Biomimetic nanomodulator reprograms glycolysis-driven immunosuppressive microenvironment to potentiate photothermal immunotherapy in cold tumors

Abstract

Aberrant glycolytic metabolic reprogramming and intrinsic low immunogenicity in “cold” tumors create a hostile immunosuppressive tumor microenvironment (ITME), severely impairing antitumor immunity. Herein, we designed a biomimetic dual-targeting nanomodulator (CM-cRGD@PBG), using mesoporous polydopamine (mPDA) nanoparticles camouflaged with cRGD-modified cancer cell membranes. This multifunctional nanomedicine synchronizes metabolic intervention with cellular immune remodeling to dismantle the ITME. Specifically, the delivered GLUT1 inhibitor BAY-876 reverses lactate-driven immunosuppression by inhibiting M2-like macrophage polarization and regulatory T cell (Treg) expansion. Complementarily, gemcitabine (GEM) is repurposed as a potent immunomodulatory agent to selectively deplete myeloid-derived suppressor cells (MDSCs) and Tregs. Notably, mPDA-mediated photothermal therapy elicits immunogenic cell death (ICD), acting as an in situ nanovaccine that facilitates dendritic cell maturation and antigen presentation. Pancreatic and lung cancer mouse models demonstrate that this “metabolism-chemotherapy-photothermal” strategy exhibits excellent tumor growth inhibition and successfully converts “cold” tumors into “hot” ones, significantly enhancing cytotoxic T lymphocyte (CTL) infiltration and effector function. Our study offers a universal metabolic and immune modulation strategy for “cold” tumors.

Graphical abstract

CM-cRGD@PBG efficiently targets and accumulates at the tumor site via homologous and active targeting. After cellular internalization, the released BAY-876 and GEM remodel the immunosuppressive tumor microenvironment (ITME) by redistributing glucose, inhibiting lactate production, and depleting immunosuppressive cells. Simultaneously, PTT induces apoptosis and triggers ICD to amplify tumor immunogenicity. This strategy reprograms the ITME from an immunosuppressive to an immune-active phenotype, thereby robustly potentiating CTL infiltration and cytotoxicity.

Highlights

• •A biomimetic nanomodulator co-loaded with BAY-876 and gemcitabine (CM-cRGD@PBG) was developed.
• •CM-cRGD@PBG exhibits excellent tumor-targeting ability and photothermal performance.
• •Enhancing in situ vaccination efficacy of PTT via glycolysis inhibition and immune modulation.
• •Under laser irradiation, CM-cRGD@PBG inhibits tumor growth via the synergistic “metabolic-chemo-photothermal” strategy.
• •The strategy reverses immunosuppressive tumor microenvironment and reinvigorates the antitumor immune responses against “cold” tumors.

1. Introduction

Immunotherapy has emerged as a transformative therapeutic paradigm for malignancies. However, its clinical efficacy remains limited in “cold” tumors characterized by “immune-desert” or “immune-excluded” phenotypes. This resistance is attributed to multiple immune evasion mechanisms, which include not only inherently low tumor mutational burden and defective antigen presentation [1,2], but also complex interactions within the tumor microenvironment (TME) that establish an immunosuppressive landscape, severely hampering the efficacy of clinical intervention [3,4]. Specifically, the substantial accumulation of immunosuppressive cell populations, including regulatory T cells (Tregs) [5], myeloid-derived suppressor cells (MDSCs) [6], and tumor-associated macrophages (TAMs) [7], cooperatively fosters an immunosuppressive niche within the TME, thereby promoting tumor growth and suppressing the effector functions of immune cells [8]. Moreover, metabolic reprogramming of diverse nutrients in tumor cells, along with the aberrant accumulation of toxic metabolic byproducts, also constitutes a critical driver of an immunosuppressive tumor microenvironment (ITME) and immunotherapy resistance [9,10]. Currently, there is significant interest in developing therapeutic approaches to overcome these immunosuppressive barriers and convert “cold” tumors into “hot” ones. The ultimate aim of such strategies is to remodel the tumor immune microenvironment to improve clinical outcomes.

Recent studies have demonstrated that the metabolic dysregulation, particularly aberrant aerobic glycolysis, is a critical driver in shaping and maintaining the immunosuppressive microenvironment and immune resistance [11,12]. To sustain rapid proliferation, many tumor cells undergo metabolic reprogramming to rely predominantly on glycolysis, a phenomenon that persists even in oxygen-rich environments [13]. This metabolic shift leads to glucose deprivation and lactate accumulation, thereby promoting M2-like TAM polarization and Treg expansion and differentiation, while severely impairing the infiltration and effector functions of cytotoxic T lymphocytes (CTLs), as well as the maturation of dendritic cells (DCs) [14,15]. Therefore, targeting glycolysis has emerged as a promising therapeutic strategy for “cold” tumors, aiming to reverse the immunosuppressive microenvironment [16].

Currently, increasing research is exploring tumor treatment strategies targeting pivotal glycolytic enzymes [17,18], yet challenges in drug specificity and metabolic plasticity severely limit their clinical translation [19]. Furthermore, “starvation therapy” based on glucose oxidase (GOx), which directly catalyzes glucose consumption, is limited in solid tumor treatment due to hypoxia and poor enzyme stability [20,21]. To address these constraints, directly inhibiting glucose transporter 1 (GLUT1) to block substrate supply at the source offers a more efficient strategy for inhibiting tumor glycolysis. Existing evidence indicates that GLUT1 is highly expressed in aggressive malignant tumors and correlates with chemotherapy resistance and poor clinical outcomes, suggesting its potential as a vital therapeutic target [22,23].

While metabolic intervention is a key factor in remodeling the tumor immune microenvironment, it alone is often insufficient to completely break down the immunosuppressive barriers in “cold” tumors. Additionally, triggering a robust antitumor immune response is often impeded by the inherently low immunogenicity of tumors. Thus, inducing immunogenic cell death (ICD) offers a strategic approach to reinvigorate antitumor immunity. The process of ICD leads to the release of tumor-associated antigens (TAAs) accompanied by damage-associated molecular patterns (DAMPs), including high-mobility group box 1 (HMGB1), calreticulin (CRT), and heat shock proteins (HSPs) [24]. These “danger signals” are critical for coordinating DC maturation and antigen presentation, ultimately eliciting effective immune responses such as CTL activation and a proinflammatory cytokine cascade [25]. Photothermal therapy (PTT) effectively triggers such ICD by utilizing photothermal agents to convert near-infrared (NIR) irradiation into localized hyperthermia [26,27]. However, the excessive inflammation post-PTT induces intratumoral recruitment of innate immunosuppressive MDSCs, thereby sustaining the inflammatory ITME [28]. Notably, low-dose GEM has been repurposed for its immunomodulatory potential to selectively deplete MDSCs and Tregs within the TME [29,30]. This chemo-immunomodulation strategy may effectively overcome the immunosuppressive barrier, facilitating efficient PTT-triggered CTL infiltration.

To effectively treat “cold” tumors, developing a combinatorial strategy that integrates metabolic remodeling, immunomodulation, and immunogenic activation is a novel approach. Emerging nanotechnology offers a versatile “all-in-one” platform to realize the “metabolic-chemo-photothermal” strategy. Mesoporous polydopamine (mPDA) is an ideal therapeutic nanocarrier owing to its excellent biocompatibility, remarkable payload capacity, and innate photothermal properties [31,32]. Furthermore, cell-based biomimetic modification has emerged as a promising paradigm for optimizing drug delivery, effectively enabling nanomedicines to evade mononuclear phagocyte system (MPS) clearance and yield superior site-specific accumulation [33]. Cell membrane-camouflaged nanoparticles can inherit the intrinsic biological properties of their parental cells. For instance, cancer cell membrane coatings enable nanomedicines to evade immune clearance and target homologous antigens [34]. Furthermore, surface-targeting modification with cRGD peptides can facilitate the delivery and deep penetration of nanomedicines within tumors by specifically recognizing α_νβ₃ integrins overexpressed on tumor neovasculature and cells [35,36], thereby improving drug efficacy and reducing systemic toxicity.

