Immediate tumor killing and long-term anti-tumor immunoreaction induced by Bufalin-loaded phototherapeutic Janus membrane in CRC postoperative therapy

Abstract

Colorectal cancer (CRC) management is significantly hindered by the challenges of early detection and the high rates of recurrence and metastasis following surgery. Systemic postoperative therapies are frequently compromised by adverse effects and drug resistance. To address this, we engineered a localized and combinatorial platform by integrating Bufalin (BU), a potent anti-tumor agent from traditional Chinese medicine, into citric acid-based carbon dots modified with hyaluronic acid (BU-CDs_CA-HA), which were subsequently incorporated into a Janus-structured membrane. This design enabled robust adhesion to post-resection tumor sites, ensuring sustained local drug release and spatially confined photothermal/photodynamic therapy (PTT/PDT) while effectively sparing adjacent healthy tissues from thermal damage. Upon near-infrared (NIR) irradiation, the synergistic BU-PTT/PDT action induced potent immunogenic cell death (ICD), as evidenced by the release of damage-associated molecular patterns (DAMPs), thereby initiating a systemic anti-tumor immune response. Concurrently, BU down-regulated heat shock proteins (HSPs) and HIF-1α expression, leading to the suppression of tumor PD-L1 and effectively countering immune escape. In orthotopic and metastatic CRC mouse models, this multifunctional Janus membrane system demonstrated remarkable efficacy in preventing local recurrence and distant metastasis. This outcome was attributed to the concerted effect of immediate cytotoxic ablation and the elicited durable anti-tumor immunity. Our work presents a novel biomaterial-based strategy that enhanced the efficacy and safety of postoperative CRC treatment, offering a versatile platform for site-specific combination therapy.

Graphical abstract

The therapeutic mechanism of CRC tumors: the BU-CDs_CA-HA-loaded Janus membrane was attached to tumor with NIR, bringing PTT and PDT. Meanwhile, BU-CDs_CA-HA was released from Janus membrane and entered tumor cells specifically. The BU-PTT/PDT combination brought an immediate tumor-killing effect through BU, heat and ROS, inducing ICD of tumor cells and thus activating anti-tumor immune response. Moreover, BU could decrease immunosuppression of TCM, supporting a long-term anti-tumor immunoreaction. The offense-defense integrated strategy could enhance the inhibition of CRC tumors and prevent the recurrence and metastasis.

1. Introduction

Colorectal cancer (CRC), one of the most incident malignant neoplasms globally, ranks as the third leading cause of cancer incidence and the second primary cause of cancer-related mortality worldwide [1,2]. For early-stage of CRC, surgical resection combined with the adjuvant therapies (e.g., radiotherapy and chemotherapy) serves as the optimal therapeutic modality. However, the insidious clinical manifestation of CRC often leads to delayed diagnosis, with most patients presenting at intermediate or advanced stages. At these stages, conventional therapeutic efficacy is substantially compromised by inherent challenges, including multi-drug resistance (MDR), radioresistance, and high risks of recurrence and metastasis in advanced tumors [3]. This clinical predicament highlights the urgent need for innovative postoperative tumor-suppressive strategies that integrate immediate therapeutic potency with long-term anti-tumor persistence.

Traditional Chinese medicine (TCM) has emerged as a promising candidate for anti-cancer drug development, attributed to its unique capacity to regulate anti-tumor immune responses through multi-channel, multi-target, and multi-level mechanisms, which coupled with low resistance induction and broad-spectrum activity compared to conventional chemical and biological agents [4,5]. Bufalin (BU), a principal bioactive component isolated from Venenum Bufonis, exerts persuasive anti-tumor effects by inducing multiple modes of tumor cell death, including apoptosis, necrosis, autophagy, and cellular senescence, thereby suppressing tumor proliferation. Furthermore, BU abrogates tumor cell migration, invasion, and angiogenesis by attenuating tumor stemness and modulating cellular phenotypes—key pathological processes driving metastatic progression [[6], [7], [8], [9], [10]]. Despite these compelling anti-tumor properties, the clinical translation of free BU is severely hindered by its inherent limitations: high systemic toxicity, poor aqueous solubility, and short in vivo half-life.

Phototherapy has garnered considerable attention in oncology due to its ability to generate localized anti-tumor effects (either hyperthermia or cytotoxic reactive oxygen species) upon stimulation with specific wavelengths of light, thereby minimizing off-target side effects associated with conventional chemotherapy and radiotherapy. Notably, both photothermal therapy (PTT) and photodynamic therapy (PDT) can induce ICD, a process that elicits adaptive anti-tumor immune responses, enabling long-term tumor surveillance [[11], [12], [13]]. However, monotherapeutic applications of PTT or PDT face inherent challenges: PTT-induced hyperthermia upregulates heat HSPs in tumor cells, which confer thermoresistance. And PDT relies on oxygen as a substrate, while anoxic tumor microenvironment (TME) would significantly attenuates its therapeutic efficacy [[14], [15], [16]]. The combination of PTT and PDT circumvents these limitations through synergistic interactions: (1) PTT-mediated local hyperthermia enhances intracellular photosensitizer delivery and accelerates tumor vascular perfusion, thereby alleviating TME hypoxia and potentiating PDT efficacy; (2) PDT-generated reactive oxygen species (ROS) disrupts HSP-mediated thermoprotective mechanisms, sensitizing tumor cells to PTT-induced hyperthermia. Among various nanoplatforms developed for combined phototherapy, carbon dots (CDs) have emerged as a versatile candidate due to their intrinsic photophysical properties, excellent aqueous solubility, superior biocompatibility, and facile surface functionalization for drug loading [[17], [18], [19]]. Specifically, CDs synthesized from citric acid (CA), a photosensitive small molecule, have been demonstrated to mediate simultaneous PTT and PDT. Under 660 nm light-emitting diode (LED) irradiation, CA-derived CDs exhibit a high photothermal conversion efficiency of 77.6 % and a singlet oxygen yield of 0.37, achieving robust tumor growth inhibition without inducing toxic side effects in adjacent healthy tissues. Given these facts, CA-derived photosensitive CDs hold significant potential as a multifunctional platform for integrating BU delivery with synergistic PTT/PDT. Such a combination strategy is anticipated to achieve: (1) immediate tumor ablation via PTT/PDT-mediated hyperthermia and ROS generation; (2) enhanced BU bioavailability and reduced systemic toxicity; (3) long-term anti-tumor immunity via ICD induction.

Herein, a BU-CDs_CA-HA-loaded Janus membrane based on interfacial polymerization with BU-PTT/PDT synergistic strategy was proposed for inhibition of postoperative tumor recurrence and metastasis. Briefly, the tumor-targeting carbon dots (CDs_CA-HA) with PTT/PDT were synthesized from CA and HA firstly. β-cyclodextrin was then grafted onto the CDs_CA-HA for loading BU (BU-CDs_CA-HA) to increase drug bioavailability, biocompatibility and tumor-targeting behavior. Finally, the muti-functional Janus membrane was prepared from BU-CDs_CA-HA, 4-arm polyethylene glycol (PEG-SS₄) and poly(lipoic acid) (PLA) by interfacial polymerization which had tissue adhesion, tumor-targeting release of BU and PTT/PDT simultaneously. Results showed that the Janus membrane not only converted UV light (808 nm) into local heat for PTT, but also generated abundant endogenous ROS for PDT to induce apoptosis and ICD of tumor cell which then released tumor-specific antigens and DAMPs to promote maturation of DCs, thus triggering an effective systemic anti-tumor immune response. Meanwhile, BU-CDs_CA-HA from Janus membrane could actively targeted to the tumor to delivery BU for down-regulation of the expression of HSP70, HSP90, HIF-1α and PD-L1, thus improving phototherapy efficiency and suppressing immune escape of tumor cells. This study might provide a new synergistic strategy for CRC postoperative treatment and a broad platform for novel application of TCM (Scheme 1).

Scheme 1.

(A) The preparation procedure of BU-CDs_CA-HA and Janus membrane; (B) The therapeutic mechanism of CRC tumors: the BU-CDs_CA-HA-loaded Janus membrane was attached to tumor with NIR, bringing PTT and PDT. Meanwhile, BU-CDs_CA-HA was released from Janus membrane and entered tumor cells specifically. The BU-PTT/PDT combination brought an immediate tumor-killing effect through BU, heat and ROS, inducing ICD of tumor cells and thus activating anti-tumor immune response. Moreover, BU could decrease immunosuppression of TCM, supporting a long-term anti-tumor immunoreaction. The offense-defense integrated strategy could enhance the inhibition of CRC tumors and prevent the recurrence and metastasis.

