Tumor Cell-Targeting and Tumor Microenvironment–Responsive Nanoplatforms for the Multimodal Imaging-Guided Photodynamic/Photothermal/Chemodynamic Treatment of Cervical Cancer

Introduction

Cervical cancer is the fourth most common gynecological malignancy in the world, causing more than 300,000 deaths annually.¹ Currently, its clinical treatment relies on surgery.² For young patients with reproductive needs, however, this approach is unsatisfactory, as cervical conization increases the risk of miscarriage due to cervical insufficiency and radical surgery renders women infertile.^3,4 Chemotherapy is a necessary adjuvant treatment for cervical cancer, but it has poor efficacy due to treatment tolerance and severe toxic side effects.⁵ Thus, the development of new therapeutic strategies for cervical cancer is of great clinical importance.

Given that the cervical anatomical structure is connected to that of the vagina, the use of optical therapies for cervical cancer have attracted extensive research attention.^6,7 Phototherapy, or the administration of photothermal therapy (PTT) and photodynamic therapy (PDT) via near-infrared (NIR) light exposure using photosensitizers (PSs) and photothermal agents, has been used widely in antitumor research.^8–10 As a local treatment method, it has the advantages of non–direct invasiveness, low drug resistance, real-time diagnosis, and good tissue selectivity; it enables photosensitive and photothermal material accumulation at irradiation sites without damaging surrounding normal tissues.^11–13 In PTT, a photothermal reagent converts absorbed NIR light energy into effective thermal energy for the killing of tumor cells through thermal ablation.^14,15 In PDT, PSs convert surrounding molecular O₂ into cytotoxic reactive oxygen species (ROS) under laser irradiation at specific wavelengths to drive cell death.^16–18 PTT and PDT have synergistic effects.^19,20 On one hand, increased hypoxia and acidification of the tumor microenvironment after PDT increase the sensitivity of tumor cells to PTT, and reduced blood flow to the tumor reduces the “heat sink” effect, which aids the achievement of the photocoagulation temperature. On the other hand, the heat generated by PTT increases blood flow, improves the O₂ supply, and significantly enhances PDT-induced injury.²¹ Just as laser irradiation with a single light source can facilitate clinical operations, the synergistic use of PDT and PTT against cervical cancer with a single light source is potentially translatable to the clinical setting.

With advances in the field of phototherapy, agents such as protoporphyrin IX, dihydrochlorin e6, and methylene blue (MB) have been approved for the clinical treatment of skin, esophageal, and lung tumors.^9,22 However, the absorption spectra of these PSs fall primarily within the visible light range (<700 nm), and high degrees of light absorption and scattering by biological tissues significantly diminish their effectiveness in deep tissue treatment. NIR light, which falls within the biological transparency window of 700–1000 nm, exhibits deeper tissue penetration and lower absorption rates.^23–25 Thus, the use of PSs with longer wavelength absorption characteristics to enhance phototherapeutic effects warrants consideration.

Indocyanine green (ICG) has a strong absorption peak between 780 and 800 nm, and thus good absorption capacity in the NIR.²⁶ Due to its low toxicity, high affinity, and unique optical properties, ICG has been approved by the US Food and Drug Administration for clinical application as a NIR organic dye with photothermal and photodynamic capabilities under single-light-source irradiation.^27–30 However, the practical application of ICG entails several challenges related to its propensity for photobleaching, low photothermal conversion efficiency, and poor photostability.^31–34 For instance, under NIR laser exposure, the quantum yield of singlet oxygen (¹O₂) generated by ICG monomers is only 0.2%.²³ Furthermore, intravenously injected ICG binds readily to plasma proteins, such as albumin and high-density (eg, α-1) lipoproteins, leading to its rapid clearance from the body and a half-life of only about 2–4 min.^35,36 ICG also lacks specificity, limiting its accumulation in tumor tissues and thereby its application in tumor phototherapy.^37,38 Thus, the development of new nanomedicines to address these issues, enhancing ICG’s stability, retention time in the body, and therapeutic effects against tumors, is urgently needed.

Recent advances in nanotechnology have made the combined use of nanomaterials and phototherapy techniques a viable approach for tumor treatment. To improve the therapeutic advantages of ICG and the effect of combined PTT/PDT, researchers have constructed novel composite materials. Xue et al³⁴ grafted ICG onto nanocarriers with magnetic (Fe₃O₄) nanoparticle cores and Prussian blue nanoshells (formed in situ) using polyethylenimine modification and electrostatic adsorption to improve drug-carrying efficiency and stability, which are conducive to ICG accumulation in tumors. This approach effectively solved the problems of free ICG molecules’ poor optical stability and low bioavailability. Chen et al³⁹ developed a Pd1.7 Bi@CeO₂–ICGcore–shell structure with dual enzyme activity, and reported that the ICG-loaded Pd1.7Bi@CeO₂ exhibited peroxidase- and catalase (CAT)-like activities and synergistic photothermal/photokinetic/chemokinetic antitumor efficacy. Its CAT-like activity resulted in in-situ O₂ generation, which alleviated tumor hypoxia and improved the PDT efficiency of the coupled ICG. Cheng et al⁴⁰ prepared a nanoscale Cu/Zn metal organic framework (MOF) by introducing Cu²⁺ ions into the precursor of ZIF-90 and explored the formation of mixed-valence (Cu^+/2+) hollow porous structures upon heating and mixed-metal, mixed-valence (Cu^+/2+/Mn^2+/4+) structures upon heating with the addition of manganese acetylacetonate [Mn(acac)₂]. They loaded the hollow structure with ICG (ICG@Mn/Cu/Zn-MOF@MnO₂) for multimodal imaging-guided PTT/PDT/chemodynamic therapy (CDT). These studies have shown that the optical stability of ICG can be improved effectively by encapsulation in nanocarriers, enabling full leveraging of this dye’s dual therapeutic advantages and improving its antitumor efficacy. The structural complexity of these multifunctional nanomedicines, however, increases the difficulty of their clinical translation. Thus, the simplification of drug carrier structures, enhancement of drug delivery efficiency, and full leveraging of combined therapeutic effects continue to be challenges.

In this study, we used hollow mesoporous manganese dioxide (HMnO₂), which has a high specific surface area and good tumor microenvironment responsiveness, as a carrier. It was modified using poly (allylamine hydrochloride) (PAH) to exploit its cationic properties for reversing the charge of manganese dioxide, thereby overcoming the electrostatic repulsion with the negatively charged ICG molecules. Additionally, hydrogen phosphate ions were added to create a counterion cohesion effect, stabilizing the surface charge state of the material, enhancing adsorption sites, and significantly enhancing the adsorption and encapsulation of ICG molecules.^35,36,41 With this design, efforts were made to enhance the stability of ICG and utilize the CAT-like activity of the carrier to decompose high concentrations of H₂O₂ in tumors into O₂, sensitizing the PDT effect of ICG under laser irradiation. Additionally, with the integration of PTT, the chemodynamic properties of the carrier were enhanced to amplify oxidative stress. The coating of the NP surfaces with HeLa cell membranes enabled the NPs to achieve immune evasion and target homologous tumors. An integrated NP platform was constructed to synergize the effects of PDT/PTT/CDT and exert multimodal imaging functions (Scheme 1). An in-vivo mouse model confirmed that M-HMnO₂@ICG exerts optimal antitumor treatment effects under NIR irradiation This simplified, cost-effective, tumor microenvironment–responsive integrated nanoplatform thus provides a basis for clinical translation.

Scheme 1 Schematic diagram of M-HMnO2@ICG preparation and multimodal imaging-guided PTT/PDT/CDT.

Materials and Methods

Materials

Tetraethyl orthosilicate (TEOS), PAH (molecular weight ≈15,000), potassium permanganate, sodium carbonate, and ammonia (NH₃·H₂O) were purchased from Innochem. ICG was purchased from ARK Pharmaceuticals. Hoechst 33,342, phenylmethylsulfonyl fluoride (PMSF), cell counting kit 8 (CCK8), and bicinchoninic acid (BCA) protein assay kits were purchased from Biosharp. Cytoplasmic protein extraction and live-staining dead-dye kits were purchased from Beyotime Biotechnology. All cell culture reagents were purchased from Procell. Deionized water was prepared using a Milli-Q system (Millipore, USA).