Herein, we developed a biomimetic nanomodulator to reverse severe immunosuppression and reinvigorate the antitumor immune response against “cold” tumors. As illustrated in Scheme 1, the nanomodulator was constructed by loading GLUT1-specific inhibitor BAY-876 and chemotherapeutic agent GEM into mesoporous polydopamine nanoparticles (mPDA NPs). Coating mPDA-BAY-876-GEM (denoted as PBG) with cRGD-modified cancer cell membrane yielded CM-cRGD@mPDA-BAY-876-GEM (denoted as CM-cRGD@PBG), which integrates three functional components: a) Dual-targeting drug co-delivery, specifically recognizing tumors via homologous and active targeting to improve drug efficacy and reduce systemic toxicity. b) Metabolic and chemo-immunomodulation, restoring glucose availability to immune cells and mitigating lactate-driven TAM polarization and Treg differentiation with BAY-876, while selectively depleting MDSCs and Tregs with GEM. c) Enhanced tumor immunogenicity by functioning as an in situ nanovaccine, inducing ICD, promoting DC maturation, and triggering potent immune responses via mPDA-mediated PTT. Taken together, the synergistic “metabolic-chemo-photothermal” strategy can trigger profound antitumor immunity to impede tumor growth. Therefore, the nanomodulator, acting as a novel immunomodulatory adjuvant, offers a new paradigm for converting immunosuppressive “cold” tumors into immune-active “hot” tumors, thereby potentiating antitumor efficacy.

Scheme 1.

Schematic illustration of the biomimetic nanomodulator-mediated synergistic “metabolic-chemo-photothermal” therapy against “cold” tumors. (A) Preparation of CM-cRGD@PBG. cRGD-modified cancer cell membranes (CM) are coated onto mPDA nanoparticles co-loaded with BAY-876 and gemcitabine (GEM). (B) Mechanism of synergistic antitumor efficacy. CM-cRGD@PBG efficiently targets and accumulates at the tumor site via homologous and active targeting. After cellular internalization, the released BAY-876 and GEM remodel the immunosuppressive tumor microenvironment (ITME) by redistributing glucose, inhibiting lactate production, and reducing immunosuppressive cell populations. Simultaneously, photothermal therapy induces apoptosis and triggers immunogenic cell death to amplify tumor immunogenicity. This strategy reprograms the TME from an immunosuppressive to an immune-active phenotype, thereby robustly potentiating cytotoxic T lymphocyte infiltration and cytotoxicity.

2. Materials and methods

2.1. Chemicals and reagents

BAY-876 and gemcitabine (GEM) were purchased from MedChemExpress (Shanghai, China). Dopamine hydrochloride and ammonia solution were obtained from Aladdin (Shanghai, China). Pluronic F127 and γ-H2AX Immunofluorescence Kit were purchased from Beyotime Biotechnology (Shanghai, China). DSPE-PEG2000-cRGDyk was obtained from Ruixi Biotechnology (Xi'an, China). 1,3,5-trimethylbenzene (TMB) was purchased from Thermo Fisher Scientific (Waltham, USA).

2.2. Bioinformatics analysis of SLC2A1 in cancers

The Pan-cancer mRNA expression profile of SLC2A1 (encoding GLUT1) was assessed using the Tumor Immune Estimation Resource (TIMER) 2.0 database (http://timer.cistrome.org/). Differential expression between tumor and adjacent normal tissues was compared across The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) datasets, with values presented as log₂ transcripts per million (TPM). Additionally, the Gene Expression Profiling Interactive Analysis (GEPIA) 2.0 database was utilized to evaluate the correlation between SLC2A1 expression and overall survival (OS) in patients with pancreatic adenocarcinoma (PAAD) and non-small cell lung cancer (NSCLC).

2.3. γ-H2AX immunofluorescence

DNA double-strand breaks were assessed by detecting γ-H2AX foci via immunofluorescence staining. Following exposure to various treatments for 72 h, Panc02 cells were fixed with 4% paraformaldehyde (PFA) and blocked for 15 min. Subsequently, γ-H2AX staining was conducted using a DNA Damage Assay Kit. The resulting fluorescence images were visualized using a confocal laser scanning microscope (CLSM, PerkinElmer ULTRAVIEW VOX, USA).

2.4. Preparation of CM-cRGD@PBG

To construct the co-delivery system (designated as PBG), mPDA nanoparticles were loaded with BAY-876 and GEM. Briefly, BAY-876 and GEM (both at 20 mg/mL in DMSO) were encapsulated within mPDA via continuous stirring for 48 h. The mixture was then centrifuged to remove free drugs. Meanwhile, cancer cell membranes (CM) were isolated from harvested tumor cells. Subsequently, the CM was incubated with DSPE-PEG2000-cRGDyk at 37 °C for 1 h to obtain cRGD-functionalized membranes (CM-cRGD). Finally, CM-cRGD was mixed with PBG nanoparticles, followed by sonication and extrusion to yield the cRGD-modified membrane-coated nanoparticles, referred to as CM-cRGD@PBG.

2.5. Colocalization analysis

To verify the integrity of the membrane coating, a colocalization assay was performed. CM-cRGD and mPDA nanoparticles were labeled with DiD (red) and DiO (green), respectively. Subsequently, the labeled components were mixed and subjected to sonication and extrusion to facilitate the coating process. Finally, the resulting nanoparticles were visualized to evaluate the colocalization of the DiD (shell) and DiO (core) signals using CLSM (PerkinElmer ULTRAVIEW VOX, USA).

2.6. SDS-PAGE analysis

To verify the successful coating of CM-cRGD on the nanoparticles, protein profiles were analyzed by SDS-PAGE. Samples including mPDA, PBG, CM-cRGD@PBG, CM and cancer cell lysates were prepared and separated on a 10% polyacrylamide gel. The gel was subsequently stained using M5 HiPer Ultrafast Protein Staining Solution. Finally, the protein bands were visualized to confirm that the biomimetic nanoparticles retained the characteristic protein profile of the source cell membranes.

2.7. In vitro cellular uptake and targeting evaluation

To evaluate cellular uptake efficiency, Panc02 cells were incubated with free DiO and DiO-labeled mPDA, CM@mPDA, or CM-cRGD@mPDA for 2 h. After washing with PBS, intracellular fluorescence was observed using CLSM. To investigate time-dependent uptake, cells were treated with DiO-labeled mPDA or CM-cRGD@mPDA for 1, 2, and 4 h, followed by CLSM imaging and flow cytometric analysis. To further investigate CM-cRGD-mediated targeting, NIH3T3 cells (normal control) and Panc02 cells were incubated with DiO-labeled CM-cRGD@mPDA. After incubation for indicated time periods (0, 2, and 4 h), the resulting intracellular fluorescence signals were visualized by CLSM and quantified by flow cytometry to compare the uptake efficiency between the two cell lines.

Additionally, to verify the broad applicability of the targeting strategy, mPDA cores and cRGD-modified Lewis lung carcinoma (LLC) cell membranes (CM-cRGD) were labeled with DiO (green) and DiI (red), respectively. The resulting nanoparticles were incubated with LLC cells for 8 h. Finally, the internalization and intracellular colocalization of the membrane and core signals were analyzed via CLSM.