2. Experiment section

2.1. Materials

Citric acid (CA) and hyaluronic acid (HA) were purchased from Beyotime Biotechnology (Shanghai). 4-arm polyethylene glycol (PEG-OH₄) and lipoic acid (LA) were purchased from Aladdin (Shanghai). Bufalin (BU) was obtained from Herbpurify (Chengdu). The live/dead cell double staining kit was acquired from the Tongren Institute of Chemistry. For detecting adenosine triphosphate (ATP) and reactive oxygen species (ROS), assay kits were purchased from Beyotime Biotechnology (Shanghai). For detecting high mobility group box-1 protein (HMGB1), assay kits were purchased from CUSABIO (Wuhan). Antibodies against CRT, Ki67, CD4, CD8, HSP70 and fluorescently labeled secondary antibodies were purchased from Abcam (UK). Antibodies against β-actin and HIF-1α were purchased from Proteintech (USA). Antibodies against CD11c, CD80 and CD86 were obtained from Biolegend (USA) and Invitrogen (USA), respectively. The terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay kit was purchased from Roche (Switzerland). The IFN-γ and TNF-α detection kit was purchased from Biodragon Biotechnology (Beijing), the IL-6 detection kit was obtained from BOSTER Biotechnology (Wuhan).

2.2. Cell lines and animals

CT26, HCT116, RAW 264.7 and human colon epithelial cells (FHC) were obtained from the Institute of Biochemistry and Cell Biology of Chinese Academy of Sciences (Shanghai). The cells were cultured in RPMI 1640 medium, supplemented with 1 % penicillin/streptomycin (v/v) and 10 % fetal bovine serum (FBS). RAW 264.7 and FHC were cultured in DMEM enriched with 10 % fetal bovine serum (FBS) and 1 % penicillin/streptomycin. The culture condition maintained under a humidified atmosphere containing 5 % CO₂ at 37 °C. Balb/c mice (6–8 weeks old, weighing 20.0 ± 0.50 g) were purchased from Bikai Keyi Biotechnology (Shanghai). All animal experiments were conducted in accordance with the regulations of the Ethics Committee of Putuo Hospital, Shanghai University of Traditional Chinese Medicine China (approval number. DWEC-A-2024-16-2-86).

2.3. Synthesis and characterization of the BU-CDs_CA-HA

First of all, to synthesize CDs_CA-HA, 3 g CA, 0.3 g HA and 5 mL ethylenediamine were dissolved in 20 mL of deionized water (DI), followed by hydrothermal reaction in an oil bath at 180 °C for 2 h. The crude product was then dissolved in 1 M NaOH solution and stirred for 30 min. Subsequently, 1 M hydrochloric acid solution was used for neutralization, and CDs_CA-HA with homogeneous size and good water-solubility was obtained by using ultrafiltration and frozen-phase separation techniques.

Next, 50 g of β-cyclodextrin was dissolved in 0.45 M NaOH solution with constant stirring until the solution was clear and slightly yellow. 25 mL of acetonitrile solution containing 12.6 g of p-Toluenesulfonyl chloride (TsCl) was then added dropwise into the β-cyclodextrin solution. The whole system was protected by N₂ and reacted in an ice bath with continuous stirring. After completion of addition, the mixture was stirred for another 4 h at room temperature. Afterwards, 20 g of ammonium chloride (NH₄Cl) was added into the mixture to adjust pH to 8.5, which was recrystallized at 4 °C for 24 h. The crystal substance was washed 3 times by DI and acetone. Finally, the products (β-cyclodextrin-Ts) was dried under vacuum for further usage [20]. To obtain β-CDs_CA-HA, 8.8 g of β-cyclodextrin-Ts was dissolved with 20 g of CDs_CA-HA in 100 mL of DMSO. Next, NHS (7.75 mmol, 0.782 g) and EDC·HCl (15.4 mmol, 2.61 g) were added to carry out for 72 h at room temperature. Finally, the cyclodextrin-modified CDs_CA-HA (β- CDs_CA-HA) were dialyzed with DI for 72 h and lyophilized.

BU-CDs_CA-HA, a tumor-targeting phototherapeutic carbon dots containing Bufalin (BU), was prepared by dissolving BU and β-CDs_CA-HA in DMSO and dialysis with DI, followed by lyophilization. The successful synthesis was determined by nuclear magnetic resonance (¹H NMR), Fourier transform infrared (FT-IR). And the particle size and distribution of carbon dots were determined by dynamic light scattering (DLS). The morphology of carbon dots was observed by transmission electron microscopy (TEM). Further, the surface structure of carbon dots was observed by XPS. The drug loading (DL) and encapsulation rate (EI) of BU were determined by UV spectrophotometer.

2.4. Testing of photothermal properties of BU-CDs_CA-HA

To evaluate the photothermal conversion performance of BU-CDs_CA-HA, different concentrations of BU-CDs_CA-HA solution (0, 0.5, 1, and 2 mg/mL) were exposed to 808 nm NIR at a power density of 1.0 W/cm² (MDL-F-808, Changchun New Industries Optoelectronics Technology Co, Ltd). The temperature change was monitored and recorded using a NIR thermal camera. At the same time, to investigate the photothermal stability of BU-CDs_CA-HA, 2 mg/mL of BU-CDs_CA-HA solution was exposed to 808 nm NIR laser for 5 min and then naturally cool to room temperature circularly. The temperature was recorded every minute for 10 cycles.

2.5. Synthesis and characterization of the Janus membrane

2.5.1. Synthesis of PEG-SS₄ and PLA

First, 10 g PEG-OH₄ and 2 g lipoic acid were dissolved in 50 mL of dichloromethane. Then, 1.85 g EDC·HCl and 0.5 g DMAP were added to activate the carboxyl group of the lipoic acid and catalyze the esterification reaction. The mixture was stirred constantly for 48 h and then concentrated to 20 mL. Excess ice ether was used to obtain precipitated PEG-SS₄ which was dried under vacuum. The structure of PEG-SS₄ was characterized by ¹H NMR.

Tris base solution was prepared by 0.8 g of Tris base 4 mL DI. Next, 2 g lipoic acid (LA) monomer was dissolved in the Tris base solution to produce a transparent yellow solution. The mixture was reacted at 70 °C for 30 min. Afterwards, 200 μL of 1-vinylimidazole was added into the mixture. After 6 h, the system was lyophilized to produce poly(lipoic acid) (PLA). The successful synthesis was characterized by XPS and the stickiness of PLA was assessed through lap shear test.

2.5.2. Preparation and characterization of Janus membranes

Firstly, solution of organic (ethyl acetate) and aqueous (DI) phases was prepared separately. Organic phase: 0.09 g of PEG-SS₄ was dissolve in 300 μL ethyl acetate for further use. Aqueous phase: 0.09 g of PEG-SS₄, 600 μg of BU-CDs_CA-HA and 0.12 g of PLA was dissolved in 300 μL DI for further use. Next, the organic and aqueous phases were sequentially poured into a polytetrafluoroethylene mould (150 mm × 150 mm × 2 mm). Due to the immiscibility of ethyl acetate and DI, the two phases would form a bilayer system within the mould. When the system was stabilized, it was exposed to 365 nm UV light for 5 min to initiate photopolymerization. After reaction, the Janus membrane was removed from the mould and dried in fume hood for 24 h. The gelation kinetics was characterized by rotational rheometer (DHR-3, Waters, America; oscillation displacement: 0.03 rad; angular frequency: 10 rad/s; during time: 500 s). And the structure of Janus membrane was observed by scanning electron microscope (SEM, SU-1500, HITACHI, Japan). Further, the sustained release behavior of BU was measured [21]. Briefly, the samples were immersed in PBS at 37 °C and the fresh PBS was changed every day for 14 days. The concentration curves of released BU were recorded by UV–visible spectrophotometer (Agilent HP 8453, USA, absorption wavelength: 298 nm).

2.5.3. Biodegradation of Janus membranes

The degradation of Janus membrane was a key factor of biosafety through subcutaneous implantation of mice models. BALB/c mice (aged 6–8 weeks, n = 5) were firstly anaesthetized with 2.5 % sodium pentobarbital and the skin incision was made on the flank to implant Janus membrane, which was followed by suture. The degradation of Janus membrane was observed and skins were collected the weekly until complete degradation. Finally, the skins were cut into slices and stained with g haematoxylin and eosin (H&E) to assess the biocompatibility of Janus membrane.

2.6. In vitro anti-tumor efficiency

2.6.1. Hemolysis test

To test the hemolysis of carbon dots (CDs), the blood of mice was collected from eyes into heparin anticoagulant tubes which were then centrifuged (2000 rpm, room temperature) for 15 min. Next, β-CDs_CA-HA and BU-CDs_CA-HA were added in PBS with different concentration, respectively. Then, the red blood cells were aspirated by pipette tips to 1.5 mL EP tube and diluted with PBS to 4 %, which was co-cultured with β-CDs_CA-HA and BU-CDs_CA-HA solution at 37 °C for 4 h (the final mixture 3 % v/v), respectively. After co-culturing, the mixture was centrifuged (3000 rpm, room temperature) for 15 min. The hemolysis was measured through the absorbance at 540 nm by a microplate reader. It was worth noted that the β-CDs_CA-HA and BU-CDs_CA-HA solution was set as background due to their own color to calculate the relative value of absorbance.