HMnO₂@ICG Preparation

First, solid SiO₂ NPs were synthesized using a previously reported method.⁴² They were washed three times with water and ethanol and then stored in water until use. To synthesize SiO₂@MnO₂ NPs, 300 mg KMnO₄ dispersed in 10 mL water was added dropwise to the NPs, followed by 1 h ultrasound treatment and then stirring overnight at room temperature. The obtained SiO₂@MnO₂ NPs were washed three times with deionized water and centrifuged at 12,000 rpm, then dissolved in a Na₂CO₃ solution at 60°C for 12 h to obtain HMnO₂ NPs. For ICG loading, these NPs (10 mg) were added to a solution containing 10 mL PAH (5 mg/mL), followed by stirring for 2 h, centrifuging at 12,000 rpm, and washing with water. Phosphate buffer (PB; 20 mM) was added to the obtained H-MnO₂/PAH solution, followed by the addition of ICG in PB (0.3 mg/mL), mixing, and stirring in the dark for 2 h. ICG was then loaded into the solution with centrifuging at 12,000 rpm for 10 min. The resulting HMnO₂@ICG mixture was washed three times with deionized water. To further investigate the effects of the loading condition on the drug encapsulation and loading rates, we used HMnO₂, HMnO₂/PAH, and HMnO₂/PAH/HPO₄^2- solutions as carriers. We added an aqueous ICG solution with the same content and stirring under the same conditions at a feeding ratio of 2:1, then collected and centrifuged the supernatant. According to the reported standard curve method,⁴³ the ICG content of the supernatant was measured using an ultraviolet/visible-near infrared (UV/Vis-NIR) spectrophotometer (UV-2600; Shimadzu Corporation, Japan) under HMnO_2, HMnO₂/PAH, and HMnO₂/PAH/HPO₄^2- addition conditions. The ICG loading capacity (LC) and encapsulation efficiency (EE) of the NPs were calculated using the following equations:

M-HMnO₂@ICG Preparation

To obtain HeLa cell membrane vesicles, human cervical cancer cells were centrifuged for 5 min (1500 rpm, 4°C) and the pure sediment was collected and resuspended in hypotonic lysis buffer containing membrane protein extraction reagent and PMSF. After 1 h incubation in an ice bath, the HeLa cells were transferred to a glass homogenizer for 40 mechanical disruptions, and the resulting mixture was centrifuged (700 × g, 10 min, 4°C). The resulting supernatant was centrifuged again at 14,000 × g for 30 min. The HeLa cell membrane fragments were collected, freeze dried, quantified, and stored at –20°C. These fragments (1 mg) were redispersed in 1× phosphate-buffered saline (PBS), sonicated for 10s, and then extruded through 800- and 400-nm porous polycarbonate membranes at least 11 times using a LiposoFast extruder (Avestin) to obtain HeLa cell membrane vesicles. The HMnO₂@ICG mixture (1 mg/mL, 0.5 mL) was added to these vesicles (2 mg/mL, 0.5 mL), followed by mixing in an ice bath, sonication for 20s, and extrusion through a 400-nm polycarbonate membrane at least 5 times. The mixture was then centrifuged (5000 × g, 30 min, 4°C) and the supernatant was discarded to obtain the final M-HMnO₂@ICG formulation.

Material Characterization

The NP morphology, size, and elemental distribution were observed under a transmission electron microscope (TEM; Tecnai G2 F20; FEI, USA). The materials’ Brunauer, Emmett, and Teller (BET) surface area and pore volume were determined using an ASAP2460 3.01 nitrogen adsorption-desorption isotherm (Micromeritics, USA). The particles’ dynamic light scattering (DLS) and zeta potentials were determined using a Zetasizer Nano ZS90 device (Malvern). X-ray photoelectron spectra collected using an ESCALAB 250Xi microprobe (Thermo Scientific) were used to analyze the NPs’ elemental composition and valence distributions. Inductively coupled plasma optical emission spectrometry (ICP-OES) was used to determine the Mn content. Fluorescence spectra were obtained using an F-2700 spectrometer (Hitachi, Japan). UV/Vis-NIR absorption spectra were measured using a UV-2600 spectrometer (Shimadzu, Japan).

Membrane Protein Characterization

The types and contents of membrane proteins in the HMnO₂@ICG NPs, HeLa cell membranes, and M-HMnO₂@ICG NPs were investigated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Total proteins were extracted from cell membranes using RIPA lysis buffer, analyzed by electrophoresis, and stained with Komas blue. Western blotting was used to identify surface proteins in the HeLa cell membranes. The Hela cells, HeLa cell membranes, and M-HMnO₂@ICG NPs were lysed using RIPA buffer containing PMSF in an ice bath for 30 min and quantified using a BCA kit. These collected proteins were mixed with up-sampling buffer, denatured by boiling in water for 10 min, and subjected to SDS-PAGE (Mini-Protean; Bio-Rad, USA). The gels were then transferred to a polyvinylidene fluoride membrane, sealed with 5% milk in an ice bath for about 1 h, and incubated overnight at 4°C with a monoantibody containing CD47, CD44, E-cadherin, and galactin-3, followed by incubation with the corresponding secondary antibody. Finally, protein signals were observed using a gel imaging system (ChemiDoc XRS; Bio-Rad).

In-Vitro Photothermal Performance Assessment

Aqueous M-HMnO₂@ICG solutions (50, 100, 150, and 200 µg/mL, 0.5 mL) were added to centrifuge tubes and irradiated for 6 min using a 1.35-W/cm² power density laser, with the real-time monitoring of temperature changes using an NIR thermal imager (FLlR Systems AB). Temperatures were recorded by direct photography with an infrared thermal imager at 0, 1, 3, and 5 min after irradiation. In addition, the temperature changes of an M-HMnO₂@ICG aqueous solution (200 µg/mL) irradiated with different laser power densities (0.5, 1.0, and 1.5 W/cm²) were investigated using the same method. To investigate the photothermal stability of M-HMnO₂@ICG, four consecutive heating–cooling cycles were performed. The aqueous M-HMnO₂@ICG solution (150 μg/mL, 1 mL) was irradiated for 10 min (808 nm, 1.35 W/cm²) and then allowed to cool naturally for 11 min. According to the temperature changes that occurred during the cooling stage, its photothermal conversion efficiency (η) was calculated as follows:^44,45

where η is the heat transfer coefficient, A is the surface area of the vessel, T_max is the maximum temperature of the M-HMnO₂@ICG aqueous dispersion, T_sur is the temperature of the surrounding environment, Qs is the power absorbed by the solvent, I is the incident laser power (1.35 W/cm²), and A_λ is the absorbance of the M-HMnO₂@ICG aqueous dispersion at 808 nm.

Assessment of M-HMnO₂@ICG’s O₂ Generation Efficiency

To test the extracellular O₂ production ability of M-HMnO₂@ICG, the fluorescence signal of proportional [Ru (dpp)₃] Cl₂ (RDPP) bursting by O₂ was examined. An RDPP (10 mM, 50 μL) ethanol solution was added to PBS containing M-HMnO₂@ICG (100 μg/mL). RDPP, RDPP + H₂O₂, RDPP + M-HMnO₂@ICG (pH 7.4), RDPP + M-HMnO₂@ICG + H₂O₂ (pH 7.4), RDPP + M-HMnO₂@ICG + H₂O₂ (pH 6.0), and H₂O₂ (100 μM) were added to the corresponding systems. At 0, 3, 6, 9, and 12 min post-reaction, each mixture was transferred to a fluorescent dish for the detection of the fluorescence intensity of RDPP (λ_ex = 455 nm, λ_em = 615 nm).

In-Vitro Photodynamic Property Assessment

To investigate the photodynamic properties of M-HMnO₂@ICG, the ¹O₂ produced in each reaction system under 808 nm laser irradiation was detected using 9,10-anthracenediyl-bis(methylene)dimalonic acid (ABDA). The samples were dispersed in 1 mL PBS containing 200 μM ABDA. Under the conditions of pH 7.4, pH 7.4 + H₂O₂, and pH 6.0 + H₂O₂, 808 nm (1.35 W/cm²) laser irradiation was applied for 2, 4, 6, 8, and 10 min. The supernatants were then collected and changes in the ABDA absorption peaks were detected using UV/Vis-NIR absorption spectroscopy.