2.8. In vivo antitumor efficacy

To assess therapeutic efficacy, subcutaneous Panc02 (4 × 10⁵ cells) or LLC (6 × 10⁵ cells) tumor models were established in C57BL/6 J mice. Once tumors reached 30-50 mm³, mice were randomized into eight groups (n = 5 or 6 per group): (Ⅰ) PBS, (Ⅱ) mPDA, (Ⅲ) BAY-876 + GEM, (Ⅳ) PBG, (Ⅴ) CM-cRGD@PBG, (Ⅵ) L (Laser only), (Ⅶ) mPDA + L, and (Ⅷ) CM-cRGD@PBG + L. Nanomedicines were administered intravenously (i.v.) every 3 days. In the PTT groups, 808 nm laser irradiation (1.5 W/cm², 5 min) was performed 8 h after injection. Body weight and tumor volume were monitored every 2 days. Tumor volume was calculated using the formula: V = 0.5 × length × width². At the experimental endpoint, tumors were dissected, photographed, and weighed.

2.9. Histological and immunohistochemical analysis

Hematoxylin and eosin (H&E) staining was performed on paraffin-embedded sections (5 μm thickness) of excised tumors and major organs to assess pathological changes, tumor necrosis, and systemic biosafety.

For immunohistochemical (IHC) analysis, tumor sections were processed for antigen retrieval and incubated with primary antibodies against CD8α, Granzyme B, CD86, CD206, and Foxp3. Subsequently, the sections were incubated with HRP-conjugated secondary antibodies and visualized using a 3,3′-diaminobenzidine (DAB) substrate. The immunostained sections were observed and imaged using a light microscope (Leica, Germany).

2.10. In vivo biosafety assessment

At the experimental endpoint, whole blood was collected from the control and CM-cRGD@PBG + L groups for hematological analysis. Furthermore, serum was separated from different groups to examine the levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST) and blood urea nitrogen (BUN) to evaluate hepatic and renal functions.

2.11. Statistical analysis

All data are expressed as mean ± SD. Differences between two groups were evaluated by unpaired two-tailed Student's t-tests, whereas multiple group comparisons were analyzed using one-way ANOVA followed by Tukey's post hoc test. Statistical significance was defined as p< 0.05 (∗p< 0.05, ∗∗p< 0.01, ∗∗∗p< 0.001, and ∗∗∗∗p < 0.0001; n. s., not significant).

3. Results and discussion

3.1. SLC2A1 overexpression correlates with poor prognosis in PAAD and NSCLC, and its inhibitor BAY-876 enhances the antitumor effect of GEM

To evaluate the clinical relevance of SLC2A1 (encoding GLUT1), its expression profile was analyzed based on datasets from the TCGA and GTEx databases. SLC2A1 is aberrantly overexpressed in multiple cancer types (Fig. 1A). Notably, in PAAD and NSCLC cohorts, SLC2A1 showed significant upregulation in tumors relative to normal tissues (Fig. 1B and C). Moreover, its expression level was negatively correlated with overall survival (Fig. 1D and E), suggesting that GLUT1 is vital for facilitating massive glucose uptake to sustain malignant proliferation. Previous studies have shown that this glycolysis dependence not only leads to treatment resistance but also creates an ITME characterized by glucose deficiency and lactate accumulation [11,37,38]. Taken together, these findings indicate that SLC2A1 could be a promising therapeutic target for low-immunogenicity “cold” malignancies, particularly PAAD and a large subset of NSCLCs. Consistent with the bioinformatics findings, immunohistochemical analysis validated the significant overexpression of GLUT1 in pancreatic tumor samples compared with adjacent normal tissues (Fig. 1F). Furthermore, using immunofluorescence and IHC, we observed high expression levels of GLUT1 on the cell membranes of both the murine pancreatic cancer cell line Panc02 and its subcutaneous tumor tissues (Fig. S1A and B). Thus, Panc02 can serve as a suitable cellular model for our subsequent study on targeting GLUT1-mediated metabolic reprogramming. GEM serves as the standard first-line therapeutic regimen for pancreatic cancer and advanced lung cancer. However, its clinical efficacy is often severely limited by drug resistance and dose-limiting systemic toxicity. Because aberrant glycolysis fuels tumor growth and undermines chemotherapy by providing energy and metabolic substrates for DNA repair [39,40], we hypothesized that targeting GLUT1 could metabolically sensitize tumor cells to GEM treatment, thereby enhancing its antitumor efficacy and minimizing systemic toxicity.

Fig. 1.

Bioinformatics analysis of SLC2A1 in pancreatic and lung cancers, with in vitro evaluation of the synergistic antitumor effects of BAY-876 and gemcitabine (GEM). (A) The mRNA expression profiles of SLC2A1 in tumor (red) versus normal (blue) tissues across different cancer types from the TIMER 2.0 database. (B) SLC2A1 expression levels in PAAD tumor samples (n = 183) compared to normal tissue controls (n = 164). (C) SLC2A1 expression levels in NSCLC tumor samples (n = 1122) compared to normal tissue controls (n = 287). (D) Overall survival analysis of PAAD patients using GEPIA 2.0, comparing patients with high (n = 89) versus low (n = 89) SLC2A1 expression. (E) Overall survival analysis of NSCLC patients using GEPIA 2.0, comparing patients with high (n = 481) versus low (n = 481) SLC2A1 expression. (F) Representative immunohistochemical staining of GLUT1 in pancreatic tumor tissues and adjacent normal tissues. Scale bar: 100 μm. (G) Inhibitory effects of combined BAY-876 and GEM on Panc02 cell proliferation at 48 and 72 h assessed by EdU assay. Blue: DAPI, Red: EdU-positive cells. Scale bar: 200 μm. (H) Intracellular ROS generation in Panc02 cells treated with combined BAY-876 and GEM using DCFH-DA probe. Blue: DAPI, Green: DCFH-DA. Scale bar: 50 μm. (I) Representative immunofluorescence images of γ-H2AX foci in Panc02 cells after indicated treatments for 72 h. Blue: DAPI, Green: γ-H2AX. Scale bar: 20 μm. Data are presented as mean ± SD. ∗∗p< 0.01, ∗∗∗p< 0.001 and ∗∗∗∗p < 0.0001.

We conducted in vitro studies using the GLUT1-specific inhibitor BAY-876 to verify this hypothesis. Glucose uptake assays demonstrated that BAY-876 inhibited glucose uptake in Panc02 cells in a concentration-dependent manner. Treatment with 40 nM BAY-876 reduced glucose uptake by more than 60%, confirming effective blockade of the glucose supply to tumor cells (Fig. S1C). We then evaluated the combined effect of BAY-876 and GEM, determining that a molar ratio of 2:1 provided optimal inhibitory activity against cell viability (Fig. S1D). This combination significantly sensitized Panc02 cells to GEM, lowering its IC50 by approximately 50% compared to GEM treatment alone (Fig. S1E and F). An EdU assay was performed to assess tumor cell proliferation across different treatments. The combination group (BAY-876 + GEM) displayed the lowest red fluorescence intensity, indicating the most potent inhibition of Panc02 cell proliferation (Fig. 1G). Intracellular reactive oxygen species (ROS) production was measured using the DCFH-DA fluorescent probe. BAY-876 induced ROS generation by blocking glycolysis, with the highest fluorescence intensity observed in the combination group, suggesting severe oxidative stress induced by the combined treatment (Fig. 1H). Given that excessive ROS can induce DNA damage, we further assessed DNA damage levels using γ-H2AX immunofluorescence staining (Fig. 1I). The BAY-876 + GEM group exhibited the strongest green fluorescence signal, indicating severe DNA damage. In conclusion, these in vitro results demonstrate that BAY-876 enhances the cytotoxic effects of GEM by restricting glucose availability and amplifying oxidative stress.