2.6.2. Cytotoxicity test

To test the biocompatibility of β-CDs_CA-HA and Janus membrane, the suspension of CT26, HCT116, RAW 264.7 and FHC cells were obtained with 1 × 10⁵ cells/mL. And 100 μL the suspension was added to each well of 96-well plates and incubated at 37 °C in a 5 % CO₂ cell culture incubator for 24 h. According to the experimental design, the medium with different concentration of β-CDs_CA-HA and Janus membrane was added into 96-well plates for further culture. After another 24 h, 100 μL CCK-8 working solution was added to each well and incubated at 37 °C to measure the OD value at 450 nm using a microplate reader to calculate the cell survival rate.

For evaluating the anti-tumor efficiency of BU-CDs_CA-HA, the suspension of CT26 cells were prepared as the same as above and divided into 6 groups (1. Control, 2. 48 °C, 3. NIR, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR). According to the experimental design, the different treatments was added in 96-well plates for further culture. After another 24 h, 100 μL CCK-8 working solution was added to each well and incubated at 37 °C to measure the OD value at 450 nm using a microplate reader to calculate the cell survival rate. For groups requiring light treatment, irradiation with 808 nm NIR for 5 min per well was carried out. Besides, CT26 and HCT116 cells were prepared in 6 groups (1. Control, 2. 48 °C, 3. BU-CDs_CA-HA, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR) and stained with Calcein-AM/PI and observed with CLSM (Carl Zeiss Germany).

2.6.3. Cellular uptake

CT26 and HCT116 cells were cultured respectively in confocal dishes with approximately 3 × 10⁵ cells per well. BU-CDs_CA-HA (2 mg/mL) was added to replace the original medium. The incubation time points were set at 0, 2, 4, and 6 h. After incubation, cells from all groups were washed with PBS for 5 min, fixed with paraformaldehyde and stained with DAPI. After 5 min incubation at room temperature, fluorescence intensity was observed using a CLSM (Carl Zeiss Germany).

2.6.4. Intracellular ROS generation

CT26 cells were cultured in confocal dishes at a density of 2 × 10⁵ cells per dish. After 24h, the cells were treated with different conditions for 6 h (1. Control, 2. 48 °C, 3. BU-CDs_CA-HA, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR). The groups requiring NIR were irradiated with 808 nm (1 W/cm²) and further incubated for 24 h. Then, 10 μM DCFH-DA solution was added to the cells and incubated at 37 °C for another 20 min. Finally, the cells were washed several times with PBS and the observed by CLSM.

2.6.5. Western blot assay of HSPs, HIF-1α and PD-L1 expression level

To explain the temperature influence on HSPs, HIF-1α and PD-L1, CT26 cells were seeded in 6-well plates (3 × 10⁵ cells/well) and treated with different temperature for 24 h (37, 40, 45 and 50 °C). Afterwards, the cells were washed three times with PBS. Besides, total cellular protein was prepared in lysis buffer and quantified using BCA protein assay kit. A total of 20 μg of protein samples were separated by 8 % SDS-PAGE and then transferred to polyvinylidene difluoride (PVDF) membranes. Subsequently, the PVDF membranes were incubated overnight at 4 °C with diluted primary antibodies such as HSP70 (1:1000), HSP90 (1:1000), PD-L1 (1:1000), HIF-1α (1:2000),followed by horseradish peroxidase-conjugated secondary antibody (1:2000) at 37 °C for 1 h. Mouse β-actin (1:20000) was used as a protein loading control. The specific protein bands were imaged using enhanced chemiluminescence (ECL).

To prove the BU-PTT/PDT synergistic effect on HSPs, HIF-1α and PD-L1, CT26 cells were inoculated in 6-well plates (3 × 10⁵ cells/well) and treated with different conditions (1. Control, 2. 48 °C, 3. BU-CDs_CA-HA, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR). The groups requiring NIR were irradiated with 808 nm (1 W/cm²) and further incubated for 24 h. The following procedures and antibody indicators were same as above.

2.7. In vitro immunity

2.7.1. ICD of CT26 cells

To assess the ICD of CT26 cells, CRT on the surface of CT26 cells was detected firstly. Briefly, 2 × 10⁵ CT26 cells were cultured in 500 μL of complete medium in 24 well plates. After 24 h, the cells were treated with different groups (1. Control, 2. 48 °C, 3. BU-CDs_CA-HA, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR). The groups requiring NIR were irradiated with 808 nm (1 W/cm²) and further incubated for 24 h. After that, cells were immobilized with 4 % paraformaldehyde for 5 min, permeabilized with 0.1 % Triton X-100 for 5 min and then rinsed three times with PBS, followed by being blocked with 5 % fetal bovine serum for 1 h and stained with Alexa Fluor 488-CRT antibody (ab196158) at 4 °C overnight. Before being observed by CLSM, the cells were rinsed with PBS and stained with DAPI for 5 min.

Further, the extracellular HMGB1 concentration and intracellular ATP level were also detected. And CT26 cells were cultured for 24 h with the same treatments as described in the CRT experiment. The supernatant was collected to measure mouse HMGB1 using an ELISA kit, and cells were harvested for ATP analysis following the instructions of manufacturer.

2.7.2. In vitro DC maturation

In this study, Balb/c mice (3–5 weeks old, male, n = 5) were humanely euthanized using the cervical dislocation method. Then, the hind legs were skinned and the bilateral femurs and tibiae were carefully extracted. The bone marrow was then harvested by inserting a needle until the cavity turns white. Subsequently, the cells were then resuspended in a medium containing 10 % FBS, 10 ng/mL GM-CSF and 1 ng/mL IL-4 in RPMI-1640. The cell concentration was adjusted to 2 × 10⁶ and cultured in transwell chambers for 72 h. Afterwards, CT26 cells that had been treated with different conditions (1. Control, 2. 48 °C, 3. BU-CDs_CA-HA, 4. CA + NIR, 5. β-CDs_CA-HA + NIR, 6. BU-CDs_CA-HA + NIR) were added to the upper chamber and co-cultured for an additional 24 h. The DCs from different groups were collected, blocked with 2 % Fc block for 5 min, and incubated with antibodies CD11c (Biolegend), CD80 (Biolegend) and CD86 (Invitrofenfor) 30 min and fianlly detected by flow cytometry.

2.8. In vivo anti-tumor efficiency of BU-CDs_CA-HA

All animal procedures were carried out in line with National Research Council's Guide for the Care and Use of Laboratory Animals as well as Institutional Animal Care and Use Committee of Putuo Hospital affiliated to Shanghai University of Traditional Chinese Medicine (Shanghai, China). All experiments were approved by the Animal Ethics Committee of Putuo Hospital affiliated to Shanghai University of Traditional Chinese Medicine (DWEC-A-2024-16-2-86). To establish the mouse subcutaneous tumor model, 1 × 10⁶ of luciferase-labeled CT26 cells were injected subcutaneously into the right side of 6-8-week-old Balb/c male mice (n = 5). When the tumor volume reached 100 mm³, the mice were randomly divided into 5 groups: 1. Control, 2. β-CDs_CA-HA + NIR, 3. BU, 4. BU-CDs_CA-HA, 5. BU-CDs_CA-HA + NIR. It was noted that β-CDs_CA-HA and BU-CDs_CA-HA was injected intratumorally and NIR (808 nm, 1.0 W/cm²) was irradiated for 5 min on day 0, 2 and 4, while BU group was injected intraperitoneally on day 1, 3 and 5. And the temperature changes were monitored and recorded throughout the process using an NIR thermal camera all groups were monitored and recorded by NIR thermal camera. Subsequently, tumor volumes and the body weight of mice were recorded every two days from the beginning of treatment. When the tumors reached about 2000 mm³, all mice were executed and the tumor tissues were collected for further analysis, including H&E staining, Ki67 proliferation index analysis, TUNEL apoptosis assay, western blot detection of HSPs, HIF-1α and PD-L1 and immunofluorescence analysis of CD4⁺ and CD8⁺ T cells.

Besides. freshly isolated spleens were cut, ground, and filtered through a 70 μm cell filter to prepare cell suspensions. For spleen cells, an additional amount of erythrocyte lysate was required. Afterwards, the cells were incubated with Fc-Block (biolgend) to avoid non-specific binding. To analyze mature DCs in lymph nodes, CD11c-Cy5.5 (biolgend), CD80-APC (biolgend), and CD86-FITC (Invitrogen) staining were used. Additionally, staining with CD3-PE (Abcam), CD4-FITC (Abcam), and CD8a-APC (Abcam) were performed to analyze helper T cells (Ths, CD3⁺CD4⁺) and CTLs, CD3⁺CD8⁺). After washing, the cells were resuspended in stain buffer and then analyzed using flow cytometry (Beckman, USA). Finally, the main organs were collected to carried out H&E staining for proving the biosafety of BU-CDs_CA-HA.