Extracellular •OH Detection

To evaluate the responsive degradation and generation of •OH in the tumor microenvironment by the M-HMnO₂@ICG NPs (100 µ g/mL), GSH (0, 1, 5, and 10 mM), H₂O₂ (4 mM), NaHCO₃/5% CO₂ buffer solution (25 mM), and MB (10 µg/mL) were reacted under a constant temperature of 37°C for 30 min. Changes in MB absorbance at 664 nm were monitored to detect the rates of •OH generation under different conditions. Then, M-HMnO₂@ICG (100 µg/mL) was placed in a buffer solution (pH 6.0 or 7.4) containing NaHCO₃/5% CO₂ and H₂O₂ (4 mM) for 30 min, and MB degradation was assessed to determine the production of • OH through an Mn²⁺-mediated Fenton-like reaction.

In vitro Release Experiment of ICG

Disperse M-HMnO₂@ICG (1 mg) in different simulated biological environments such as pH 7.4; pH 6.0; pH 6.0 + 1 mM H₂O₂; pH 6.0 + 10 mM GSH, and incubate at 37°C for various times. Then, centrifuge the solution, collect 1 mL of the supernatant for analysis by UV/Vis-NIR absorption at 800 nm, and replace with fresh buffer.

Cell Culture

HeLa cells and mouse monocyte macrophage leukemia (RAW264.7) cells were purchased from Xiangya Cell Center (Changsha, China). The cells were cultured in modified Eagle medium (MEM) and Roswell Park Memorial Institute 1640 medium, respectively, supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin supplementation under 5% CO₂ at 37°C.

Assessment of in-vitro Targeting Capacity and Macrophage Phagocytic Escape

HeLa or RAW264.7 cells were inoculated in 35-mm confocal dishes (4 × 10⁴ cells/dish) and incubated overnight, followed by the addition of 100 μg/mL HMnO₂@ICG or M-HMnO₂@ICG for 6 h. Before the experiment, the unabsorbed NPs were washed three times with PBS, 4% paraformaldehyde was added, and the mixtures were stored at 4°C for 1 h. The excess paraformaldehyde was aspirated and placed in 5 μg/mL Hoechst dye for 10 min, followed by three washes and observation under a confocal laser microscope (LSM780NLO; Zeiss, Oberkochen, Germany). Fluorescence quantification was performed using a flow cytometer (LSRFortessa; BD, USA) and the method described above.

Cytotoxicity and Inflammation Factor Detection Experiment

The cytotoxicity and therapeutic effects of the nanomaterials were tested by CCK8 assay. HeLa cells were incubated in 96-well plates at 37°C for 24 h. Then, 100 μL fresh medium containing the same concentration gradients of M-HMnO₂@ICG and HMnO₂@ICG was added to each experimental well. After 6 h incubation, the cells were gently washed and incubation was continued for 12 h. Then, 10 μL CCK8 was introduced into each well and incubation was continued for 1 h. Cell viability was determined by measuring UV–Vis absorbance at 450 nm with a microplate reader (Unlimited M200 pro; Tecan, AT). To verify the materials’ therapeutic effects, cells were cultured as described above. They were then co-incubated with M-HMnO₂@ICG under pretreatment conditions for 6 h, followed by laser irradiation (808 nm, 1.35 W/cm², 10 min) and incubation for another 12 h. Finally, cell viability was assessed using the CCK8 method and synergistic treatment effects were evaluated. Use polymerase chain reaction to detect the effects of the photothermal effect of nanomaterials on inflammation in the microenvironment. Seed HeLa cells and RAW 264.7 macrophages separately in a 24-well plate at densities of 4×10⁴ and 1×10⁴ cells per well, respectively, and culture for 24 hours. After that, add different formulations to the HeLa cells and incubate for 6 hours, wash three times with PBS, and then treat the HeLa cells with an 808 nm NIR laser at an intensity of 1.35 W/cm² for 10 minutes. Collect the supernatant from the HeLa cells and use it to culture the RAW 264.7 macrophages for 24 hours. Then, collect the resulting RAW 264.7 cells for total RNA extraction, and further perform RT-PCR detection using a qRT-PCR kit. The primer sequences are designed according to Table S1.

Live-Dead Cell Staining and Assessment of Apoptosis

Cytotoxicity was evaluated in the control, NIR, HMnO₂@ICG, ICG + NIR, HMnO₂@ICG + NIR, and M-HMnO₂@ICG + NIR groups using calcein-acetoxymethyl ester (AM)/propidium iodide (PI) dye. Each well of a 24-well culture plate containing 500 μL medium was inoculated with 4×10⁴ HeLa cells. When the cell confluence had reached about 70%, fresh medium containing different NPs was added and the cells were co-incubated for 6 h. Then, the cells were irradiated (808 nm, 1.35 W/cm², 10 min), followed by incubation for 12 h. They were then stained for 30 min using a calcein-AM/PI double staining kit and photographed under a fluorescence microscope. To quantify apoptosis, cells were cultured and treated as described above, then stained using an annexin V-Alexa Fluor 647/PI apoptosis detection kit and analyzed by flow cytometry.

Intracellular Oxygen, Glutathione, and Reactive Oxygen Species Detection

To verify intracellular O₂ production, HeLa cells were first inoculated in 24-well plates for 24 h, and then co-incubated with 10 μg/mL RDPP for 4 h. After being washed three times with PBS, they were incubated with fresh medium containing 100 μg/mL M-HMnO₂@ICG or HMnO₂@ICG for 8 h. Intracellular O₂ levels were determined by fluorescence based on RDPP signals (λ_ex/λ_em = 488/600–700 nm).

To determine the effect of nanomaterial-generated O₂ on hypoxia-inducible factor-1α (HIF-1α) expression, HeLa cells were inoculated in 6-well plates at a density of 1×10⁵ cells/well with 2 mL MEM for 24 h, then incubated with intact MEM containing M-HMnO₂@ICG or HMnO₂@ICG (100 μg/mL) for 8 h. After two washes with PBS, the cells were lysed and intracellular HIF-1α levels in cells treated with PBS and the two NPs were determined by Western blotting as described in our previous work.⁴⁶

To determine the intracellular GSH content, HeLa cells were inoculated in 6-well plates at a density of 1×10⁵ cells/well and incubated for 24 h. After incubation with PBS and M-HMnO₂@ICG (50 or 100 μg/mL) for 6 h, the medium was removed and the cells were washed three times with PBS. Then, 80 μL Triton-X-100 lysis buffer (0.4%) was added to the lysed cells. The lysate was centrifuged at 6000 rpm for 5 min, and 50 μL supernatant was mixed with 200 μL 5.5′-dithiobis-(2-nitrobenzoic acid) (0.5 mM); Absorbance at 410 nm was measured using a microplate reader.

For the ROS assay, HeLa cells were inoculated in 24-well plates at a density of 4×10⁴ cells/well for 24 h. Then, samples in the control, NIR, HMnO₂@ICG, ICG + NIR, HMnO₂@ICG + NIR, and M-HMnO₂@ICG + NIR treatment groups were co-incubated with the respective NPs for 6 h. The cells were washed with PBS and stained with diacetyldichlorofluorescein (DCFHDA, 10 μM) at 37°C for 30 min, then laser irradiated (808 nm, 1.35 W/cm², 10 min), washed three times with PBS, and observed under a fluorescence microscope. In addition, HeLa cells were inoculated in 24-well plates, and ROS production in the same treatment groups was detected by flow cytometry.

Animal and Tumor Models

Nude BALB/c mice (6–8 weeks old, 18–20 g) were obtained from Hunan Silaike Experimental Animal Co. All animal experiments were approved by the Laboratory Animal Ethics Committee of Central South University (no. CSU-2023-0227) and were conducted in accordance with the Law of the People’s Republic of China on the Use of Laboratory Animals. HeLa cells were collected and dispersed in PBS (3 × 10⁷/mL), then injected subcutaneously into the right axillae of the mice to establish a tumor-bearing model. When the tumor volume had reached 100 mm³, the mice were used in the experiments.