3.2. Fabrication and characterization of the biomimetic nanomodulator CM-cRGD@PBG

mPDA nanoparticles were synthesized using a modified template method. The biomimetic nanomodulator, CM-cRGD@PBG, was fabricated by encapsulating BAY-876 and GEM into the mPDA pores, followed by coating with cRGD-modified cancer cell membranes (CM-cRGD) to enable dual-targeting co-delivery of drugs to tumor sites (Scheme 1A). Morphological characterization via SEM and TEM revealed that the synthesized mPDA possessed a uniform spherical morphology with a distinct mesoporous structure (Fig. 2A and B), providing an ideal carrier for therapeutic agents. BAY-876 and GEM were efficiently encapsulated into the mPDA pores via π-π stacking and hydrogen bonding interactions. Following drug encapsulation, the mesoporous channels of PBG became less distinct, and the average hydrodynamic diameter increased from 164.1 nm to 189.8 nm (Fig. 2C). This was accompanied by a shift in zeta potential from −25.5 mV to −31.2 mV (Fig. 2D), indicating successful drug loading. The encapsulation efficiencies for BAY-876 and GEM were determined to be 99.9% and 70.4%, respectively. TEM imaging of CM-cRGD@PBG revealed a classic “core-shell” nanostructure. After biomimetic membrane modification, the hydrodynamic diameter of CM-cRGD@PBG increased by approximately 20 nm to 212.3 nm, while the zeta potential shifted to −15.3 mV. HAADF-STEM mapping confirmed the elemental composition of CM-cRGD@PBG and distribution of C, N, O, F, and P in CM-cRGD@PBG (Fig. 2E). In addition, the membranes and nanoparticles were labeled with DiD (red) and DiO (green), respectively. The distinct colocalization of the cRGD-modified membranes and the nanoparticles was observed by CLSM (Fig. 2F), confirming the successful membrane modification of mPDA. Moreover, SDS-PAGE analysis confirmed that CM-cRGD@PBG retained the characteristic membrane protein profile (Fig. 2G).

Fig. 2.

Characterization and photothermal performance of CM-cRGD@PBG. (A) SEM images of mPDA, PBG, and CM-cRGD@PBG. Scale bar: 500 nm. (B) TEM images and corresponding size distribution of mPDA, PBG, and CM-cRGD@PBG. Scale bar: 100 nm. (C, D) Hydrodynamic diameters (C) and zeta potentials (D) of mPDA, PBG, and CM-cRGD@PBG. (E) HAADF-STEM image and the corresponding EDS elemental mapping of C, N, O, F, and P for CM-cRGD@PBG. (F) CLSM images showing the colocalization of DiD-labeled CM-cRGD (red) and DiO-labeled mPDA (green). Scale bar: 10 μm. (G) SDS-PAGE protein profiles of mPDA, PBG, CM-cRGD@PBG, cancer cell membranes (CM), and cancer cell lysates. (H) Temperature elevation curves of 100 μg/mL CM-cRGD@PBG under varying power densities of 808 nm laser irradiation for 10 min. (I) Corresponding infrared thermal images at different time points. (J) Photothermal stability of CM-cRGD@PBG (100 μg/mL) under 808 nm laser irradiation at 1.5 W/cm² over four consecutive laser on/off cycles. Data are presented as mean ± SD.

Subsequently, we evaluated the physiological stability and biocompatibility of CM-cRGD@PBG. When dispersed in PBS at 4 °C for 48 h, the hydrodynamic diameters of bare mPDA and PBG nanoparticles significantly increased due to aggregation (Fig. S2A and B). In contrast, CM-cRGD@PBG exhibited enhanced stability in various physiological solutions (Fig. S2C and D). Its hydrodynamic diameter and polydispersity index (PDI) did not fluctuate significantly over 8 days. A hemolysis assay was performed to determine the hemolysis rate at varying concentrations of CM-cRGD@PBG. Compared with the positive control group, CM-cRGD@PBG showed excellent hemocompatibility with negligible hemolysis (<5%) (Fig. S3). Furthermore, cytotoxicity assays revealed that incubation with 200 μg/mL mPDA did not significantly compromise cell viability in either normal or tumor cells (Fig. S4). Therefore, CM-cRGD@PBG serves as a stable and biocompatible nanoplatform for effective drug delivery.

3.3. CM-cRGD@PBG exhibits excellent photothermal performance and robust photothermal stability

Beyond serving as a drug-delivery nanocarrier, mPDA is widely used for PTT owing to its intrinsic NIR absorption [41]. We measured the photothermal performance of the biomimetic nanomodulator using an infrared thermal imaging camera and a thermocouple probe. As shown in Fig. S5A and B, PBG and CM-cRGD@PBG exhibited negligible difference in photothermal heating capacity compared to mPDA. Upon laser irradiation at the same concentration, the temperature of all samples increased by approximately 23 °C. Moreover, the photothermal effect of CM-cRGD@PBG was dependent on both power density and concentration (Fig. 2H, I and Fig. S5C). Furthermore, CM-cRGD@PBG and mPDA showed no significant deterioration in heating performance during four successive laser irradiation cycles, indicating their superior photothermal stability (Fig. 2J and Fig. S5D). Collectively, these results indicate that CM-cRGD@PBG serves as an efficient and stable PTT agent for tumor therapy.

3.4. Dual-targeting strategy facilitates efficient cellular uptake and tumor accumulation

As illustrated in Fig. 3A, we designed a dual-targeting strategy to achieve precise tumor accumulation by integrating cancer cell membrane-mediated homologous targeting with α_νβ₃ integrin-specific active targeting, thereby facilitating efficient delivery of therapeutic agents for PTT. We investigated the cellular internalization of CM-cRGD@PBG using CM-cRGD@mPDA-DiO, a fluorescent surrogate prepared by loading DiO into mPDA. CLSM revealed that CM@mPDA-DiO displayed more intense green fluorescence in Panc02 cells, whereas the signals from DiO and mPDA-DiO control groups were faint (Fig. 3B). Notably, the dual-targeted CM-cRGD@mPDA-DiO displayed the strongest intracellular fluorescence, suggesting that cRGD modification further enhanced receptor-mediated endocytosis. CLSM and flow cytometry subsequently confirmed that the targeted internalization of CM-cRGD@mPDA-DiO was time-dependent, with significantly higher internalization efficiency at 1, 2, and 4 h compared to bare mPDA (Fig. 3C and Fig. S6C). We then evaluated the uptake efficiency of CM-cRGD@mPDA-DiO in Panc02 cells and NIH3T3 fibroblasts (murine fibroblast cell line). CLSM and flow cytometry revealed substantially stronger fluorescence in tumor cells than in normal cells (Fig. S6A and B), which can be attributed to its specific tumor-targeting capability. To demonstrate the broad applicability of the drug-delivery nanoplatform, we additionally investigated its cellular uptake in Lewis lung carcinoma (LLC) cells. The LLC-derived cell membrane and mPDA NPs were labeled with DiI (red) and DiO (green), respectively. After 8 h of incubation with LLC cells, significant colocalization of fluorescence signals was observed within the cells, indicating efficient internalization mediated by homologous and active targeting (Fig. S7).

Fig. 3.

Tumor-targeting capability and synergistic cytotoxicity of CM-cRGD@PBG. (A) Schematic illustration of the dual-targeting cellular uptake mechanism. (B) Representative fluorescence images of cellular uptake in Panc02 cells incubated with different formulations for 2 h. Blue: DAPI, Green: DiO. Scale bar: 200 μm. (C) Time-dependent uptake of mPDA or CM-cRGD@mPDA in Panc02 cells at 1, 2, and 4 h. Blue: DAPI, Green: DiO-labeled nanoparticles. Scale bar: 25 μm. (D, E) Ex vivo biodistribution images (D) and semi-quantitative analysis (E) of major organs (heart, liver, spleen, lung, and kidney) and tumors from Panc02 tumor-bearing mice at 24 h post-injection (n = 4). (F) Quantitative analysis of the apoptosis rates of Panc02 cells following various treatments (n = 3). (G) Representative flow cytometry plots of Panc02 cell apoptosis following indicated treatments. (H) Live/Dead cell fluorescence staining images of Panc02 cells subjected to different treatments. Green: Calcein-AM (live cells), Red: PI (dead cells). Scale bar: 200 μm. Data are presented as mean ± SD. ∗∗p< 0.01, ∗∗∗p < 0.001; n. s. Indicates not significant.