2.9. Evaluation of anti-tumor efficiency and immunity of BU-CDs_CA-HA-loaded Janus membrane

To establish the mouse subcutaneous bilateral tumor model, 1 × 10⁶ of luciferase-labeled CT26 cells were injected subcutaneously into the right side of 6-8-week-old Balb/c male mice. When the subcutaneous tumor volume of mice reached about 100 mm³, they were randomly divided into 4 groups (n = 5): 1. Control, 2. BU-CDs_CA-HA + NIR, 3. Janus, 4. Janus + NIR. The mice in each group were anaesthetized with 2.5 % sodium pentobarbital and underwent tumor resection, leaving about 1 % residual tumor tissue to simulate a realistic residual microscopic tumor tissue during surgery. In addition, colon cancer tumor tissue of the same size as the residual tissue was inoculated subcutaneously on the left side to establish a distal metastatic tumor model. For the Janus group, tumors on the right side were routinely treated with Janus membrane; for the BU-CDs_CA-HA group, BU-CDs_CA-HA solution was injected subcutaneously into the mice. NIR was irradiated for 5 min on day 0, 2, and 4 after surgery (808 nm, 1 W/cm²). Tumor volumes and the body weight of mice were recorded every two days from the beginning of treatment. The formula for calculating tumor volume was: volume = (long diameter × short diameter × short diameter)/2. At the end of experiment, the bilateral tumors were collected for H&E staining, Ki67 proliferation index analysis, TUNEL apoptosis assay, western blot detection and immunofluorescence analysis of CD4⁺ and CD8⁺ T cells.

Freshly isolated spleens were cut, ground, and filtered through a 70 μm cell filter to prepare cell suspensions. For spleen cells, an additional amount of erythrocyte lysate was required. Afterwards, the cells were incubated with Fc-Block (biolgend) to avoid non-specific binding. To analyze MDSC in spleen, CD11b-PE (biolgend) and Gr-1-FITC (biolgend) staining were used.

2.10. In vivo cytokine assay

In the study, serum samples were collected from mouse subcutaneous tumor model and subcutaneous bilateral tumor model and the levels of TNF-α (EK0527), IFN-γ(BDEL-0054), and IL-6 (EK0411) were measured using ELISA kits as the instructions provided by the manufacturer.

2.11. Biosafety

The main organs of mouse, including heart, liver, spleen, lungs, and kidneys were collected to evaluate the tissue integrity. It was fixed in 10 % formalin, embedded in paraffin, and 5 μm sections were prepared. The slices of each group were then stained with H&E and observed under a microscope (Leica). It was noted that the protection of Janus membrane for scald was also performed by subcutaneous material implantation in mice with an 808 nm laser (1 W/cm²) for 5 min, including Control, BU-CDs_CA-HA + NIR, CDs membrane + NIR and Janus membrane + NIR. The skins of mice were recorded by photos.

2.12. Statistical analysis

The data are presented as the mean ± standard deviation (SD). The photos were representative ones from at least 5 samples of every study. One-way analysis of variance (ANOVA) was used to reveal statistical differences of multi-group comparisons (n = 5); pairwise differences were evaluated with two-tailed Student's t tests. Statistical significance was defined as ∗ p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. All statistical analyses were performed using GraphPad Prism 8.0.

3. Results

3.1. Preparation and characterization of BU-CDs_CA-HA

3.1.1. Chemical structure of β-CDs_CA-HA

In order to successfully prepare drug-loaded carbon dots (BU-CDs_CA-HA), it was first to synthesize carbon dots CDs_CA-HA modified by β-cyclodextrin (β-CDs_CA-HA), whose appearance and chemical structuren Fig. S1A and Fig. 1A. Fig. 1B revealed that the chemical shift at 4.89 ppm was belonged to the proton peak signal of H₍₁₎ in β-cyclodextrin, and the chemical shift at 3.98–3.50 ppm was belonged to the proton signal peak of H₍₂₎-H₍₆₎ in β-cyclodextrin. At the same time, the FT-IR spectrum showed that the broad peaks in the range of 3000–3700 cm⁻¹ in CDs_CA-HA were attributed to N-H, C-H and O-H stretching vibrations after hydrothermal reaction and the absorption peak at 1656 cm⁻¹ was attributed to the carboxyl group on the surface of CDs_CA-HA. And the signal peak at 1556 cm⁻¹ were attributed to the formation of C=C by dehydration of CA and HA, which was the main structure of CDs_CA-HA. Furthermore, the absorption peaks at 1027 cm⁻¹, 1380 cm⁻¹ and 1650 cm⁻¹ were attributed to C=O, N-H and -CONH, respectively, indicating the successful modification of β-cyclodextrins (Fig. 1C). These results illustrated that there were a large number of reactive groups in CDs_CA-HA and β-CDs_CA-HA, which endowed the carbon dots with good water solubility and fluorescence properties. As shown in Fig. 1D, the broad peak at 25^o in the XRD spectrum of β-CDs_CA-HA indicated its internal graphene-like structure. Fig. 1E and Fig. S1B–D explained the surface structure of β-CDs_CA-HA, consisted of C_1s (285.1 eV: 284.7 and 288.4 eV, attributed to C=C/C-C and C=O/C=N), N_1s (400.0 eV: 399.5 and 400.6 eV, attributed to N-H and N-C), and O_1s (532.1 eV: 530.8 and 532.2 eV, attributed to C=O and C-O). The above results are consistent with the FT-IR results.

Fig. 1.

The characterization of CDs: (A) the chemical structure scheme, (B) ¹H NMR and (C) FT-IR spectra of β-CDs_CA-HA; (D) XRD and (E) XPS spectra of β-CDs_CA-HA; (F) Encapsulation rate (EI) and drug loading rate (DL) of BU-CDs_CA-HA, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; (G–K) TEM photos of and size measurement of β-CDs_CA-HA and BU-CDs_CA-HA. Bar: 20 nm for G and I, 5 nm for H and J; (L–N) The photothermal conversion and stability of BU-CDs_CA-HA; (O, P) Singlet oxygen generation of CA and BU-CDs_CA-HA under NIR irradiation and dark field.

3.1.2. BU-loading efficiency

Due to the host-guest interaction between the β-cyclodextrins and BU, as well as the good solubility of carbon dots, BU-CDs_CA-HA could improve the bioavailability of BU in the in vivo environment (Fig. S1E). To explore the optimal encapsulation (EI) and drug-loading (DL) of BU-CDs_CA-HA, different mass ratios of BU to β-CDs_CA-HA were tested, and the concentration standard curve of BU was built in Fig. S1F. As shown in Fig. 1F, EI increased with the addition of β-CDs_CA-HA, while DL showed a trend of increase first and then decrease. Because more β-CDs_CA-HA meant more BU could be loaded in carbon dots, indicating the increasing EI. However, it also reduced DL after peak value due to the excessive proportion of β-CDs_CA-HA mass. In this paper, when the mass ratio of BU:β-CDs_CA-HA reached 1:10, both EI and DL the optimal value, 91.2 ± 1.85 % and 8.4 ± 0.44 %, respectively. Compared to the conventional BU-loaded liposomes and polymer nanoparticles whose EI and DL were about 80 % and 4 %, the drug loading capacity of BU-CDs_CA-HA had significant advantages [22]. Therefore, the mass ratio of BU:β-CDs_CA-HA at 1:10 was chosen for all subsequent experiments and are denoted as BU-CDs_CA-HA.

3.1.3. Morphology of BU-CDs_CA-HA

TEM was carried out to observe the morphological changes before and after drug loading of carbon dots. Fig. 1G–J proved that both β-CDs_CA-HA and BU-CDs_CA-HA were spherical nanoparticles and well-dispersed in aqueous solution. As shown in Fig. 1K, the size of BU-CDs_CA-HA increased slightly after loading BU compared to β-CDs_CA-HA, and the size distribution was homogeneous in both cases, proving the good batch stability.

3.1.4. Photothermal and photodynamic properties of BU-CDs_CA-HA

To examine the photothermal properties of BU-CDs_CA-HA, heating curves of BU-CDs_CA-HA solution with different concentrations (0, 0.5, 1, and 2 mg/mL) were recorded under irradiation of 808 nm NIR laser (1 W/cm², 5 min). The results showed that BU-CDs_CA-HA exhibited both time and concentration-dependent photothermal behaviors (Fig. 1L). The temperature of 2 mg/mL samples increased from 24.0 °C to 53 °C under irradiation, while the PBS was only increased from 24.0 °C to 29.7 °C under the same condition (Fig. 1M). Moreover, the heating-cooling cycle curves also proved that the BU-CDs_CA-HA solution still could reach 48 °C stably under NIR after 10 cycles, demonstrating a good photothermal stability (Fig. 1N). On the other hand, CA is a photosensitive small molecule that could be dissociated to produce ROS upon irradiation with NIR. Fig. 1O–P showed both CA and BU-CDs_CA-HA could produce singlet oxygen under NIR. These results suggested that BU-CDs_CA-HA possessed the potential for both PTT/PDT synergistic therapy.