Hemolysis Assay

Blood samples were obtained from normal nude BALB/c mice and centrifuged (3000 rpm, 4°C). The obtained erythrocytes were washed several times until the supernatant was clear, and a 2% (v/v) erythrocyte suspension was prepared with PBS. This suspension (200 µL) was added to a series of equal volumes of M-HMnO₂@ICG at concentrations of 12.5, 25, 50, 100, and 200 µg/mL and co-incubated at 37°C for 2 h. It was then centrifuged at 10,000 rpm, and 100 µL supernatant was transferred to each well of a 96-well plate. Absorbance at 540 nm was measured and the hemolysis rate was calculated using the optical density (OD) and the following formula: hemolysis rate (%) = OD_sample – OD_{negative control} / OD_{positive control} – OD_{negative control} × 100.

In-Vivo and Ex-Vivo Fluorescence Imaging

Tumor-bearing mice were injected in the tail vein with 100 µL HMnO₂@ICG and M-HMnO₂@ICG at the same concentration (400 µg/mL). At 0, 12, 24, 48, and 72 h post-injection, the mice were anesthetized with gas and in-vivo fluorescence imaging was performed using a small-animal imaging system (Ani View 600; λ_ex/λ_em = 745/840 nm). For ex-vivo fluorescence imaging, all mice were executed 72 h after injection, and the tumors and major organs (heart, liver, spleen, lungs, and kidneys) were dissected, washed with PBS, and examined using the in-vivo imaging system.

In-Vivo and in-vitro Magnetic Resonance Imaging

The in-vitro T1-weighted imaging properties of M-HMnO₂@ICG were examined using a 7.0-T MRI scanner. Samples containing different concentrations (0.1725, 0.345, 0.69, and 1.38 mM) of Mn²⁺ were dispersed in PBS (pH 6.0 or 7.4), and GSH (10 mM) or H₂O₂ (1 mM) was added. T1 relaxation rates were obtained by detecting the T1 signal intensity in each group. For in-vivo MRI examination, M-HMnO₂@ICG (20 mg/kg) was injected (intravenous or intratumor) into HeLa tumor (~100 mm³)–bearing mice. For the intratumor injection model, MRI was performed on mice that had not been injected and on treated mice at 30 min and 1 h after injection. For the intravenous injection model, MRI was performed in the tumor region on mice that had not been injected and on treated mice 24 h after injection.

Biodistribution Measurements

Hela tumor-bearing mice were divided into three groups (n=3), and dissected at 12, 24, and 48 hours post-injection to collect tumor tissues. Subsequently, all tumor tissues were treated with concentrated nitric acid and heated until the solution became clear. The manganese content in the solution was determined using an Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES), and the manganese content in the collected mouse tumor tissues was calculated.

In-Vivo Photothermal Imaging

HMnO₂@ICG or M-HMnO₂@ICG (Mn²⁺ concentration, 20 mg/kg) was injected into tumor-bearing mice. Twenty-four hours after injection, the mice were anesthetized for tumor site irradiation (808 nm, 1.35 W/cm², 3min) and infrared thermal imaging (ONE PRO; FLIR). Tumor site warming was monitored.

In-Vivo Examination of Antitumor Treatment Effects

HeLa tumor (~100 mm³)–bearing nude BALB-c mice were divided randomly into six treatment groups (n = 3 each): PBS, NIR, ICG + NIR, HMnO₂@ICG, HMnO₂@ICG + NIR, and M-HMnO₂@ICG + NIR. The nanomedicines were injected through the tail vein on days 0 and 5 of the treatment cycle, according to the previously determined time window of maximum drug accumulation. Twenty-four hours after injection, all animals were anesthetized and irradiated (1.35 W/cm², 3 min). The animals’ weight and tumor volume were monitored and recorded every other day during the treatment period. After 14 days, the mice were euthanized and representative tumor tissue samples were collected, photographed, and weighed. The tumor tissues and major organs of mice from the different treatment groups were stained with hematoxylin and eosin (H&E), and mouse tumor tissues were stained by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) to detect apoptosis. To detect HeLa cell hypoxia, tumor tissues from each treatment group were stained with anti–HIF-1α antibody, washed with PBS to remove the primary antibody residue, and then incubated with goat anti-rabbit immunoglobulin G at 37°C for 1 h. Fluorescence was then observed, with green fluorescence representing hypoxic areas.

Biosafety Analysis

To examine the biosafety of M-HMnO₂@ICG, mice that had undergone 14 days of the different treatments were euthanized, blood samples were collected, and serum specimens were obtained by centrifugation. Biochemical assays were performed to determine the aspartate aminotransferase, alanine aminotransferase, total protein, albumin, blood urea nitrogen, creatinine, total cholesterol, and triglyceride levels.

Statistical Analysis

The data are expressed as means ± standard deviations. GraphPad Prism 8 was used for the statistical calculations and plotting. One-way and two-way analyses of variance were used to compare data from the various treatment groups. P < 0.05 was considered to indicate significance.

Results and Discussion

M-HMnO₂@ICG NP Preparation and Characterization

Firstly, an oxidation-reduction reaction occurs between KMnO₄ and SiO₂, resulting in the formation of a MnO₂ layer on the surface of well-dispersed SiO₂NPs (Figure 1A), thus creating SiO₂-MnO₂ NPs (Figure 1B). Then, utilizing the strong alkaline properties of Na₂CO₃ as an etching agent, the internal silicon shell is dissolved to prepare HMnO₂ NPs. TEM observation confirmed that all prepared NPs had regular spherical structures and sizes of about 110 nm. The HMnO₂ NPs had typical hollow structures (Figure 1C). Elemental scanning analysis revealed significantly reduced silicon elemental signals in the HMnO₂ structure relative to SiO₂-MnO₂ (Figure S1), confirming the successful etching of the silicon shells. Significant Mn elemental signals were observed in the HMnO₂ structure (Figure S2), confirming the synthesis of the MnO₂ shells. BET calculations yielded a specific surface area of 178 m²/g and average pore size of 4 nm for HMnO₂ (Figure 1D), conforming to a typical type-IV Langmuir isotherm and indicating the presence of a large hollow structure and abundant mesopores to facilitate drug adsorption and loading.

Figure 1 (A–C) Transmission electron microscopic images of SiO2, SiO2@MnO2, and HMnO2 nanoparticles. (D) N2 adsorption/desorption isotherms and pore size distribution curve (inset) for synthesized HMnO2. (E) Transmission electron microscopic image of HMnO2@ICG. (F) Elemental mapping images and energy-dispersive spectra of Mn, O, N, P, and S in HMnO2@ICG. Scale bar: 50 nm. (G) Supernatant UV/Vis-NIR absorption after ICG loading of HMnO2@ICG under different addition conditions. (H) Drug loading and encapsulation rates of ICG-loaded NPs under different conditions. (I) Transmission electron microscopic image of M-HMnO2@ICG. (J) Hydrodynamic dimensions of HMnO2@ICG and M-HMnO2@ICG, determined by DLS measurement. (K) Surface charges of HMnO2, HMnO2/PAH, HMnO2@ICG, and M-HMnO2@ICG. Data are expressed as means ± standard deviations (n = 3). (L) Full XPS characteristics of M-HMnO2@ICG. (M) UV/Vis-NIR absorption spectra of HMnO2, ICG, HMnO2@ICG, and M-HMnO2@ICG. (N) SDS-PAGE results for HeLa cell membranes, M-HMnO2@ICG, and HMnO2@ICG. (O) Western blots of HeLa cells, their membranes, and M-HMnO2@ICG.

After HMnO₂@ICG synthesis, TEM observation revealed blurring of the hollow NP structure and cluster-like aggregation around the core-shell structure (Figure 1E). Elemental mapping images and corresponding X-ray energy-dispersive spectra confirmed the existence and spatial distribution of Mn, O, N, P, and S in HMnO₂@ICG (Figure 1F), thereby confirming phosphate-mediated ICG encapsulation.