Furthermore, we evaluated the tumor-targeting capability of CM-cRGD@PBG by fluorescence imaging. DiR-labeled nanoparticles were prepared by encapsulating the NIR fluorescent probe DiR into the mPDA. At 24 h post-intravenous injection, the major organs and tumors were harvested for ex vivo imaging. Tumors of the CM-cRGD@mPDA-DiR group showed greater fluorescence intensity than those in the mPDA-DiR and CM@mPDA-DiR groups, thereby resulting in reduced accumulation in major organs (Fig. 3D and E). This enhanced targeted delivery and tumor accumulation can be attributed to its dual-targeting modification.

3.5. CM-cRGD@PBG potentiates metabolic-chemotherapeutic cytotoxicity via synergistic photothermal therapy in vitro

We assessed the level of cell apoptosis induced by CM-cRGD@PBG with or without laser irradiation. In Panc02 cells, the CM-cRGD@PBG group induced an apoptosis rate of 21.9%, which was higher than that of the BAY-876 + GEM group (18.5%). Notably, the CM-cRGD@PBG + L group showed a profoundly elevated apoptotic rate of 63.8%, significantly higher than that of mPDA + L group (41.7%), confirming that the synergistic combination of targeted drug delivery and PTT effectively promoted tumor cell apoptosis (Fig. 3F and G). Similarly, LLC cells treated with CM-cRGD@PBG + L also exhibited a significant apoptotic rate (∼47.4%) (Fig. S8). Furthermore, cell viability assays (CCK-8) showed that CM-cRGD@PBG + L reduced Panc02 viability to < 5% after 48 h of incubation (Fig. S9). Live/Dead assays confirmed that the CM-cRGD@PBG + L treatment could rapidly trigger extensive cell death, compared to the control treatments (Fig. 3H). Therefore, CM-cRGD@PBG-mediated PTT can significantly potentiate the direct antitumor efficacy of the BAY-876 and GEM combination in vitro.

3.6. CM-cRGD@PBG inhibits tumor glycolysis to alleviate lactate-driven M2-like polarization of BMDM in vitro

BAY-876 can disrupt tumor glycolysis by specifically blocking GLUT1 while sparing immune cells which primarily depend on GLUT3 [42,43]. The underlying mechanism mediated by BAY-876 is illustrated in Fig. 4A. We evaluated the extent of glycolysis inhibition following the experimental scheme depicted in Fig. 4B. Glucose uptake was assessed by measuring the glucose concentration in the culture medium of Panc02 cells following various treatments (Fig. 4C). In the BAY-876, BAY-876 + GEM, PBG, and CM-cRGD@PBG groups, the supernatant contained substantially higher glucose level than that of the controls. Notably, the CM-cRGD@PBG group exhibited a 2.05-fold higher glucose level than the control group, demonstrating its potent inhibitory effect on glucose uptake. Since glucose serves as an upstream substrate for glycolysis, blocking glucose uptake reduces lactate production, an end product of glycolysis. As shown in Fig. 4D, lactate secretion in the CM-cRGD@PBG group decreased by ∼88.0% compared to that of the control group. Since tumor cells predominantly depend on glycolysis for energy, CM-cRGD@PBG treatment reduced intracellular ATP levels to 37.1% of the control level (Fig. 4E). These results suggest that CM-cRGD@PBG can effectively inhibit tumor glycolysis by reducing glucose uptake, thereby decreasing lactate production and depriving tumor cells of their energy supply.

Fig. 4.

CM-cRGD@PBG inhibited tumor glycolysis to alleviate lactate-driven M2-like polarization of BMDMs in vitro. (A) Schematic mechanism of BAY-876-mediated GLUT1 blockade. (B) Schematic workflow of the metabolic analysis experiments. (C, D) Glucose concentrations (C) and lactate levels (D) in the cell culture supernatants of Panc02 cells treated with indicated formulations (n = 3). (E) Intracellular ATP levels in Panc02 cells following treatment with indicated formulations (n = 3). (F) Schematic illustration of the BMDM polarization assay using tumor-conditioned medium (TCM) collected from different treatment groups. (G) Representative CLSM images of CD206 (M2 marker) expression on BMDMs following a 24 h incubation with TCM from indicated groups. Scale bar: 25 μm. (H) Flow cytometric analysis of CD206 expression on BMDMs treated with TCM for 24 h. Data are presented as mean ± SD. ∗p< 0.05, ∗∗p< 0.01, and ∗∗∗p < 0.001.

Lactate is a key driver of the ITME, particularly by promoting TAM polarization toward the pro-tumoral M2 phenotype [44,45]. To investigate the modulation of the M2-like macrophage phenotype by CM-cRGD@PBG, bone marrow-derived macrophages (BMDMs) were incubated with tumor-conditioned medium (TCM) collected from Panc02 cells pretreated with different formulations (Fig. 4F). For BMDMs incubated with TCM from the control group (characterized by high lactate levels), immunofluorescence and flow cytometry analyses revealed high CD206 expression (43.8%), a typical marker of M2-like BMDMs. Conversely, CM-cRGD@PBG treatment significantly decreased the expression of CD206 on BMDMs to 18.4% (Fig. 4G and H). To further investigate this phenotypic shift, qRT-PCR analysis demonstrated that the mRNA level of the M1 marker Il6 in the CM-cRGD@PBG group showed a 56.6-fold increase, whereas the expression of the M2 marker Arg1 was significantly lower than that in the control (Fig. S10). Collectively, these results demonstrate that CM-cRGD@PBG successfully reprograms the lactate-driven TME, thereby repolarizing macrophages and remodeling the ITME.

3.7. CM-cRGD@PBG synergized with laser irradiation induces ICD to promote DC maturation and sensitizes tumor cells to CTL-mediated cytotoxicity in vitro

Next, we further investigated the immunomodulatory potential of CM-cRGD@PBG on DC maturation and CTL effector functions in vitro (Fig. 5A). PTT has been reported to induce ICD and release DAMPs, including CRT and HMGB1, to promote DC maturation, thereby eliciting a robust antitumor immune response. To demonstrate the superior capability of CM-cRGD@PBG + L in inducing ICD, the expression of ICD markers CRT and HMGB1 was assessed by immunofluorescence staining. As revealed in Fig. 5B, the CM-cRGD@PBG + L group displayed the most intense green fluorescence, while negligible CRT signal was detected in the control and mPDA groups. In addition, significant reduction in nuclear HMGB1 fluorescence was found in the mPDA + L and CM-cRGD@PBG + L groups compared to other controls (Fig. 5C). Collectively, these data verified that CM-cRGD@PBG + L synergistically potentiated ICD via driving CRT translocation to the cell surface and accelerating HMGB1 release. To examine whether the released DAMPs could effectively activate the immune response, bone marrow-derived dendritic cells (BMDCs) were co-cultured with Panc02 cells from different treatment groups for 36 h (Fig. 5D and Fig. S11). Flow cytometry analysis showed that the proportion of mature DCs in the CM-cRGD@PBG + L group reached 50.9%, which was 1.32 and 1.80 times that of the BAY-876 + GEM and the control group, respectively. These results indicate that the potent ICD induced by CM-cRGD@PBG + L can effectively promote DC maturation, bridging the innate and adaptive immunity.

Fig. 5.

In vitro ICD induction and immune activation elicited by the metabolic-chemo-photothermal synergistic strategy. (A) Schematic illustration of the experimental design for evaluating ICD induction and CTL-mediated cytotoxicity. (B, C) Representative CLSM images showing surface CRT exposure (B) and release of HMGB1 from the nucleus (C) in Panc02 cells following indicated treatments. Scale bar: 25 μm. (D) Flow cytometric analysis of DC maturation (CD80⁺CD86⁺) following a 36 h co-culture with tumor cells pretreated with indicated treatments. (E) Representative fluorescence images of intracellular ROS levels in Panc02 cells using the DCFH-DA probe. Blue: DAPI, Green: DCFH-DA. Scale bar: 50 μm. (F) Flow cytometry analysis of intracellular ROS levels in Panc02 cells. (G, H) Quantitative analysis of the apoptosis rates of Panc02 cells following indicated treatments in the absence (G) or presence (H) of CTLs (n = 3). Data are presented as mean ± SD. ∗∗p< 0.01 and ∗∗∗p < 0.001.