3.2. Preparation and characterization of Janus membrane

3.2.1. Synthesis and characterization of PEG-SS₄ and poly(lipoic acid)

In order to successfully construct Janus membranes through interfacial polymerization, lipoic acid was first grafted to the end of PEG-OH₄ to synthesize PEG-SS₄, taking advantage of ring-opening polymerization of disulfide bond under UV light (Fig. S2A). Fig. S2B showed that peak a at 1.62 ppm and peak b at 1.74 ppm are attributed to the proton signals of the methylene group on the carbon chain of lipoic acid (4H, -CH₂), respectively, whereas the peak c1 and c2 at 1.96 ppm and 2.45 ppm are attributed to the proton signals in methylene group on the dithiolane of lipoic acid (1H, -CH₂). It was also observed that PEG-SS₄ showed a distinct characteristic absorption peak of disulfide bonding at 280 nm compared with PEG-OH₄ in the UV absorption spectrum, thus indicating the successful synthesis of PEG-SS₄ (Fig. S2C).

Lipoic acid was found to exhibit good stickiness after polymerization. Therefore, it could be chosen to construct tissue-adhesive Janus membrane. In Fig. S2D, the narrow spectrum of S_2p could be divided into two significant energy peaks, 164.6 eV and 163.4 eV, which were attributed to the S-S bond and the C-S bond in the dithiolane, while the original two energy peaks became significant three peaks (164.1 eV, 163.0 eV and 161.7 eV) attributed to the S-S bond, C-S bond and C-S-C bond in the main chain of poly(lipoic acid) after polymerization (Fig. S2E).

The adhesion property of poly(lipoic acid) (PLA) solution with different concentration was characterized to measure the loading force and shear modulus by overlapping shear test at room temperature (Fig. S2F). Fig. S2G exhibited that the adhesive strength increased significantly with the up-regulated concentration of PLA solution and the shear force was 1.3 ± 0.16 N, 6.9 ± 1.35 N, 23.3 ± 2.16 N and 50.4 ± 2.52 N. Accordingly, the shear modulus was calculated from the shear and tensile displacement curves and obtained as 3.3 ± 1.16 kPa, 10.6 ± 3.53 kPa, 173.2 ± 8.66 kPa and 90.0 ± 4.50 kPa (Fig. S2H). From the above results, 40 % of PLA showed the highest shear modulus and further increase in concentration rather might weaken the shear modulus. Thus, the feed rate of PLA in Janus membrane was chosen as 40 %.

3.2.2. Gelation of Janus membrane

To explore the optimal proportion of two layers in Janus membrane preparation, the gelation kinetics of hydrogels (DI water) and organogel (EA) with different composition ratio was tested under UV irradiation through rotational rheometer (Table S1). As shown in Fig. 2A–F, all groups could occur cross-linking under UV irradiation due to the ring-opening polymerization of dithiolane in LA. The highest storage modulus (G′) of hydrogels and organogel were achieved when the concentration of PEG-SS₄ reached 0.3 g/mL (DI-2 and EA-2). Meanwhile, the further increase of PEG-SS₄ would not improve the G’ obviously. Based on this result, the formulation of DI-2 and EA-2 was chosen for the following Janus membrane preparation. At the same time, the shear resistance of the Janus membrane was decided by the synergistic effect of PLA concentration and mechanical properties of the membrane (Fig. 2G–H). The maximum load force and shear modulus were 20.65 ± 1.56 N and 220.73 ± 11.4 kPa respectively, in agreement with the trend of the gelation kinetics. It indicated that the 30 % PEG-SS₄ brought the strongest hydrogel network to cooperate with PLA, thus producing good adhesion.

Fig. 2.

(A–F) The gelation kinetics of 2 different parts of Janus membrane in EA and DI solution respectively; (G, H) Shear force and modulus of hydrogels with 40 % PLA and different PEG-SS₄ content by lap shear test to evaluate adhesion, ∗∗∗p < 0.001; (I) The preparation scheme of Janus membrane through interfacial polymerization under UV irradiation; (J–L) Digital and SEM photos of PEG layer, adhesion layer and cross section of Janus membrane. Bar: 5 mm for J, 200 μm for K and 50 μm for L; (M) Release curves of BU from Janus membrane with or without CDs under first-order kinetic model; (N) Release curve of BU from Janus membrane with CDs after 24 h under linear fitting; (O) The digital photos of Janus membrane under the skin of mice at different time points, bar: 4 mm; (P) H&E staining of the skin covered the Janus membrane samples in mice models, bar: 200 μm.

3.2.3. Characterization of Janus membrane

Janus membrane was prepared by interfacial polymerization between EA solution of PEG-SS₄ and DI water solution of PEG-SS₄, BU-CDs_CA-HA and PLA under UV irradiation (Fig. 2I). As shown in Fig. 2J–K, the appearance of PEG layer was white and translucent, while the adhesion layer was brown due to BU-CDs_CA-HA, indicating its photothermal conversion ability. In SEM photos, the cross section of PEG layer was relatively rough compared with adhesion layer which was smooth, resulting from the different solubility of PEG-SS₄ in EA and DI. In Fig. 2L, it was obvious that the interfacial section was appressed without delamination. And it was apparent that there was a transition between two layers in SEM photos.

Fig. 2M–N illustrated the drug release kinetics of Janus membrane, it was evident that the Janus membrane without CDs as BU carriers (Free BU) exhibited a burst release behavior resulting from that BU molecules could diffuse directly into the external solution due to solvent exchange. And about 80 % of BU was released in 24 h. Whereas, BU in the Janus membrane containing BU-CDs_CA-HA needed to break through the hydrogel-CDs double barriers to diffuse into the external solution. And only 47 % of BU was released in the first 24 h, significantly reducing the burst release behavior. At the later stage, there was a linear and constant release curve which was expected to realize the long-term controlled release of BU.

To evaluate the biodegradation of Janus membrane, the in vivo subcutaneous degradation in mice models were built. As shown in Fig. 2O, subcutaneous implantation of Janus membrane began to collapse in vivo after 1 week and gradually degraded until disappearing completely after 5 weeks. H&E staining of the skin from implantation sites of mice showed good structural integrity throughout the entire period. It was clear that there was no obvious necrosis, oedema and inflammation of skin in 5 weeks (Fig. 2P) resulting from a fact that PEG and poly(lipoic acid) were often used in biomedical field as non-immunogenic biomaterials. And their degradation product could be excreted quickly from the body due to good water-solubility. Thus, the prepared Janus membrane showed the outstanding biodegradability and biocompatibility without causing immune rejection in vivo.

3.3. In vitro anti-tumor activity of BU-CDs_CA-HA

The cytotoxicity of β-CDs_CA-HA and β-CDs_CA-HA-loaded Janus membrane was assessed at first to prove the biosafety of materials. Fig. 3A indicated that β-CDs_CA-HA and β-CDs_CA-HA-loaded Janus membrane had no remarkable lethality to CT26 cells, as well as HCT116 cells. Meanwhile, the similar results were also shown on normal cells, RAW 264.7 and FHC cells (Fig. S3A). The cell viability of CT26, HCT116, RWA 264.7 and FHC cells was all higher than 90 % in all groups, which demonstrated the good biocompatibility of β-CDs_CA-HA and β-CDs_CA-HA-loaded Janus membrane. Further, no obvious hemolysis was observed in hemolysis test, proving the good blood compatibility of β-CDs_CA-HA and BU-CDs_CA-HA (Fig. 3B). On the other hand, the tumor-affinity of BU-CDs_CA-HA was important to the anti-tumor effect, due to the existence of HA which is specific to the CD44 receptor on the tumor surface. The uptake assay confirmed that BU-CDs_CA-HA could enter into both CT26 and HCT116 cells. And the strongest intracellular fluorescence intensity was observed after 6 h (Fig. 3C and Fig. S3B–D). The results provided data support for the surgical procedures of in vivo experiments.

Fig. 3.