Under consistent reaction conditions, we investigated the effect of different carrier modifications on the drug loading capacity and encapsulation efficiency of ICG. We observed that, compared to the samples with HMnO₂ and HMnO₂/PAH as carriers, the group treated with HMnO₂/PAH/HPO₄^2- as the carrier exhibited the lowest ultraviolet/visible-near infrared in the supernatant (Figure 1G). The UV/Vis-NIR absorption of the supernatant indicated the presence of unencapsulated (free) ICG. Quantitative analysis was conducted by examining the linear relationship between the characteristic peak and concentration of ICG dissolved in PB (Figure S3). We found that these rates were significantly higher for HMnO₂/PAH/HPO₄^2- (31.72% and 95.14%, respectively) than for HMnO₂ (5.43% and 24.16%, respectively) and HMnO₂/PAH (17.21% and 68.58%, respectively; Figure 1H), attributable mainly to the high ICG affinity of the counterion aggregation effect formed by phosphate/PAH.

TEM observation showed that the encapsulated HMnO₂@ICG NP surfaces had a clear lipid layer with a thickness of about 8.6 nm (Figure 1I). DLS measurement showed that the NP size had increased from 240 to 258 nm after cell membrane encapsulation (Figure 1J). The synthesis of M-HMnO₂@ICG was verified by zeta potential monitoring. The potential of HMnO₂@ICG changed from +21 to –10 mV, suggesting the occurrence of a flip relative to the HMnO₂/PAH potential that was attributed to successful ICG loading. The M-HMnO₂@ICG NP charge was further reduced to –29 mV relative to that of HMnO₂@ICG after cell membrane encapsulation, which may be attributed to the cell membrane surface charge (Figure 1K).

X-ray photoelectron spectroscopy (XPS) detected absorption peaks for Mn, O, N, P, and S in M-HMnO₂@ICG (Figure 1L). The high-resolution XPS spectrum of Mn2p showed peaks at 642.3 and 653.3 eV (Figure S4) corresponding to the bonding energies of Mn2p_3/2 and Mn2p_1/2 of MnO₂, respectively, the satellite peak of manganese (Mnsat)is located approximately 3 eV higher in energy than the main peak, indicating the presence of manganese in a higher oxidation state.⁴⁷

The high-resolution XPS spectrum of S2p showed 2p_1/2 and 2p_3/2 peaks corresponding to the bonding energy of the sulfonic acid moiety⁴⁸ (-SO₃H, ~169 eV; Figure S5), with the N and P peaks attributed to the PAH/phospholipid bilayer-coated cell membranes and the presence of HPO₄^2-, respectively (Figure S6 and S7). UV/Vis-NIR absorption spectral peaks for the M-HMnO₂@ICG and HMnO₂@ICG NPs were consistent with those for the HMnO₂ NPs at 300–500 nm and ICG at 700–850 nm (Figure 1M). In particular, the red-shifted ICG peaks of HMnO₂@ICG were ascribed to the conversion of ICG monomers to J aggregates via non-covalent chemistry.⁴⁹ The characteristic ICG absorption peaks of M-HMnO₂@ICG were consistent with those of HMnO₂@ICG, indicating that the ICG-loaded NPs maintain functional integrity after physical coextrusion and cell membrane encapsulation.

The integrity of surface proteins after HMnO₂@ICG NP encapsulation was examined further. The protein bands of the M-HMnO₂@ICG NPs were almost the same as those of the HeLa cell membranes; no protein band was detected for HMnO₂@ICG (Figure 1N). Several marker proteins were detected on the HeLa cell surfaces, including isotype-targeting highly dependent cell adhesion molecules (eg, epithelial cell adhesion molecule, galectin-3, CD44, and CD47). CD44, a signature tumor-cell transmembrane glycoprotein, is involved in heterogeneous tumor cell adhesion.⁵⁰ CD47, a transmembrane glycoprotein that is expressed and up-regulated widely in tumor cells, binds to signal-regulatory protein-α on the surfaces of macrophages and signals to inhibit macrophage phagocytosis, thereby playing an important role in immune escape.^51,52 These proteins were preserved on the surfaces of the HeLa cells, their membranes, and the M-HMnO₂@ICG NPs, indicating that NP preparation by co-extrusion did not disrupt their presence (Figure 1O). Western blotting confirmed that these homologous targeting- and immune tolerance–associated proteins had been transferred from the cell membranes to the NP surfaces, providing the basis for the NPs’ action.

M-HMnO₂@ICG Photothermal, Photodynamic, and Chemical Kinetic Properties, and Degradation Properties Reflecting Tumor Microenvironment Responsiveness

As the ICG loaded on M-HMnO₂@ICG showed significant absorption in the 700–900-nm NIR spectral region, indicating that this NP had the property of a photothermal reagent, we systematically investigated its in-vitro photothermal properties. The photothermal conversion temperatures of M-HMnO₂@ICG solutions with different concentrations under 6 min NIR irradiation showed obvious concentration dependence; a temperature of 25°C was achieved with a concentration of 200 μg/mL, relative to 3°C for water (control; Figure 2A). The relationship between warming and laser power was also investigated. The temperature of M-HMnO₂@ICG (150 μg/mL) under 6 min irradiation increased with the laser power density (Figure 2B), indicating that this NP effectively converted 808 nm laser energy into thermal energy. We further investigated the photothermal stability and conversion efficiency of M-HMnO₂@ICG. In the on/off cycle irradiation (808 nm) experiment, M-HMnO₂@ICG showed no obvious temperature attenuation after four cycles, indicating good photothermal stability (Figure 2C). Photothermal conversion efficiency is the efficiency with which a photothermal agent converts absorbed energy from near-infrared light into thermal energy, and is a key parameter for evaluating the performance of photothermal agents. The photothermal conversion efficiency of the M-HMnO₂@ICG solution, calculated by fitting the curve of temperature change over time during cooling, was 34.62% (Figure 2D). An infrared thermal imager was also used to obtain photothermal images of M-HMnO₂@ICG solutions with different concentrations. The solution color changed from light-yellow to light-red with increasing concentration and irradiation time (Figure 2E), confirming the preparation’s photothermal conversion efficiency.

Figure 2 (A) In-vitro photothermal properties of M-HMnO2@ICG nanoparticles; temperature curves for different concentrations of M-HMnO2@ICG under 808-nm laser irradiation for 6 min. (B) Temperature curves for M-HMnO2@ICG solution (150 μg/mL) under 808-nm laser irradiation at different power densities. (C) Photothermal conversion of M-HMnO2@ICG solution (150 μg /mL) over four on/off irradiation (808 nm) cycles. (D) Photothermal properties of aqueous M-HMnO2@ICG solution and their linear relationship with time; temperature curve for M-HMnO2@ICG (150 μg/mL) under laser irradiation (808 nm, 1.35 W/cm2, 600 s; inset). (E) Infrared thermography of aqueous M-HMnO2@ICG solutions under 808-nm (1.35 W/cm2) laser irradiation for 0–5 min. (F) Fluorescence (RDPP probe) under different conditions reflecting M-HMnO2@ICG O2 generation and (G) O2 production over time. (H) Variation in ABDA absorbance of M-HMnO2@ICG nanoparticles in pH 6.0 + H2O2 solution with different laser irradiation intervals. (I) Normalized ABDA absorbance under different conditions. UV–Vis absorption spectra of supernatants after (J) warming in different pH buffer solutions and (K) incubation with different GSH concentrations. (L) MB degradation and photograph of the same samples after different temperature treatments (inset). (M) Nanoparticle degradation under different conditions, observed by transmission electron microscopy.