To compensate for the ATP shortage caused by BAY-876, tumor cells enhance mitochondrial oxidative phosphorylation (OXPHOS), thus elevating ROS levels. As shown in the fluorescent images (Fig. 5E), intracellular ROS levels were markedly elevated in the BAY-876, PBG, and CM-cRGD@PBG groups compared to the control and mPDA groups. Panc02 cells treated with CM-cRGD@PBG + L displayed the highest fluorescence, validating that combined metabolic and photothermal stress synergistically boosts ROS accumulation. Flow cytometry analysis also revealed a significantly increased proportion of ROS-positive cells in the CM-cRGD@PBG group (30.8%) compared to the control group (12.0%). When combined with photothermal stimulation, this proportion increased to 45.0% (Fig. 5F). Existing evidence suggests that the ROS induced by glycolysis inhibition enhances tumor cell susceptibility to TNF-α-dependent CTL cytotoxicity [46]. To validate this mechanism directly, we examined the combined antitumor effect of BAY-876 and TNF-α. As expected, the addition of TNF-α to BAY-876-treated Panc02 or LLC cells significantly promoted tumor cell death compared to TNF-α alone (Fig. S12). To determine whether BAY-876-induced metabolic stress enhances the antitumor effect of T cells, we employed a co-culture system. Flow cytometry analysis showed that the CTL-mediated tumor cell apoptosis rate in the BAY-876 group rose to 18.6% from 9.1% in the control group (Fig. 5G, H and Fig. S13). Additionally, the CTL-mediated apoptosis rate reached 18.8% in the CM-cRGD@PBG group and peaked at 21.3% in the CM-cRGD@PBG + L group. These results suggest that BAY-876 can significantly boost CTL-mediated killing while causing negligible toxicity to T cells (Fig. S14).

In summary, these findings demonstrate that CM-cRGD@PBG + L not only induces ICD but also metabolically sensitizes tumor cells to CTL-mediated cytotoxicity. By remodeling the ITME and enhancing both tumor immunogenicity and susceptibility to CTL cytotoxicity, the “metabolic-chemo-photothermal” strategy can facilitate multiple phases of the cancer-immunity cycle, with significant potential to elicit robust antitumor immune responses.

3.8. CM-cRGD@PBG potently inhibits Panc02 tumor growth via synergistic metabolic-chemo-photothermal therapy in vivo

To determine the optimal therapeutic window for PTT, an IVIS imaging system was utilized to track the biodistribution of DiR-labeled CM-cRGD@mPDA in Panc02-bearing mice at predetermined time points. The optimal time for laser irradiation was determined to be 8 h post-injection, as evidenced by the peak tumor fluorescence signal of the CM-cRGD@mPDA group at this time point (Fig. S15A). Quantitative analysis of tumor fluorescence indicated that CM-cRGD@mPDA-DiR was 3.65- and 2.01-fold that of free DiR and mPDA-DiR, respectively (Fig. S15B and C), which was attributed to homologous and active tumor targeting. Based on the aforementioned results, we further evaluated the photothermal conversion efficiency of the nanomodulator in vivo. Mice were intravenously administered with PBS, mPDA, PBG, or CM-cRGD@PBG. At 8 h post-injection, the real-time temperature at the tumor sites was recorded during laser irradiation. While the PBS group exhibited a temperature increase insufficient to reach the therapeutic threshold, the tumor temperature in the CM-cRGD@PBG group rose rapidly, increasing by approximately 14.6 °C to reach a peak temperature of about 50 °C, indicating sufficient potential to induce thermal damage to tumor tissues (Fig. S16 A and B).

To investigate the synergistic efficacy of the “metabolic-chemo-photothermal” strategy in a non-immunogenic “cold” tumor model, we established a subcutaneous Panc02 pancreatic cancer model in C57BL/6 J mice. Based on the identified PTT window, the in vivo treatment regimen was designed as illustrated in Fig. 6A. Panc02 tumor-bearing mice were randomized into groups to receive the indicated treatments, with laser exposure applied 8 h post-injection for the laser groups. Given the superior tumor-targeting capability and photothermal conversion efficacy observed in CM-cRGD@PBG, we evaluated its antitumor efficacy. Tumor growth curves in the mPDA and L groups were comparable to those of the control group, suggesting that mPDA or laser irradiation monotherapy resulted in minimal therapeutic efficacy with no significant difference from the control (Fig. 6B–D and Fig. S17). In contrast, the mPDA + L group exhibited moderate tumor inhibition, confirming the efficacy of mPDA-mediated PTT for tumor suppression. Benefiting from its superior tumor-targeting capability, the CM-cRGD@PBG group displayed significantly better tumor inhibition than the BAY-876 + GEM and PBG groups. Furthermore, by integrating metabolic-chemotherapy and PTT, the CM-cRGD@PBG + L group achieved the most pronounced tumor growth inhibition, with a suppression rate of ∼95.8%. Photographs and the average weight of the dissected tumors further confirmed the excellent antitumor therapeutic efficacy of the CM-cRGD@PBG + L treatment.

Fig. 6.

The synergistic antitumor efficacy of CM-cRGD@PBG under laser irradiation in the subcutaneous Panc02 tumor model. (A) Schematic illustration of the experimental timeline for the establishment of the subcutaneous Panc02 tumor model and the therapeutic regimen. (B) Photographs of dissected tumors from each group at the study endpoint. (C) Tumor growth curves of mice receiving different treatments (n = 6). (D) Tumor weights after different treatments (n = 6). (E) Body weight changes of mice receiving different treatments throughout the treatment period (n = 6). (F) Representative H&E staining images of tumor sections from different groups. Scale bar: 200 μm. (G) Ki67 immunohistochemical staining of tumor sections. Scale bar: 100 μm. (H) TUNEL immunofluorescence staining of tumor sections. Scale bar: 50 μm. Data are presented as mean ± SD. ∗p< 0.05, ∗∗∗p< 0.001; n. s. Indicates not significant. Ⅰ: PBS; Ⅱ: mPDA; Ⅲ: BAY-876 + GEM; Ⅳ: PBG; Ⅴ: CM-cRGD@PBG; Ⅵ: L; Ⅶ: mPDA + L; Ⅷ: CM-cRGD@PBG + L.

To evaluate antitumor efficacy, tumor sections from each group were processed for H&E, Ki67, and TUNEL staining at the study endpoint. H&E staining revealed that the tumors in the CM-cRGD@PBG + L group exhibited the most extensive tissue necrosis, characterized by distinct chromatin condensation and cell shrinkage (Fig. 6F). In contrast to the CM-cRGD@PBG and mPDA + L groups, the CM-cRGD@PBG + L group displayed the lowest level of Ki67-positive cells, indicating the most significant suppression of tumor cell proliferation (Fig. 6G). Consistent with these findings, TUNEL staining exhibited the most intense green fluorescence in the CM-cRGD@PBG + L group, indicating a significant increase in tumor cell apoptosis (Fig. 6H). Throughout the treatment, no significant loss of body weight was recorded in any group (Fig. 6E). To assess the biocompatibility of the CM-cRGD@PBG + L treatment, routine blood and biochemistry tests were performed. All hematological parameters exhibited no significant differences compared to the control group, suggesting no significant hematotoxicity of the CM-cRGD@PBG + L treatment (Table S1). Key functional indicators for the liver and kidneys, including ALT, AST and BUN, remained within reference ranges (Fig. S18). Additionally, H&E staining of major organs showed no severe pathological changes in any group (Fig. S19). These results indicate that CM-cRGD@PBG + L can provide an effective therapeutic strategy with excellent biosafety for “cold” tumors.