The biosafety, tumor-targeting and anti-tumor effect of carbon dots: (A) The cell viability of CT26 and HCT116 cells co-cultured with β-CDs_CA-HA and β-CDs_CA-HA-loaded Janus membrane; (B) The blood compatibility of β-CDs_CA-HA and BU-CDs_CA-HA; (C) Specific absorption of BU-CDs_CA-HA co-cultured with CT26 cells (blue: DAPI, green: BU-CDs_CA-HA), bar: 20 μm; (D) Immunofluorescent staining of ROS CT26 cells with different treatments, bar: 20 μm; (E) The anti-tumor ability of BU-CDs_CA-HA under NIR irradiation through synergistic effect of BU and PTT/PDT, ∗∗p < 0.01, ∗∗∗∗p < 0.0001; (F, G) Live/Dead staining photos of CT26 cells with different treatments and the quantitative analysis, bar: 200 μm for F; (H, I) Western blot test of heat shock proteins (HSP70, HSP90), HIF-1α and PD-L1 level and the quantitative analysis, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To investigate the photodynamic ability of BU-CDs_CA-HA in CT26 cells in vitro, the level of cellular ROS in different groups by ROS probe assays was examined. In Fig. 3D and Fig. S3E, there was strong ROS fluorescent signals in CT26 cells in CA + NIR group, β-CDs_CA-HA + NIR group and BU-CDs_CA-HA + NIR group, which meant the PDT potential of BU-CDs_CA-HA compared with the control, 48 °C (only heat) and BU-CDs_CA-HA groups (no ROS). As reported, photosensitive CA could generated ROS under NIR. Thus, β-CDs_CA-HA and BU-CDs_CA-HA prepared from CA also possessed photodynamic property [18,19,[23], [24], [25]]. Subsequently, the in vitro anti-tumor activity of BU-CDs_CA-HA under NIR was evaluated by CCK-8 assay. Fig. 3E proved that single heat or ROS (48 °C and CA + NIR) could not achieve the best outcome. And PTT/PDT (β-CDs_CA-HA + NIR) actually increase the therapy efficiency. However, BU-CDs_CA-HA + NIR had the strongest killing effect on CT26 cells due to the synergistic function of BU and PPT/PDT, resulting in 10.9 ± 1.29 % of survival rate. Meanwhile, Live/Dead staining of CT26 cells in different groups also showed there was a large number of dead CT26 cells (red) in BU-CDs_CA-HA + NIR group, accounting for 74.0 ± 11.02 % of the total cells (Fig. 3F–G). The similar results were achieved in HCT116 cells (Fig. S3F and G). Above data indicated that BU-CDs_CA-HA could significantly promote the apoptosis of colorectal cancer cells under NIR.

Further, the underlying mechanism was studied by Western blot test to analyze the expression level of proteins associated with therapy efficiency. As shown in Fig. 3H–I, the level of HSP70,HSP90, HIF-1α and PD-L1 were inhibited apparently in BU-CDs_CA-HA + NIR group. According to the articles, intracellular HSPs would be over-expressed to reduce the cellular damage caused by thermal effect and intracellular HIF-1α would be up-regulated in an anoxic microenvironment, thus limiting the phototherapy efficiency (Fig. S3H and I) [[26], [27], [28]]. Moreover, HSPs and HIF-1α would synergistically influence the level of PD-L1 which could bind to the programmed death receptor 1 (PD-1) on the surface of T cells, thus transmitting inhibitory signals to suppress the activity, proliferation and cytokine secretion of T cells, reduce the recognition and attack of tumor cells by NK cells and allow tumor cells to evade surveillance and attack of immune system [[29], [30], [31]]. BU, one of the most potent toxic ligands for killing tumor extracted from traditional Chinese medicine venenum bufonis, showed excellent cytotoxicity to CT26 cells. Meanwhile, BU also might promote phototherapy and immunotherapy through down-regulating the expression of HSPs, HIF-1α and PD-L1 [9,32]. Therefore, BU-CDs_CA-HA + NIR showed outstanding in vitro anti-tumor outcome through the BU-PTT/PDT synergistic effect.

3.4. In vitro immune response

PTT and PDT could both induce immunogenic death (ICD) of tumor cells which would release damage-associated molecular patterns (DAMPs), including calreticulin (CRT), high-mobility-group protein B1 (HMGB1) and adenosine triphosphate (ATP), activating the subsequent immune response [[33], [34], [35], [36]]. As shown in Fig. 4A and B, the signal intensity of CRT (green) was significantly enhanced in the groups with phototherapy. Meanwhile, intracellular ATP level was inhibited and extracellular HMGB1 concentration was increased most notably in BU-CDs_CA-HA + NIR (Fig. 4C and D). As a kind of partner protein mainly existing in the endoplasmic reticulum, CRT would be transferred to the cytomembrane and act as a "eat me" signal when ICD occurs. Then, ATP and HMGB1 would be released from the damaged cell sequentially to promote DCs maturation to engulf tumor cells. Thus, according to the above results, BU-CDs_CA-HA were proved to induce ICD of tumor cells effectively through the synergistic effect of BU and PTT/PDT.

Fig. 4.

In vitro immune response of CT26 cells after synergistic therapy of BU and PTT/PDT through BU-CDs_CA-HA under NIR: (A, B) CRT fluorescence labeling and intensity on cytomembrane after different treatments, bar: 100 μm; (C, D) Intracellular ATP level and extracellular HMGB1 concentration of CT26 cells after different treatments, ∗∗∗∗p < 0.0001; (E) The scheme of DCs maturation by supernate of CT26 cells with different treatments; (F) The analysis of DCs maturation by flow cytometer.

As the most powerful antigen-presenting cells (APCs) in the immune system, DCs play the role of an "immune hub" in anti-tumor immunity. They achieve the surveillance and elimination of tumors by recognizing tumor antigens, initiating immune responses, and regulating the immune microenvironment [37,38]. After ICD, DCs would be matured by DAMPs, such as ATP and HMGB1. Thus, in vitro DCs maturation culture was carried out by supernatant of CT26 cells after ICD (Fig. 4E). In Fig. 4F, it was clear that the maturation degree of DCs in BU-CDs_CA-HA + NIR group was the highest, 82.0 % due to the synergistic effect of BU and PTT/PDT, which was in agreement with the results of ICD of CT26 cells.

By reason of the foregoing, BU-CDs_CA-HA-mediated synergistic effect of BU and PTT/PDT not only could cause immediate killing of tumor cells, but also might realize the long-term immunological therapy to prevent the recurrence and metastasis by inhibiting immune escape and activating the anti-tumor immunity.

3.5. In vivo anti-tumor efficiency

3.5.1. In vivo anti-tumor efficiency of BU-CDs_CA-HA

To assess the in vivo anti-tumor efficiency of BU-CDs_CA-HA, Balb/c mice subcutaneous CRC models were built. And the surgery and treatment procedures were shown in Fig. 5A.

Fig. 5.

The evaluation of in vivo anti-tumor efficiency of BU-CDs_CA-HA: (A) The surgery and treatment procedure of subcutaneous CRC tumor models (n = 5); (B) Temperature change of different groups after treatments; (C) Tumor volume change within the therapy period, ∗∗∗∗p < 0.0001; (D) Living imaging of subcutaneous tumors in mice models at 16th day; (E, F) The gross observation and weight analysis of CRC tumors of all groups after euthanizing the mice. The dotted circles mean that the tumors were ablated completely; (G) H&E, Ki67 and TUNEL staining of tumor tissues to assess the proliferation and apoptosis tumor cells, bar: 100 μm for H&E and TUNEL staining and 20 μm for Ki67 staining; (H, I) Western blot test and quantitative analysis of HSPs, HIF-1α and PD-L1, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001.

The photothermal temperature was kept at 48 °C, same as in vitro experiments (Fig. 5B and Fig. S4A). During therapy period, the tumor volume was measured every 2 days to monitor the progress of treatment. In Fig. 5C–D, it was apparent that the tumors were inhibited effectively in β-CDs_CA-HA + NIR and BU-CDs_CA-HA + NIR groups due to PTT/PDT combined therapy and BU-PTT/PDT synergistic effect, respectively. And the tumor signal in BU-CDs_CA-HA + NIR group was almost invisible at the 16th day. More visually, Fig. 5E and F illustrated that the tumor volume in β-CDs_CA-HA + NIR had been decreased obviously compared with control groups, while most tumor tissues were almost disappeared in BU-CDs_CA-HA + NIR group. In addition, the BU-PTT/PDT synergistic effect reduced cell proliferation and increased apoptosis (Fig. 5G). And no physiological morphology change was noticed in the main organs by H&E staining and the body weight of mice was stable, suggesting the biosafety of the prepared biomaterials and therapy strategies (Fig. S4B–C). Similar to the in vitro trials, BU-CDs_CA-HA + NIR group displayed the most reduction of HSPs, HIF-1α and PD-L1 (Fig. 5H–I).

These results not only proved the immediate, safe and efficient anti-tumor outcome by BU-PTT/PDT synergistic effect, but also indicated the promising capacity of long-term immunotherapy.

3.5.2. BU-CDs_CA-HA promoted anti-tumor immunity

The anti-tumor immunity must play a key role during the whole therapy period due to the long-term inhibition of tumor recurrence even though NIR had been stopped after the 4th day. Therefore, for a comprehensive explanation in immune mechanism of BU-CDs_CA-HA, a series of detailed studies in anti-tumor immune response of BU-PTT/PDT synergistic effect were carried out.