Next, We used RDPP as a probe to investigate the O₂ production capacity of M-HMnO₂@ICG NPs. As the fluorescence of RDPP is quenched upon reaction with O₂, decreases in its fluorescence signal can be used as O₂ level markers. Among treatment groups, RDPP + M-HMnO₂@ICG + H₂O₂ (100 μM) showed the most pronounced decrease in fluorescence in the reaction system with buffer solution of pH 6.0, demonstrating the NP’s ability to produce O₂ (Figure 2F). We further investigated reaction kinetics and observed no significant fluorescence quenching within 12 min for RDPP, RDPP + H₂O₂, or RDPP + M-HMnO₂@ICG (pH 7.4; Figure 2G). At pH 7.4, the fluorescence quenching of RDPP + M-HMnO₂@ICG + H₂O₂ was slow relative to that at pH 6.0, consistent with the postulated reaction of MnO₂ with H₂O₂ to generate more O₂ in an acidic environment. These findings indicate that the NPs self-produced O₂ in the in vitro-simulated tumor microenvironment, providing for ICG-mediated PDT. ICG produces ROS (¹O₂) under photodynamic action. We used ABDA as a probe to examine the photodynamic performance of ICG. ¹O₂ can oxidize ABDA to peroxides, reducing its UV absorption at 380 nm. At pH 7.4, the typical ABDA absorption peak in the M-HMnO₂@ICG solution remained almost unchanged with 10 min laser irradiation (Figure S8), indicating that the NP had weak photodynamic properties. With the addition of H₂O₂ to the solution, the ABDA absorption peak decreased more (Figure S9) because the small amount of O₂ generated by catalysis under this condition enhanced the photodynamic efficiency of ICG. At pH 6.0, the ABDA absorption peak of M-HMnO₂@ICG decreased significantly with increasing irradiation time (Figure 2H). UV absorption was measured at 380 nm for relative quantification and the more intuitive characterization of the photodynamic properties of ICG under different conditions (Figure 2I). These results indicate that M-HMnO₂@ICG produced O₂ through CAT-like activity in the acidic, H₂O₂-rich tumor microenvironment, providing substrates for ICG and enhancing its photodynamic effects.

In addition to increasing the photodynamic effect through catalytic O₂ production, MnO₂ can be degraded to generate Mn²⁺. MB was used as an indicator for ·OH detection to study the chemical kinetic properties of M-HMnO₂@ICG. H₂O₂ and M-HMnO₂@ICG were added to NaHCO₃ buffer containing MB and incubated for 30 min, and the MB absorbance in the supernatant was then detected. The fluorescence signal was significantly weaker with pH 6.0 buffer solution than with pH 7.4 buffer solution (Figure 2J), indicating that the chemical kinetic effect was stronger in an acidic simulated tumor microenvironment. Absorbance did not differ significantly between the MB alone (control) and MB + H₂O₂ + HCO₃⁻ groups, indicating that no ·OH was generated without the addition of nanomaterials. As MnO₂ can degrade rapidly, producing Mn²⁺, under GSH-rich conditions in cells, we investigated the effect of nanomaterials and GSH reactions on MB degradation. The UV-Vis absorption of MB decreased gradually, reflecting more Mn²⁺ production and thus a stronger chemical kinetic effect, with increasing GSH concentration (Figure 2K). The rates of Fenton-like reactions and ·OH generation increased with the temperature (Figure 2L), indicating that the photothermal effect of ICG has a synergistic effect on chemical kinetics.

To visually assess M-HMnO₂@ICG degradation in the simulated tumor microenvironment, we incubated the NP in buffer solutions of pH 7.4, pH 6.0, pH 6.0 + H₂O₂, and pH 6.0 + GSH for 2 h and then observed them under a TEM. The structure of M-HMnO₂@ICG remained relatively intact in the pH 7.4 buffer solution (Figure 2M), demonstrating stability during in-vivo cycling. In the slightly acidic (pH 6.0) environment, the NPs tended to collapse, losing their original morphology. After the addition of H₂O₂ or GSH, the NP morphology changed significantly, with degradation into smaller nanoclusters. Additionally, under conditions of pH 6, pH 6.0 + H₂O₂, and pH 6.0 + GSH, the cumulative release of ICG is significantly greater than under conditions of pH 7.4 (Figure S10). These results indicate that the designed nanomaterials have the property of tumor microenvironment–responsive release, enabling effective synergistic PTT/PDT/CDT.

In-Vitro Tumor Targeting and Immune Escape Properties of M-HMnO₂@ICG

In this study, NPs were coated with tumor cell membranes with the aim of enabling them to evade macrophage-mediated nonspecific clearance, prolong the circulation time in the bloodstream, enhance interaction with tumor cells, and achieve homologous targeted delivery to tumor cells. To verify the effectiveness of this approach, we used RAW cells as phagocytes and HeLa cells as targets to study phagocytic escape and homologous uptake of M-HMnO₂@ICG NPs by confocal laser scanning microscopy. In examining the efficacy of cell membrane encapsulation, we used HMnO₂@ICG without encapsulation as a control.

We used the spontaneous fluorescence of ICG to trace the NPs and 4’,6-diamidino-2-phenylindole staining of nuclei to localize cells. HeLa cells treated with M-HMnO₂@ICG showed significantly more red fluorescence than did those incubated with HMnO₂@ICG (Figure 3A), confirming the good targeting of M-HMnO₂@ICG to homologous tumor cells. Mouse macrophages (RAW264.7 cells) co-incubated with HMnO₂@ICG showed weak red fluorescence and those co-incubated with M-HMnO₂@ICG showed almost none, indicating that the designed NPs have a certain immune escape effect (Figure 3B). HeLa and RAW cells were co-incubated with HMnO₂@ICG and M-HMnO₂@ICG, respectively, and nanomaterial uptake was examined using flow cytometry. The fluorescence intensity of HeLa cells co-cultured with M-HMnO₂@ICG was 26.9% greater than that of those co-cultured with HMnO₂@ICG (Figure 3C). The fluorescence intensity of RAW264.7 cells co-cultured with M-HMnO₂@ICG was only 52.6% of that of cells co-cultured with HMnO₂@ICG (Figure 3D). The fluorescence imaging and flow cytometry results (Figure 3E)were consistent and indicate that the proteins on the tumor cell membrane surfaces prevented rapid in-vivo NP clearance through immune escape, enhancing uptake in tumor cells through homologous targeting and thereby achieving enhanced antitumor efficacy.

Figure 3 (A and B) Confocal laser scanning microscopic images of HeLa and RAW264.7 cells treated with HMnO2@ICG and M-HMnO2@ICG nanoparticles for 6 h (scale bar = 20 μm). (C and D) Flow cytometry images of the relative fluorescence of the same treated cells. Relative fluorescence intensity and (E) fluorescence quantification for the same treated cells. Data are expressed as mean ± SD (n = 3). Statistical significances were calculated via one-way analysis of variance (ANOVA).*P < 0.05,**P < 0.01, ***P < 0.001.

In-Vitro Antitumor Properties

The test tube–level experiments showed that the NPs prepared in this study had synergistic PTT/PDT/CDT properties. The effect of M-HMnO₂@ICG NPs on HeLa cell viability was detected by CCK8 assay to evaluate the in-vitro antitumor activity of the NPs. In the absence of light, M-HMnO₂@ICG and HMnO₂@ICG showed no significant cytotoxicity at a concentration of 100 µg/mL, confirming their biosafety. At a concentration of 200 µg/mL, a significant cell-killing effect was observed, attributable mainly to the NPs’ chemical kinetic effects. The cell-killing effect of M-HMnO₂@ICG was significantly stronger than that of HMnO₂@ICG, reflecting the former’s stronger tumor cell targeting (Figure 4A). Under laser irradiation (808 nm, 1.35 W/cm², 10 min), M-HMnO₂@ICG had a concentration-dependent antitumor effect, suggesting a combined therapeutic effect. To further distinguish PTT, PDT, and CDT contributions, we pretreated the cells with an ice bath or vitamin C. The untreated group without laser irradiation represented the CDT effect, and the difference in cell survival between the ice bath–treated and non–laser-irradiated groups was considered to reflect the PDT contribution. The vitamin C treatment was considered to contribute to the PTT effect due to the elimination of photodynamic and chemodynamic effects. The ice bath and vitamin C pretreatments increased cell survival, confirming the synergistic contributions of photothermal, photodynamic, and chemokinetic effects to the antitumor activity of the NPs (Figure 4B). The half-maximal inhibitory concentrations (IC₅₀s) of M-HMnO₂@ICG (chemodynamic contribution), M-HMnO₂@ICG + ice bath (photodynamic and chemodynamic contributions), and M-HMnO₂@ICG + vitamin C (photothermal contribution) were 273.4, 104.9, and 59.0 µg/mL, respectively (Figure 4C). The IC₅₀ of M-HMnO₂@ICG + NIR was the lowest (34.7 µg/mL), reflecting the combined therapeutic effect of M-HMnO₂@ICG with homologous targeting to promote PTT/PDT/CDT under light conditions. Considering that photothermal therapy raises the temperature of local tumor tissues and potentially triggers inflammatory responses, to assess the impact of photothermal therapy on inflammation within the tumor microenvironment, we examined the effects of treated HeLa cells on inflammatory responses in RAW 264.7 macrophages. Using real-time polymerase chain reaction, we measured the levels of inflammatory factors released by treated cells. As depicted in Figure S11, under near-infrared light conditions, both HMnO₂@ICG and M-HMnO₂@ICG did not significantly upregulate the mRNA expression of pro-inflammatory factors (TNF-α, IL-1β, and IL-6), in contrast to the ICG+NIR group, which exhibited increased expression of IL-1β and IL-6. This finding underscores the advantage of using the nanomaterial M-HMnO₂@ICG over solely using ICG in the tumor microenvironment under near-infrared light exposure.