3.9. CM-cRGD@PBG potentiates robust antitumor immunity by combining metabolic-chemo ITME remodeling with photothermal immunogenicity enhancement

Encouraged by the potent in vitro immunomodulatory effects and in vivo antitumor efficacy, we further investigated the antitumor immune response induced by CM-cRGD@PBG + L in vivo. To elucidate the synergistic effect mediated by the combination of BAY-876 and GEM and the ICD effect, we performed IHC staining and flow cytometry to comprehensively analyze the proportions and functions of immunosuppressive subsets and effector T cells following different treatments. As a predominant immunosuppressive population within the ITME, M2-like TAMs are key drivers of tumor growth and suppress antitumor immunity. As shown in Fig. 7A and B, representative IHC images revealed that the proportion of M1-like macrophages (CD86⁺) distinctly increased in the BAY-876-containing groups, whereas that of M2-like macrophages (CD206⁺) significantly decreased. Flow cytometric quantification further demonstrated that, compared to the control group (28.7%), the percentage of M2-like macrophages significantly decreased to 10.2% and 8.12% in the CM-cRGD@PBG and CM-cRGD@PBG + L groups, respectively (Fig. S20A and S21). The shift from M2-like to M1-like macrophages indicated that BAY-876 effectively reduced lactate-driven M2-like TAM polarization through metabolic modulation. Subsequently, we evaluated MDSC infiltration, which creates an immunosuppressive barrier hindering CTL infiltration and function. Compared to the control group (44.2%), CM-cRGD@PBG treatment effectively reduced the frequency of MDSCs to 34.1%. The CM-cRGD@PBG + L treatment further reduced the proportion of MDSCs to 29.7%, compared with the L (52.6%) and mPDA + L (44.9%) groups (Fig. 7C and D). This suggested that the targeted delivery of GEM successfully countered the inflammation-driven MDSC recruitment caused by photothermal stress. Additionally, IHC images showed the lowest Treg infiltration into tumor tissues in the CM-cRGD@PBG + L group (Fig. S22). These results indicate that CM-cRGD@PBG effectively reshapes the ITME by combining BAY-876-mediated metabolic regulation with GEM-mediated immunomodulation.

Fig. 7.

CM-cRGD@PBG under laser irradiation effectively remodeled the immunosuppressive tumor microenvironment and elicited robust antitumor immunity in Panc02 tumor-bearing mice. (A, B) Representative immunohistochemical images of M1-like TAMs (A, CD86⁺) and M2-like TAMs (B, CD206⁺) in tumor tissues. Scale bar: 100 μm. (C, D) Representative flow cytometry plots (C) and quantitative analysis (D) of tumor-infiltrating MDSCs (n = 5). (E) Quantitative analysis of DC maturation in tumor tissues following indicated treatments (n = 4). (F) Quantitative analysis of tumor-infiltrating CD8⁺ T cells (n = 5). (G, H) Representative immunohistochemical images of CD8⁺ T cell infiltration (G) and Granzyme B expression (H) in tumor tissues. Scale bar: 100 μm. (I) The ratio of CD8⁺/CD4⁺ T cells in tumor tissues following indicated treatments (n = 5). (J) Serum levels of IFN-γ in mice from each group were determined by ELISA (n = 3). (K, L) Quantitative analysis (K) and representative flow cytometry plots (L) of CD8⁺ T cells in spleens of the tumor-bearing mice after different treatments (n = 4). Data are presented as mean ± SD. ∗p< 0.05, ∗∗p< 0.01, and ∗∗∗p< 0.001. Ⅰ: PBS; Ⅱ: mPDA; Ⅲ: BAY-876 + GEM; Ⅳ: PBG; Ⅴ: CM-cRGD@PBG; Ⅵ: L; Ⅶ: mPDA + L; Ⅷ: CM-cRGD@PBG + L.

Building on the successful remodeling of the ITME, we further analyzed the proportion of mature DCs in tumors. Flow cytometry analysis showed that the CM-cRGD@PBG + L treatment had the highest proportion of mature DCs, which was 1.86 times that of the control group (Fig. 7E, Fig. S20B and S23). This indicates that CM-cRGD@PBG + L-mediated PTT can effectively trigger ICD to improve the maturation of DCs, thereby priming the recruitment and activation of CTLs. We then evaluated the infiltration levels of CD8⁺ and CD4⁺ T lymphocytes in tumors and spleens. As shown in Fig. 7F, the CM-cRGD@PBG + L group presented the most abundant tumor-infiltrating CD8⁺ T cells, reaching levels 2.19 and 3.49 times greater than those of the control and L groups, respectively. Additionally, representative IHC images of the tumor sections showed the strongest CD8-positive signals in the CM-cRGD@PBG + L group (Fig. 7G). Furthermore, compared to other groups, the CM-cRGD@PBG + L treatment showed significant intratumoral expression of Granzyme B (Fig. 7H), confirming that this strategy could successfully trigger robust CTL-mediated cytotoxicity. The intratumoral CD8⁺/CD4⁺ T cell ratio was also significantly higher in the CM-cRGD@PBG + L group, indicating effective activation of antitumor immunity (Fig. 7I). Serum IFN-γ levels were quantified across treatment groups. The IFN-γ levels in the CM-cRGD@PBG and CM-cRGD@PBG + L groups were 1.77-fold and 2.22-fold higher than those of the control group, respectively (Fig. 7J). Additionally, the populations of CD8⁺ T and CD4⁺ T cells in the spleens were substantially increased following the CM-cRGD@PBG + L treatment (Fig. 7K, L and Fig. S24, S25).

In summary, the “metabolic-chemo-photothermal” strategy can facilitate the targeted co-delivery of BAY-876 and GEM to reprogram the ITME in “cold” tumors. This approach creates a favorable environment to boost PTT-induced immunity, ultimately eliciting a strong antitumor immune response.

3.10. CM-cRGD@PBG-mediated photothermal therapy inhibits tumor growth in the LLC cold tumor model via ITME remodeling and robust immune activation

Encouraged by the antitumor efficacy in the pancreatic cancer model, we expanded our evaluation to the LLC lung cancer model. LLC is classified as an immune “cold” tumor with a severely immunosuppressive microenvironment. LLC tumor-bearing mice were treated following the regimen shown in Fig. 8A. Consistent with the findings in the Panc02 model, the combination of metabolic therapy and the chemo-immunomodulatory effect of chemotherapy suppressed tumor growth. The CM-cRGD@PBG + L group showed better antitumor effect, achieving the most significant tumor growth inhibition. Remarkably, the CM-cRGD@PBG + L treatment caused complete tumor regression in 40% of the mice, as indicated by the tumor growth curves and weights (Fig. 8B–D and Fig. S26). Tumor sections from each treatment group were subjected to histological analysis using H&E, Ki67, and TUNEL staining (Fig. 8F–H). H&E staining of the excised tumors revealed the most severe necrotic damage in the CM-cRGD@PBG + L group. Consistently, the tumors in this group showed the lowest Ki67 expression and the highest proportion of TUNEL-positive apoptotic cells. Overall, these results demonstrate the versatility and strong antitumor effects of the “metabolic-chemo-photothermal” strategy across different “cold” tumor models. Moreover, no substantial fluctuations in body weight were recorded among any of the treatment groups over the entire therapeutic duration, indicating the satisfactory biocompatibility and negligible systemic toxicity of this nanomodulator-mediated combined therapy (Fig. 8E).

Fig. 8.