DCs, as the most important APC, is the key link to initiate T cell adaptive immunity. As displayed in Fig. 6A and D, the maturation of DCs in lymph nodes was relatively low in Control group (6.6 ± 1.64 %), while the level could be increased by BU (BU: 12.8 ± 3.78 %; BU-CDs_CA-HA: 14.0 ± 2.76 %) and phototherapy (β-CDs_CA-HA + NIR: 27.6 ± 2.34 %), similar as the in vitro trials. Furthermore, BU-CDs_CA-HA + NIR group showed a 42.3 ± 6.18 % DCs maturation which was 6.6-fold higher than that of Control group. In Fig. 6B and E-F, the number of activated CD4⁺ and CD8⁺ T cells in the spleen of the BU-CDs_CA-HA + NIR group was 3.2-fold and 4.2-fold higher than that of Control respectively. In addition, natural killer (NK) cells are another type of effector cells with a direct killing effect to tumor cells. It was clear that the proportion of locally activated NKs increased 5.8 folds in the BU-CDs_CA-HA + NIR group compared to Control group (Fig. 6C and G). These results revealed that BU-CDs_CA-HA-mediated BU-PTT/PDT synergistic effect could effectively boost not only DCs maturation, but also the stimulation of CD4⁺/CD8⁺ T cells and NK cells, thus activating an in vivo immune response.

Fig. 6.

In vivo study in anti-tumor immune response by analysizing the maturation, activation and infiltration of immune-related cells, as well as the level of immune-related cytokines. (A) The maturation of DCs in lymph nodes of mice; (B) The activation of CD4⁺/CD8⁺ T cells in spleens in of mice; (C) The activation of NK cells from spleens of mice. (D–G) The quantitative analysis of flow cytometry, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001. (H) Immunofluorescence photos of CD4⁺ and CD8⁺ T cell infiltrating tumor tissues of mice, bar: 100 μm. (I–K) The quantitative analysis of immune-related cytokines: IL-6, TNF-α and IFN-γ, ∗p < 0.5, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001.

Moreover, immunofluorescence analysis further proved there was a high level of infiltration of CD4⁺/CD8⁺ T cells in tumor tissues in BU-CDs_CA-HA + NIR group, meaning CD4⁺/CD8⁺ T cells could especially and effectively act on tumor tissues after being activated (Fig. 6H). Besides, the cytokine levels of interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), and interleukin 6 (IL-6) were also remarkably up-regulated in BU-CDs_CA-HA + NIR group, demonstrating that BU-CDs_CA-HA-mediated BU-PTT/PDT synergistic effect triggered a powerful anti-tumor immune response (Fig. 6I–K).

All these data indicated that BU-CDs_CA-HA, as BU carriers and key component of phototherapy, could release BU sustainably and realize PTT and PDT under NIR in tumor tissues especially, achieving efficient and immediate tumor-killing outcome. More importantly, BU-PTT/PDT synergistic effect also induced high-level anti-tumor immunoreaction through activating immune response and inhibiting immune escape.

3.5.3. In vivo anti-tumor efficacy of BU-CDs_CA-HA-loaded Janus membrane

To assess the immediate tumor-killing effect and long-term immune therapy of BU-CDs_CA-HA-loaded Janus membrane, the bilateral subcutaneous tumor models were built (Fig. 7A). When the subcutaneous in-situ tumors on the right side of mice reached about 100 mm³, they were removed about 99 % of original volume. The residual tumors were randomly divided into 4 groups: Control, BU-CDs_CA-HA + NIR, Janus and Janus + NIR groups for comparing the tumor inhibition of different treatments. Besides, the contralateral tumors with the same size as the residual in-situ ones were inoculated subcutaneously on the left side to evaluate the immunoreaction efficiency due to a fact that the therapeutic effect on contralateral tumors was completely resulted from in-situ tumor inhibition. And the temperature in BU-CDs_CA-HA + NIR and Janus + NIR groups was kept at 48 °C under NIR (Fig. 7B and Fig. S5A).

Fig. 7.

The assessment of immediate tumor-killing effect and long-term immunoreaction on anti-tumor efficiency induced by Janus membrane: (A) The surgery procedure of bilateral subcutaneous tumor models (n = 5); (B) Temperature change of different groups after treatments; (C, D) The volume change of in-situ and contralateral tumors during the therapy period, ∗p < 0.05, ∗∗∗p < 0.001; (E) Living imaging of in-situ and contralateral tumors at the 18th day; (F–I) The gross observation and weight analysis of in-situ and contralateral tumors of all groups after euthanizing the mice, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; (J) H&E, Ki67 and TUNEL staining to assess the cell proliferation and apoptosis of tumor tissues, bar: 100 μm for H&E, 20 μm for Ki67 and 50 μm for TUNEL; (K, L) Western blot test and quantitative analysis of HSPs, HIF-1α and PD-L1, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

During the whole therapy period, it was apparent that the in-situ and contralateral tumors were inhibited significantly in BU-CDs_CA-HA + NIR and Janus + NIR groups compared with the others. And tumors in Janus + NIR groups were prevented from growth remarkably (Fig. 7C–D). At the 18th day, the living imaging photos also displayed there was no significant signal of in-situ and contralateral tumors in mice of Janus + NIR group (Fig. 7E). After collecting these tumor tissues, gross observation and tumor weight both showed that BU-PTT/-DT synergistic effect could inhibit in-situ and contralateral tumors efficiently at the same time.

It's worth noting that BU-CDs_CA-HA of BU-CDs_CA-HA + NIR group were injected to the surface of in-situ tumors in this experiment, instead of intratumor injection. Thus, it might be removed much more quickly than Janus + NIR group. Whereas, in-situ and contralateral tumors were almost ablated completely in Janus + NIR group due to the persistent local BU-PTT/PDT synergistic effect through firmly adhesion of Janus membrane to the surface of the tumor, as well as the long-term immunoreaction induced by ICD of tumor cells (Fig. 7F–I). The results suggested that Janus membrane-mediated BU-PDT/PTT synergistic effect activated systemic immunity in mice, resulting in a favorable inhibitory effect on in-situ tumors and distal metastasis. The tissue sections of tumors displayed that there were remarkable tumor cell proliferation and little apoptosis in control and Janus groups due to the lack of effective inhibitory method, while the introduction of BU-PTT/PDT synergistic effect combined drug, heat and ROS, as well as the activation of immune response, preventing tumor cell proliferation and enhancing apoptosis (Fig. 7J).

Meanwhile, Janus + NIR also could down-regulate the expression level of HSPs, HIF-1α and PD-L1 significantly, proving the high-efficiency PTT/PDT and enhanced anti-tumor immunity (Fig. 7K–L). Besides, similar to the in vivo anti-tumor efficiency of BU-CDs_CA-HA, there was no physiological morphology change in the main organs by H&E staining and the body weight of mice was stable, proving the biocompatibility of BU-CDs_CA-HA-loaded Janus membrane (Fig. S5B–C).

Last but not the least, the skin in the local area of the tumor would be damaged due to the high temperature of PTT, which was observed in the in vivo anti-tumor efficiency of BU-CDs_CA-HA. However, Janus membrane could protect the tumor-local skin from empyrosis through the PEG layer which prevent the heat conduction. Fig. S5D showed that BU-CDs_CA-HA + NIR and CDs membrane + NIR both scalded the skins because of the direct contact with BU-CDs_CA-HA and CDs membrane of 48 °C, while the skin in Janus membrane + NIR was almost undamaged, suggesting its thorough biosafety and practicality.

3.5.4. Promoted anti-tumor immunity of Janus membrane

To evaluate the anti-tumor immunity induced by BU-PTT/PDT from Janus membrane under NIR, immune-related indicators were analyzed in detail. As shown in Fig. 8A–G, the cell expression level of CD80⁺/CD86⁺ was significantly increased in the Janus+NIR group compared with the others, meaning the higher maturation of DCs. And the CD4⁺/CD8⁺ splenic infiltrating T cells had the most significant up-regulation in the proportion of T cells, with 2.7-fold and 2.6-fold higher than CD3⁺/CD4⁺ T cells and CD3⁺/CD8⁺ T cells of Control group. Besides, there was a 5.4-fold rise in the proportion of locally activated NKs in the Janus + NIR group compared to Control group, suggesting Janus-mediated BU-PTT/PDT synergistic effect triggered an innate immune response in vivo. On the other hands, the tumor microenvironment (TCM) is immunosuppressive by releasing soluble cytokine mediators and attracting immunosuppressive cell types, such as myeloid-derived suppressor cells (MDSCs).

Fig. 8.

In vivo anti-tumor immune mechanisim analysis of BU-CDs_CA-HA-loaded Janus membrane under NIR: (A–I) Flow cytometry and quantitative analysis of DCs, CD4⁺/CD8⁺ T cells, NK cells and MDSCs, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; (J) Immunofluorescent staining of CD4⁺/CD8⁺ T cells in tumor tissues to illustrate the enhanced tumor-specific identification and infiltration of T cells, bar: 50 μm; (K–L) Elisa test of proinflammatory cytokin (IL-6, TNF-α and IFN-γ) to assess the inflammatory activity, ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Depletion of immunosuppressive cells indicates an improvement of anti-tumor efficacy. Therefore, the proportion of MDSCs from spleens of mice was analyzed. As shown in Fig. 8H–I, the proportion of MDSCs in the Janus + NIR group was reduced by 2.9-fold compared with Control group, which meant that BU-CDs_CA-HA-loaded Janus membrane could not only alleviate the immunosuppressive microenvironment, but also improve the anti-tumor immune response through BU-PTT/PDT synergistic effect. Meanwhile, immunofluorescence analysis proved the infiltration of CD4⁺ and CD8⁺ T cells to tumor tissues in Janus+NIR group (Fig. 8J). Furthermore, there was enhanced expression of pro-inflammatory cytokines, including IFN-γ, IL-6 and TNF-α, in Janus + NIR group (Fig. 8K–M), indicating the active inflammatory activities due to the Janus membrane-mediated immune response.