Figure 4 (A) Effects of different concentrations of M-HMnO2@ICG and HMnO2@ICG on HeLa cell viability. (B) HeLa cell viability after treatment with M- HMnO2@ICG at different concentrations and under different conditions. (C) IC50s, calculated using the data illustrated in panel B. (D) Live/dead HeLa cell staining after different treatments (scale bar: 50 μm). Apoptosis as determined by (E) flow cytometry and (F) quantitative analysis after different treatments. Data are expressed as mean ± SD (n = 3). Statistical significances were calculated via one-way analysis of variance (ANOVA). *P < 0.05, **P < 0.01, ****P < 0.0001.

The NPs’ antitumor activity was visualized by live/dead cell staining. After NIR irradiation, M-HMnO₂@ICG caused a large number of cell deaths (red fluorescent labeling), further confirming its combined therapeutic and targeting effects (Figure 4D). To further explore the tumor cell death pathway, we used flow cytometry and an annexin V-APC/PI apoptosis detection kit. As expected, M-HMnO₂@ICG + NIR triggered the strongest cell-autonomous programmed death, with an apoptosis rate of 59.6%, exceeding that observed for HMnO₂@ICG + NIR without tumor cell membrane encapsulation (Figure 4E and F).

Cellular-Level Validation of the Antitumor Potentiation Mechanism of M-HMnO₂@ICG

At the test tube level, we demonstrated that MnO₂ enhanced the effect of ICG via self-oxygenation and chemical kinetics. We further validated this effect at the cellular level, using RDPP to examine O₂ production was significantly reduced in HMnO₂@ICG- and M-HMnO₂@ICG–treated cells incubated for the same period of time compared with the control group (Figure 5A), suggesting that MnO₂ NPs efficiently produce O₂ in tumor cells by catalyzing endogenous H₂O₂. The O₂ production effect was more pronounced in the M-HMnO₂@ICG group than in the HMnO2@ICG group, which may be attributed to the former’s superior cellular uptake effect. HIF-1α is a marker of the intracellular O₂ level; its level correlates negatively with the intracellular O₂ concentration. The alteration of the intracellular O₂ level was further clarified by measuring HIF-1α expression. The intracellular HIF-1α level was significantly decreased after treatment with M-HMnO₂@ICG and HMnO₂@ICG (Figure 5B), further confirming the NPs’ ability to alleviate tumor hypoxia via self-oxygenation.

Figure 5 (A) Fluorescence images of RDPP staining of HeLa cells under different treatments (scale bar: 200 μm). (B) HIF-1α expression in Hela cells under different treatments and quantitative analysis. (C) Relative GSH levels in HeLa cells after treatment with different concentrations of M-HMnO2@ICG. Fluorescence images of DCFHDA staining of HeLa cells after different treatments (D) and corresponding flow cytometry assays (E) and relative fluorescence intensity (F). The data represent mean ± SD, one-way ANOVA, *P < 0.05, ****P < 0.0001.

High GSH concentrations in tumor cells protect against oxidative stress damage, thereby reducing the tumor cell–killing effect of ROS. MnO₂ has been shown to react with GSH, reducing its intracellular level and enhancing the therapeutic effect of oxidative stress.⁵³ To demonstrate this effect, we incubated HeLa cells with different concentrations of M-HMnO₂@ICG and measured the intracellular GSH level by sulfhydryl tracking assay. M-HMnO₂@ICG reduced the intracellular GSH concentration in a concentration-dependent manner (Figure 5C). Thus, it weakens tumor cells’ antioxidant defense systems, facilitating their killing through ROS generation.

We used a DCFHDA probe to detect the ROS production in HeLa cells by different NPs. The green DCFHDA fluorescence in the cells was very weak under the control and irradiation-only conditions (Figure 5D). Under laser irradiation, free ICG induced a weak fluorescent signal in HeLa cells. For HMnO₂@ICG in the absence of light, some fluorescent signal was generated due to the occurrence of Mn²⁺-mediated Fenton-like reactions. Green fluorescence was strongest in the irradiated M-HMnO₂@ICG group, reflecting effective cellular internalization and intracellular self-oxygenation for sensitized PDT. The analysis of green fluorescence by flow cytometry (Figure 5E) and quantification (Figure 5F) produced similar results. Together, these results indicate that M-HMnO₂@ICG effectively targeted tumor cells and produced O₂ intracellularly for PDT sensitization, achieving PTT and Mn²⁺-mediated Fenton-like effects.

In-Vivo Tumor Targeting Properties of M-HMnO₂@ICG

After clarifying the antitumor activity of M-HMnO₂@ICG at the cellular level, the performance of M-HMnO₂@ICG was further investigated at the animal level. M-HMnO₂@ICG showed <5% hemolysis, even at a concentration of 200 μg/mL (Figure S12), indicating its good biocompatibility with tail vein injection. A female nude BALB/c mouse tumor model was constructed, and the tumor-targeting properties of the nanopreparations were first evaluated using ICG fluorescence. Using an in-vivo small-animal imager, dynamic fluorescence changes in the mice after tail vein injection were observed (Figure 6A). Twelve hours after injection, fluorescence at tumor sites was detected in the M-HMnO₂@ICG and HMnO₂@ICG groups. The tumor-site fluorescence intensity peaked at 24 h and then decreased gradually, indicating that the optimal time for laser irradiation is 24 h after injection. The tumor-tissue fluorescence intensity was greater in the M-HMnO₂@ICG group than in the HMnO₂@ICG group at all timepoints, and some fluorescence was detected at 72 h in the former group, suggesting stronger tumor targeting and retention of M-HMnO₂@ICG. The mice were executed and their major organs and tumor tissues were collected for ex-vivo imaging, which confirmed that tumor-tissue fluorescence was stronger in the M-HMnO₂@ICG group (Figure 6B). Quantification confirmed that fluorescence was significantly more intense in the M-HMnO₂@ICG group than in the HMnO₂@ICG at different timepoints (Figure 6C). Twenty-four hours after administration, tumor accumulation in the M-HMnO₂@ICG group was 1.1 times that in the HMnO₂@ICG group. The targeting ability of M-HMnO₂@ICG is attributable to its HeLa cell membrane encapsulation, which prevented rapid clearance during circulation and enabled preferential accumulation at tumor sites through homologous targeting. To verify these results, we used ICP-OES to measure Mn concentrations in tumor tissues at different timepoints after M-HMnO₂@ICG injection. The highest level was measured at 24 hours after injection (Figure 6D).

Figure 6 (A) In-vivo fluorescence imaging of HeLa tumor–bearing mice at different timepoints after the intravenous injection of HMnO2@ICG and M-HMnO2@ICG, with the subcutaneous tumor area indicated by a white dashed line. (B) Ex-vivo imaging of tumors and major organs collected from these mice at 72 h after injection. (C) Quantification of fluorescence at tumor sites after injection. (D) Mn content in tumor tissues at different timepoints after the intravenous injection of M-HMnO2@ICG. (E) T1-weighted MR images of M-HMnO2@ICG dissolved under different pH and redox conditions. (F) Longitudinal T1 relaxation rates versus Mn2+ concentrations under different conditions. (G) In-vivo MR images of HeLa tumor-bearing mice at different timepoints after the intratumoral injection of M-HMnO2@ICG, with the subcutaneous tumor area indicated by a red dashed line. (H) In-vivo T1-weighted MR images were obtained 24 h after the tail vein injection of M-HMnO2@ICG, with the subcutaneous tumor area indicated by a red dashed line. (I) Photothermograms of HeLa tumor-bearing mice in different treatment groups under laser irradiation (808 nm, 1.35 W/cm2).