CM-cRGD@PBG under laser irradiation exhibited synergistic antitumor efficacy in the subcutaneous LLC tumor model. (A) Schematic illustration of the experimental timeline for the establishment of the subcutaneous LLC tumor model and the therapeutic regimen. (B) Tumor growth curves of mice receiving different treatments (n = 5). (C) Photographs of dissected tumors from each group at the study endpoint. (D) Tumor weights after different treatments (n = 5). (E) Body weight changes of mice throughout the treatment period (n = 5). (F) Representative H&E staining of tumor sections from different groups. Scale bar: 200 μm. (G) Ki67 immunohistochemical staining of tumor sections. Scale bar: 100 μm. (H) TUNEL immunofluorescence staining of tumor sections. Scale bar: 50 μm. Data are presented as mean ± SD. ∗∗∗p< 0.001. Ⅰ: PBS; Ⅱ: mPDA; Ⅲ: BAY-876 + GEM; Ⅳ: PBG; Ⅴ: CM-cRGD@PBG; Ⅵ: L; Ⅶ: mPDA + L; Ⅷ: CM-cRGD@PBG + L.

We then evaluated the remodeling of the tumor immune microenvironment. IHC staining showed that the CM-cRGD@PBG (+L) treatment increased the proportion of M1-like antitumor macrophages while significantly reducing M2-like pro-tumor macrophages (Fig. 9A and B). These results were further confirmed by quantitative flow cytometry analysis. The M1/M2 ratio in the CM-cRGD@PBG group showed a 1.65-fold increase compared to the control, whereas the CM-cRGD@PBG + L group achieved a ratio 1.74 times higher than that of the L group (Fig. 9C and Fig. S20A). These results indicated that CM-cRGD@PBG could significantly promote the polarization of macrophages towards the M1 phenotype. We also assessed the infiltration of MDSCs in tumor tissues after different treatments. Similar to the findings in the Panc02 model, the CM-cRGD@PBG + L group had the lowest percentage of immunosuppressive MDSCs, effectively reversing the ITME (Fig. S20A and S27). The proportion of mature DCs in both tumor tissues and lymph nodes was significantly increased by CM-cRGD@PBG + L treatment relative to the control group, with respective rises from 10.4% to 18.9% and from 4.61% to 31.5%, respectively (Fig. 9D, E and Fig. S20B, S28). Additionally, IHC staining and quantitative flow cytometry analysis showed that the CM-cRGD@PBG + L group displayed the highest percentage of CD8⁺ T cells infiltrating the tumor, which increased to 25.1% from 7.24% in the control group (Fig. 9F,H and Fig. S20B, S29A). Meanwhile, these tumor-infiltrating CD8⁺ T cells exhibited strong Granzyme B expression, indicating effective CTL activation (Fig. 9G). The CM-cRGD@PBG + L group also exhibited an elevated accumulation of intratumoral CD4⁺ T cells, whereas the population of immunosuppressive Tregs declined substantially (Fig. 9I and Fig. S29B and C). Regarding the systemic immune response, the CM-cRGD@PBG + L treatment significantly increased the frequencies of CD8⁺ and CD4⁺ T cells in the spleens, representing 2.28-fold and 2.77-fold increases relative to the control group, respectively (Fig. 9J, K and Fig. S30).

Fig. 9.

CM-cRGD@PBG under laser irradiation induced the remodeling of the immunosuppressive tumor microenvironment and antitumor immunity in LLC tumor-bearing mice. (A, B) Representative immunohistochemical images of M1-like TAMs (A, CD86⁺) and M2-like TAMs (B, CD206⁺) in tumor tissues. Scale bar: 100 μm. (C) The ratio of M1 (CD86⁺)/M2 (CD206⁺) TAMs in tumor tissues following indicated treatments determined by flow cytometry (n = 3). (D, E) Quantitative analysis of DC maturation in tumor tissues (D) and lymph nodes (E) following indicated treatments (n = 3). (F, G) Representative immunohistochemical images of CD8⁺ T cell infiltration (F) and Granzyme B expression (G) in tumor tissues. Scale bar: 100 μm. (H, I) Quantitative flow cytometric analysis of tumor-infiltrating CD8⁺ (H) and CD4⁺ (I) T cells after different treatments (n = 3). (J, K) Quantitative analysis (J) and representative flow cytometry plots (K) of CD8⁺ T cells in spleens of the tumor-bearing mice after different treatments (n = 4). Data are presented as mean ± SD, ∗p< 0.05, ∗∗p< 0.01, and ∗∗∗p< 0.001. Ⅰ: PBS; Ⅱ: mPDA; Ⅲ: BAY-876 + GEM; Ⅳ: PBG; Ⅴ: CM-cRGD@PBG; Ⅵ: L; Ⅶ: mPDA + L; Ⅷ: CM-cRGD@PBG + L.

Taken together, these results demonstrate that the nanomodulator-mediated “metabolic-chemo-photothermal” combination strategy can effectively reprogram the ITME and stimulate robust antitumor immunity across multiple “cold” tumor models, highlighting its potential as a versatile nanomedicine platform for therapeutic applications.

4. Conclusion

We developed a biomimetic nanomodulator, CM-cRGD@PBG, to enhance antitumor immunity against “cold” tumors. This system employs a dual-targeting strategy that integrates homologous targeting via cancer cell membranes and active targeting through cRGD modification, enabling efficient delivery of BAY-876 and GEM to the tumor site. CM-cRGD@PBG not only directly inhibits rapid tumor proliferation but also significantly reprograms the tumor immune microenvironment. The released BAY-876 deprives tumor cells of glucose and subsequent energy supplies by blocking GLUT1. This metabolic intervention reduces interstitial glucose consumption and lactate accumulation, thereby reversing the ITME by suppressing M2-like TAM polarization and Treg expansion. GEM has been repurposed as an immunomodulator at sub-clinical chemotherapeutic doses to enhance antitumor immunity by selectively depleting immunosuppressive MDSCs and Tregs, rather than relying on its direct cytotoxicity. Given that PTT-mediated tumor ablation often induces a pro-inflammatory microenvironment and recruits MDSCs, GEM acts as a complement to PTT by alleviating potential immunosuppression. While the potential of BAY-876 to sensitize tumors to standard chemotherapy deserves further study, this work focuses on the synergy between BAY-876 and low-dose GEM in modulating the ITME. By dismantling both metabolic and cellular barriers, this combination strategy successfully converts “cold” tumors into immunologically “hot” tumors, resulting in extensive CTL infiltration and persistent effector function. This optimized immune microenvironment creates favorable conditions for the therapeutic efficacy of mPDA-mediated PTT. Beyond directly ablating the primary tumor, PTT induces ICD, thereby functioning as an in situ nanovaccine that primes systemic antitumor responses. The ICD process releases abundant TAAs and DAMPs, which promote DC maturation and further amplify CTL-mediated antitumor immunity. Taken together, the combination of CM-cRGD@PBG and PTT effectively remodels the tumor immune microenvironment and enhances tumor immunogenicity, converting “cold” tumors into “hot” ones through integrated immunomodulation and eliciting potent T cell-mediated immunity. As a result, this combination regimen exhibits strong antitumor efficacy and may induce durable immune responses. Furthermore, PTT facilitates nanomedicine accumulation and deep infiltration into the tumor core by inducing vasodilation and enhancing vascular permeability, and alleviates interstitial fluid pressure (IFP) through extracellular matrix disruption, further amplifying synergistic antitumor efficacy [[47], [48], [49]]. In summary, this study integrates metabolic reprogramming with immunosuppressive TME remodeling, establishing a systemic immunomodulation strategy that offers a promising avenue for reinvigorating antitumor immunity.

Declaration of generative AI and AI-assisted technologies in the writing process

Gemini was used to assist with language refinement in this manuscript. The authors confirm that the text has been comprehensively reviewed and edited by them, and they bear full responsibility for its publication.

CRediT authorship contribution statement

Xu Zhao: Conceptualization, Investigation, Methodology, Software, Writing – original draft, Writing – review & editing. Ying Yang: Investigation, Methodology, Software. Yanan Niu: Investigation, Methodology. Junya Feng: Investigation, Methodology. Wei Yuan: Funding acquisition, Supervision, Writing – original draft, Writing – review & editing. Mingyang Liu: Supervision, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.