (c) Distant tumor size curves of mice. (d) Weight of mice. Data are presented as mean ± SD (n = 4). (e) Schematic diagram of lung metastasis inoculation. (f) Baseline image of PET/MRI. (g‐h) PET/MRI images of whole body, in situ tumors, and lung metastases in PBS and PRO‐P‐L groups of mice. (i) SUV_bw calculation formula. (j) Quantification of TLG and MTV in situ tumors and lungs of mice in PBS and PRO‐P‐L groups. (k) Lung images and H&E after different drug treatments. Data are presented as mean ± SD (n = 3), scale bar: 100 µm.

Because advanced TNBC is frequently associated with lung metastasis, a metastatic model was generated through tail vein injection of 1 × 10⁶ 4T1 cells (Figure 9e). This model was subsequently used to evaluate the antimetastatic potential of PRO‐P. In addition, the tumor metastasis and metabolism were quantitatively evaluated using FDG‐PET/MRI, characterized explicitly by TLG and MTV. Baseline imaging (primary tumor ≈100 mm³) demonstrated elevated fluorodeoxyglucose (FDG) uptake in the primary foci (Figure 9f); a 12‐day assessment of the metastatic cell inoculation indicated that the primary foci in the PBS group exhibited a distinctive “ring” uptake pattern attributable to central necrosis resulting from proliferation exceeding vascular supply capacity, with peripheral survivors displaying high uptake and central necrotic regions showing low uptake. In contrast, tumors in the PRO‐P‐L group were considerably smaller, stable, and exhibited a mildly uniform uptake. This finding indicated significantly suppressed proliferation of breast cancer lung metastatic lesions in the PRO‐P‐L treatment group, while the PBS group exhibited metastatic nodules exceeding 3 mm with localized radio‐concentration. In contrast, the PRO‐P‐L group displayed no discernible metastatic nodules and a uniform physiological backdrop (Figure 9g–j; Figure S17). Post‐scan, H&E slices were extracted from the pulmonary tissues of mice (Figure S18). In vitro validation conducted 17 days post‐inoculation revealed that the lung surfaces of the PBS and statin groups were extensively populated with metastatic nodules. But the PRO‐P‐L group exhibited minimal or no metastatic nodules. H&E staining confirmed that the PRO‐P‐L group maintained a preserved lung tissue architecture and demonstrated the least tumor infiltration (Figure 9k). The findings indicate that PRO‐P‐L proficiently suppresses the proliferation and metastatic dissemination of distant tumors by restructuring critical subpopulations within the tumor immune milieu and creating a systemic immunological barrier via PROTAC‐mediated degradation of HMGCR.

3.9. Quantitative Proteomics Provides a Comprehensive Mechanistic Landscape of Lipid Metabolism Regulation and In Vivo Antitumor Activity of PRO‐P‐L

High‐throughput quantitative proteomic analysis was performed to further understand the therapeutic mechanisms by comparing PRO‐P‐L to PBS group. Totally, 9,627 proteins were quantified from two groups of tumor tissues in our dataset acquired by nanoLC‐MS/MS system using data‐independent acquisition (DIA) mode. By principal component analysis (PCA), these two groups of proteomics datasets were separated clearly (Figure 10a), showing the significantly differential landscape of proteins expressed in tumor tissue during the PRO‐P‐L therapeutic strategy. In total, we found that 354 differentially expressed proteins (DEPs) (FC>2, p<0.05) between above two groups (Figure 10b), including 113 upregulated and 241 downregulated species in PRO‐P‐L group compared to PBS group. By clustering the 354 DEPs, it showed similar expression levels among biological replicates in each group demonstrating the high data quality, and distinctive levels between PRO‐P‐L and PBS groups demonstrating PRO‐P‐L strategy promotes differential expression of proteins (Figure 10c), revealing the consistency with PCA of comparative proteomics in Figure 10a. Based on GO enrichment analysis in the term of biological process using the DEPs, we found several pathways that are closely related to the antitumor effects (Figure 10d), and individual terms were further profiled in Figure 10e. Focusing on the downregulated pathways, the one, generation of precursor metabolites and energy, demonstrated the similar therapeutic effects with reporting decreased pathway in breast cancer gene expression profiles related with neoadjuvant therapy with letrozole [37], in which the disrupted lipid homeostasis as demonstrated above might mediate the downregulated generation of energy [38] (Figure 10e). In addition, disrupted lipid homeostasis were also demonstrated by the Gene Set Enrichment Analysis (GSEA) of expressed protein landscapes between PRO‐P‐L and PBS groups as negative enrichment scores, including the pathways such as glycerolipid metabolic process and phospholipid dephosphorylation (Figure S19). The downregulation of mitochondrial respiratory chain complex assembly would affect the mitochondrial energy metabolism that is consistent with the decreased energy generation and further inhibit the tumor progression [39]. The downregulated proteins were also enriched in oxidative phosphorylation (OXPHOS) and negative regulation of response to reactive oxygen species that are also related to the decreased generation of energy. On the contrary, we found the upregulated proteins that are enriched in response to hydrogen peroxide, response to metal ion and T cell proliferation after PRO‐P‐L treatment. Hydrogen peroxide could be generated by numerous mechanisms, such as photodynamic therapy [40, 41], and it has been reported that moderate concentrations of H₂O₂ induce apoptosis, while higher concentrations induce necrosis [42], which could be promoted by the upregulated response to hydrogen peroxide. It is reported that radiation could induce high intracellular concentration of metal ions and cause ionic overloading that might binds to inappropriate receptor sites. This process might upregulate the response to metal ion and further lead to signaling disorders, toxic effects and cell death [43]. As demonstrated above, ratio of CD4⁺ CD8⁺ T cells was markedly elevated in the PRO‐P‐L group, which is also verified by the upregulated T cell proliferation in proteomics dataset.