The MnO₂ in the NP structure degrades responsively in the tumor microenvironment to produce Mn²⁺, which is considered to be a T1 developer for MRI due to its occupation of five unpaired 3d electrons. Thus, a 7.0-T small-animal MRI system was used to evaluate the tumor-targeting and responsive degradation abilities of M-HMnO₂@ICG in vitro. Weak MRI signals were observed for M-HMnO₂@ICG in a pH 7.4 solution, whereas the concentration-dependent enhancement of MRI signals was observed for M-HMnO₂@ICG in a pH 6.0 solution containing GSH or H₂O₂ (Figure 6E), confirming the NPs’ responsive degradation in the tumor microenvironment. Longitudinal relaxivity coefficients for M-HMnO₂@ICG in solutions at pH 7.4, pH 6.0 containing H₂O₂ and pH 6.0 containing GSH were 0.26, 0.59, and 1.2 mM⁻¹·s⁻¹, respectively (Figure 6F). These results suggest that M-HMnO₂@ICG releases Mn²⁺ in response to the tumor microenvironment, which enables tumor diagnosis by MRI. To verify this effect in vivo, MRI signals were recorded at 0, 30, and 60 min after the intratumoral injection of M-HMnO₂@ICG. Clear MRI signals were observed at 30 min after injection, and their intensity increased gradually over time (Figure 6G). Subsequently, we evaluated MRI effects at tumor sites in vivo after tail vein injection of M-HMnO₂@ICG. No significant difference was observed between the tumor tissues before M-HMnO₂@ICG injection and other tissues. The T1-weighted MRI signals of the tumor tissues were significantly enhanced at 24 h after injection (Figure 6H), further indicating that NPs effectively accumulated and released Mn²⁺ at the tumor sites. Taking advantage of the photothermal conversion ability of M-HMnO₂@ICG, we also monitored the NPs’ tumor-targeting enrichment and photothermal properties in vivo using NIR thermography. PBS, HMnO₂@ICG, and M-HMnO₂@ICG were injected via the tail vein into tumor-bearing mice, and laser irradiation (808 nm, 1.35 W/cm²) was performed for 3 min 24 h later. Compared with PBS (control), M-HMnO₂@ICG and HMnO₂@ICG significantly increased the tumor tissue temperature, indicating that the NPs had good photothermal conversion abilities in vivo. The photothermal conversion effect of M-HMnO₂@ICG was stronger than that of HMnO₂@ICG (51.3°C vs 46.3°C, Figure 6I) due to the former’s stronger tumor-targeting ability.

In-Vivo Antitumor Treatment and Biosafety

The antitumor effects of M-HMnO₂@ICG were evaluated in vivo. When the subcutaneous graft tumor volume had reached 100 mm³, the tumor-bearing mice were divided randomly into six groups (PBS, NIR, ICG + NIR, HMnO₂@ICG, HMnO₂@ICG + NIR, and M-HMnO₂@ICG + NIR). The treatments were administered by tail vein injection on days 0 and 5, and 3 min laser irradiation of the tumor regions was performed at 24 h after injection (Figure 7A). The body weights of the mice and the tumor volumes were monitored at 2-day intervals for 14 days after the first injection. The body weights of the mice did not change significantly over time in any group, indicating that no treatment had an acute toxic effect (Figure 7B). The tumor volumes increased rapidly over the 14 days in the control and NIR groups (Figure 7C), indicating that cervical cancer is extremely proliferative and that laser irradiation alone has no therapeutic effect on it. The therapeutic effect was limited in the HMnO₂@ICG-alone group, indicating that CDT alone is not effective. Similarly, poor tumor inhibition was observed in the ICG + NIR group, indicating that PTT and PDT via ICG alone is not sufficient to inhibit tumor growth under hypoxic conditions. A significant tumor inhibitory effect was observed in the M-HMnO₂@ICG + NIR group relative to the other groups, due to homologous targeting to enhance tumor accumulation and PDT/CDT/PTT synergism. The mice were executed after treatment, and the tumor tissues were collected for photography (Figure 7D) and weighing (Figure 7E). The results were consistent with the in-vivo observations, indicating that M-HMnO₂@ICG + NIR had the strongest therapeutic effect.

Figure 7 (A) Schematic diagram of the treatment of tumor-bearing mice. Mouse body weights (B) and relative tumor volumes (C). (D) Photographs of tumors collected from the mice on day 14 after intravenous injection. (E) Tumor tissue weights. Data are expressed as mean ± SD (n = 3). **P < 0.01, ***P < 0.001. (F–H) H&E, and TUNEL staining and immunofluorescence images of HIF-1α (green; blue, nuclei) in tumor tissues collected after treatment (scale bar: 50 μm). (I–L) Post-treatment alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), creatinine (CRE), total cholesterol (TC), triglyceride (TG), albumin (ALB), and total protein (TP) levels in HeLa tumor-bearing mice. (1) PBS group, (2) NIR group, (3) HMnO2@ICG group, (4) ICG + NIR group, (5) HMnO2@ICG + NIR group, (6) M-HMnO2@ICG + NIR group. Error bars represent results from three measurements.

To further validate the synergistic strategy for PDT enhancement, we examined in-vivo hypoxia levels via immunofluorescence staining of mouse tumor tissue sections for HIF-1α. Due to hypoxia in the tumor microenvironment, the PBS and NIR groups showed strong green fluorescence. Green fluorescence (and thus HIF-1α expression) was significantly reduced in the HMnO₂@ICG group, suggesting that these NPs could enhance PDT. Fluorescence was also reduced in the M-HMnO₂@ICG + NIR group relative to the control group (Figure 7F), attributable to the enhancement of PDT by self-oxygenation and CDT by the warming effect of photothermal conversion. H&E staining of the tumor tissues revealed the apoptosis of almost all tumor cells in the M-HMnO₂@ICG + NIR group (Figure 7G). TUNEL analysis showed certain degrees of apoptosis in the ICG, HMnO₂@ICG, and HMnO₂@ICG + NIR groups relative to the control and NIR groups, with the strongest apoptosis-indicating fluorescence observed in the M-HMnO₂@ICG + NIR group (Figure 7H). Based on these findings, as well as preferential accumulation at tumor sites and effective, homologous, and passive tumor targeting due to NP encapsulation, M-HMnO₂@ICG + NIR contributed to the synergistic antitumor effect of CDT/PTT/PDT.

Serum biochemical indexes did not change in any treatment group (Figure 7I–L), indicating that the treatments had excellent biosafety. H&E staining of major organs revealed no significant morphological change or inflammatory infiltration (Figure S13), confirming the long-term biosafety of the nanomaterials. The in-vivo antitumor treatment effects were consistent with the in-vitro results.

Conclusion

In this study, we used hollow mesoporous MnO₂ as a carrier to construct a simple, economical, green, and convenient nanoplatform for efficient ICG loading. We demonstrated that the MnO₂ degraded responsively in the tumor microenvironment to effectively release aggregated ICG, consume intracellular GSH in tumor cells, and produce Fenton-like effects while catalyzing O₂ production in tumor cells for sensitization to ICG-mediated PDT under laser irradiation. In addition, the good photothermal properties of ICG contributed to the enhancement of ROS-generating antitumor strategies. The HMnO₂@ICG NPs homologously targeted tumor cells and immune escape due to their HeLa cell membrane surface coatings. We systematically validated the good antitumor effects of M-HMnO₂@ICG at the in-vivo and in-vitro levels. We designed the nanoplatform to enable turn-on with MRI and accurate guidance with fluorescence and photothermal imaging. Thus, the M-HMnO₂@ICG NPs offer a new strategy for the active targeting of cervical cancer via synergistically enhancing PTT/PDT/CDT and the use of the tumor cell membrane camouflage. As multi-modal imaging-guided tumor microenvironment–responsive nanoplatforms, they provide new hope for the effectiveness of ROS-related treatment modalities.