Considering the above results, the volume and weight of tumors managed to be dramatically decreased by Janus membrane under NIR. Meanwhile, the in vivo systemic anti-tumor immunity due to ICD of tumor cells was obviously promoted by BU-PTT/PDT synergistic effect. More importantly, the immunosuppression microenvironment was also significantly alleviated, which was crucial to maintain a long-term anti-tumor immunoreaction. Therefore, it could be concluded that the prepared BU-CDs_CA-HA-loaded Janus membrane could not merely realize the immediately killing effect to tumors, but also induce the long-term anti-tumor immunoreaction to prevent recurrence and metastasis, thus strengthening the therapeutic outcomes.

4. Discussion

CRC presented significant challenges in clinical therapy owing to its high incidence and delayed screening. The current clinical methods might induce residual tumors, multi-drug resistance, radiotherapy resistance and immune rejection at the middle or late stage of disease, resulting in the high risk of recurrence and metastasis [39,40]. Phototherapy have been proved to show an immediate tumor-killing effect, including photothermal and photodynamic therapies (PTT/PDT). However, heat shock proteins (HSPs) and HIF-1α secreted by tumor cells from high temperature and insufficient oxygen supply of tumor microenvironment (TME) both limited the therapeutic efficiency. Meanwhile, the enhancement of HSPs and HIF-1α would further up-regulate the expression level of PD-L1 which would bind to PD-1 receptor of immune cells to mediate immune escape, thus impairing the long-term anti-tumor immune activity [41,42]. Traditional Chinese medicine has always been seen as a promising strategy to solve the limited therapeutic outcomes in cancers because of its broad-spectrum property and low drug resistance. Moreover, it can regulate tumor immune responses through multiple pathways, targets and levels to kill tumor cells and facilitate immunotherapy [43,44]. Thus, how to construct an integrated material with immediate tumor-killing effect and long-term anti-tumor immunity is still an urgent demand. In our work, BU-CDs-_CA-HA exhibit remarkable drug-loading capacity and sensitive photothermal/photosensitive properties (Fig. 1F and 1L–P). After combination with Janus membrane which showed a double-layer structure from interfacial polymerization, the multifunctional system could locally adhere to residual tumors to perform synergistic anti-tumor strategy, containing drug, heat and ROS. At the same time, the other layer of Janus membrane could protect surrounding tissues from thermal damage due to its hydrogel network. As the active component, BU-CDs-_CA-HA exhibit good biocompatibility, ultrasmall size (<10 nm) and tumor targeting, which could specifically inhibit proliferation and promote apoptosis of tumor cells through BU-PTT/PDT synergistic effect, followed by inducing the ICD of tumor cells and thus initiating and enhancing anti-tumor immunoreaction (Fig. 3, Fig. 4). Overall, BU-CDs-_CA-HA-loaded Janus membrane showed immediate and intensive tumor-killing effect and long-term anti-tumor immunoreaction to prevent recurrence and metastasis of CRC.

To elucidate the underlying anti-tumor mechanism of BU-PTT/PDT synergistic strategy, we conducted comprehensive in vitro trials and tumor models. It revealed that BU-CDs_CA-HA-loaded Janus membrane could locally adhere to residual tumor to release BU-CDs_CA-HA and perform PTT/PDT, inducing the death of tumor cells efficiently through BU, thermal damage and ROS, while the high temperature was controlled by the other layer of Janus membrane to protect surrounding tissues. Meanwhile, BU-CDs_CA-HA decreased the protein level associated with PTT/PDT resistance and immune escape, including HSP70, HSP90, HIF-1α and PD-L1. HSPs and HIF-1α are secreted by tumor cells when them are in a high-temperature and hypoxia TME respectively, which limited the therapeutic efficiency. Besides, the simultaneous up-regulation of HSPs and HIF-1α would further induce the rise of PD-L1 which would bind to PD-1 receptor of T cells to weaken the systemic immunity and enhance immune escape. Thus, inhibition of HSPs and HIF-1α expression could not only strengthen PTT/PDT outcomes, but also improve systemic immunity through subsequent decrease of PD-L1 [45]. The results suggested that BU exerted obvious down-regulation effect on HSP70, HSP90 and HIF-1α, as well as the following PD-L1, indicating the muti-function of BU-CDs_CA-HA which might be the novel therapeutic strategy of CRC postoperative treatment.

For tumor therapy, the initiating and maintaining of immunoreaction should be both paid a large attention to and equally important as immediate tumor-killing effect owing to its close relation to long-term recurrence and metastasis [46]. Based on reports and our data, BU-CDs_CA-HA from Janus membrane could induce immunogenic cell death (ICD) of tumor cells to release DAMPs to initiate immunoreaction through BU-PTT/PDT synergistic effect. Thus, a large amount of matured DCs, CD4⁺/CD8⁺ T cells and NK cells were activated. Moreover, given the inhibition effect of BU on HSPs, HIF-1α and PD-L1, more CD4⁺/CD8⁺ T identified and attacked tumor cells. And the MDSCs were reduced, which could have produced an immunosuppressive TME. Animal models also proved both local tumors and distal tumors were inhibited and diminished. These results indicated that there was an active and long-term anti-tumor immunoreaction following immediate tumor-killing effect closely.

Compared with the current strategy for CRC postoperative therapy, including radiotherapy, chemotherapy and immunotherapy, BU-CDs_CA-HA-loaded Janus membrane exhibited some advantages such as non-toxicity, biocompatibility, immediate tumor-killing effect and enhanced long-term anti-tumor immunity. In the future, it still needs cooperation with more advanced clinical technologies to realize clinical translation, while these findings have built a strong foundation for its clinic potential and provided a new platform for expanding the modern application of TCM.

5. Conclusion

This work proposed a biocompatible BU-CDs_CA-HA-loaded Janus membrane with sustained and tumor-specific release of BU, as well as phototherapy, to provide immediate tumor-killing effect and long-term anti-tumor immunoreaction for CRC inhibition. Janus membrane was built through interfacial polymerization and could tightly attach to the tumors. BU-CDs_CA-HA loaded BU through β-cyclodextrin, improving the solubility and bioavailability of drug. Besides, PTT and PDT were combined under NIR due to photothermal conversion of the dark brown appearance and photosensitiveness of CA. In vitro and in vivo data both revealed BU-PTT/PDT synergistic effect could promote apoptosis, inhibit proliferation and induce ICD of CT26 cells, thus inhibiting tumors and activating anti-tumor immune response. More importantly, BU could decrease the expression of PD-L1 and the proportion of MDSCs, reducing the risk of immune escape and alleviating the immunosuppressive TME, which are benefit to maintain an anti-tumor immunoreaction in a long term. In conclusion, the offense-defense integrated strategy not only improved therapeutic outcomes, but also expanded the application field of traditional Chinese medicine.

CRediT authorship contribution statement

Xi Yu: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Yujie He: Writing – review & editing, Writing – original draft, Resources, Methodology, Investigation, Conceptualization. Jiani Zhong: Investigation, Formal analysis. Yue Li: Software, Formal analysis, Data curation. Yuan Zhu: Software, Formal analysis. Yingying Chen: Visualization, Data curation. Jingbo Yin: Writing – review & editing, Validation, Supervision. Zeting Yuan: Writing – review & editing, Validation, Supervision. Peihao Yin: Writing – review & editing, Validation, Supervision, Resources.

Ethics approval and consent to participate

All the mice used in the experiment were carefully bred at Putuo Hospital of Shanghai University of Traditional Chinese Medicine. The experiment followed the ethical norms and received the ethical approval number: DWEC-A-2024-16-2-86.

Declaration of competing interest

I, Xi Yu, declare that there are no conflicts of interest in relation to the manuscript titled "Immediate tumor killing and long-term anti-tumor immunoreaction induced by Bufalin-loaded phototherapeutic Janus membrane in CRC postoperative therapy" submitted to Materials Today Bio. I confirm that the results and interpretations reported in the manuscript are original and have not been plagiarized.

I certify that I have read and understand the [Materials Today Bio] conflict of interest policy, and I understand that failure to disclose a conflict of interest may result in the manuscript being rejected or retracted.

I also certify that I have disclosed any financial or non-financial relationships that may be interpreted as constituting a conflict of interest in relation to this manuscript. I understand that this information will be subject to peer review, and I am willing to provide further information or clarification if required.

I confirm that I have no known conflicts of interest that would influence the results or interpretation of the data presented in this manuscript, and I understand that failure to disclose a conflict of interest is unethical and may result in sanctions being imposed on me.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

All data are available in the main text, supporting information and are also on request from the corresponding author.