FIGURE 10.

Comparative proteomics between PRO‐P‐L and PBS groups. (a) Principal component analysis (PCA) of the quantitative proteomics (n = 4). (b) The volcano plot of differentially expressed proteins (DEPs) between PRO‐P‐L and PBS groups. (c) The heatmap of significant differential proteins (354 species). (d) GO enrichment of DEPs in the term of biological process (BP). (e) Chord diagram of the enriched BP terms with the associated DEPs. (f) The box plots of the differentially expressed proteins between PRO‐P‐L and PBS groups.

In particular, we also found lots of functional proteins that have been demonstrated to inhibit the tumor progress previously (Figure 10f). For examples, downregulated Atox1 was reported to reduce the migration of MDA‐MB‐231 cells [44]; PRO‐P‐L treatment resulted into a 3.47‐fold decreasing of Hmga2, whose downregulation was demonstrated to suppress breast cancer cell growth previously [45, 46]; as mentioned above, Mki67, whose downregulation showed the inhibited cell proliferation, was significantly downregulated in PRO‐P‐L group; in addition, it is speculated that Wtap promotes breast cancer growth, while we found the downregulation of Watp in the tumor tissues after PRO‐P‐L therapy [47]; Thus, we conclude that PRO‐P‐L therapeutic strategy is potentially efficient to inhibit breast cancer.

Collectively, this study introduces a multi‐omics‐guided, self‐delivering PROTAC‐photosensitizer nanoplatform that couples laser‐gated HMGCR degradation with mevalonate/CoQ10‐GPX4 axis collapse to reprogram lipid metabolism toward ferroptosis and augment tumor‐confined photoimmunotherapy in triple‐negative breast cancer. Nevertheless, our work has several limitations, including reliance on subcutaneous 4T1 models, use of 660 nm excitation with limited tissue penetration, and the lack of comprehensive pharmacokinetic, off‐tumor HMGCR depletion, and long‐term safety evaluations in larger or orthotopic models. From a translational perspective, advancing PRO‐P‐like platforms will require clinically compatible irradiation strategies (e.g., NIR‐I/II or interstitial delivery), scalable manufacturing, and Good Laboratory Practice (GLP)‐grade evaluation of pharmacokinetics, biodistribution, immunogenicity, long‐term biosafety, and off‐tumor HMGCR depletion. In parallel, regulatory and clinical development will depend on validation in orthotopic and patient‐derived TNBC models, deployment of deeper‐penetrating or X‐ray‐activated PDT, and integration of PROTAC‐based photoimmunotherapy with approved regimens while carefully monitoring target selectivity, chronic toxicity, and long‐term immune modulation.

4. Conclusion

In conclusion, this study identifies HMGCR as a lipid‐metabolic and immunologic chokepoint in triple‐negative breast cancer through bulk‐ and single‐cell transcriptomics, then neutralizes it with PRO‐P, a 170 nm self‐assembled nanomedicine that fuses an HMGCR‐targeting PROTAC and a Ppa photosensitizer. Laser‐activated PRO‐P achieves catalytic degradation of HMGCR, collapses the mevalonate/CoQ10‐GPX4 axis, and redirects lipids into ferroptosis while simultaneously supplying photodynamic ROS. The dual modality delivers 0.28 µg/mL IC₅₀ in vitro, 92.5% tumor inhibition, and complete metastatic clearance in vivo, with negligible systemic toxicity. Immune profiling confirms durable expansion of cytotoxic T cells and dendritic‐cell maturation alongside T‐reg suppression. Collectively, PRO‐P converts lipid addiction into a redox‐immunologic vulnerability and offers a low‐toxicity, translatable strategy for refractory TNBC and other lipid‐driven malignancies.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: advs73441‐sup‐0001‐SuppMat.docx.

Data Availability Statement

These data were derived from the following resources available in the public domain: Gene Expression Omnibus database, https://www.ncbi.nlm.nih.gov/geo/ and Cancer Genome Atlas (TCGA) database, https://portal.gdc.cancer.gov/.