Gelation embolism agents suppress clinical TACE-incited pro-metastatic microenvironment against hepatocellular carcinoma progression

Summary

Background

Current embolic agents in transcatheter arterial chemoembolization (TACE) of hepatocellular carcinoma (HCC) encounter instability and easy leakage, discounting TACE efficacy with residual HCC. Moreover, clinical TACE aggravates hypoxia and pro-metastatic microenvironments, rendering patients with HCC poor prognosis.

Methods

Herein, we developed Zein-based embolic agents that harness water-insoluble but ethanol-soluble Zein to encompass doxorubicin (DOX)-loaded mesoporous hollow MnO₂ (HMnO₂). The conditions and capacity of HMnO₂ to generate reactive oxygen species (ROS) were assayed. Mechanical examinations of Zein-HMnO₂@DOX were performed to evaluate its potential as the embolic agent. In vitro experiments were carried out to evaluate the effect of Zein-HMnO₂@DOX on HCC. The subcutaneous HCC mouse model and rabbit VX2 HCC model were established to investigate its anti-tumor and anti-metastasis efficacy and explore its potential anti-tumor mechanism.

Findings

The high adhesion and crosslinking of Zein with HMnO₂@DOX impart Zein-HMnO₂@DOX with strong mechanical strength to resist deformation and wash-off. Zein gelation and HMnO₂ decomposition in response to water and acidic tumor microenvironment, respectively, enable continuous DOX release and Fenton-like reaction for reactive oxygen species (ROS) production and O₂ release to execute ROS-enhanced TACE. Consequently, Zein-based embolic agents outperform clinically-used lipiodol to significantly inhibit orthotopic HCC growth. More significantly, O₂ release down-regulates hypoxia inducible factor (HIF-1α), vascular endothelial growth factor (VEGF) and glucose transporter protein 1 (GLUT1), which thereby re-programmes TACE-aggravated hypoxic and pro-metastatic microenvironments to repress HCC metastasis towards lung. Mechanistic explorations uncover that such Zein-based TACE agents disrupt oxidative stress, angiogenesis and glycometabolism pathways to inhibit HCC progression.

Interpretation

This innovative work not only provides a new TACE agent for HCC, but also establishes a new strategy to ameliorate TACE-aggravated hypoxia and metastasis motivation against clinically-common HCC metastasis after TACE operation.

Funding

Excellent Young Science Fund for National Natural Science Foundation of China (82022033); National Natural Science Foundation of China (Grant No. 82373086, 82102761); Major scientific and technological innovation project of Wenzhou Science and Technology Bureau (Grant No. ZY2021009); Shanghai Young Top-Notch Talent.

Research in context.

Evidence before this study

Transcatheter arterial chemoembolization (TACE) consisting of starvation therapy and chemotherapy is a recommended therapy for progressive HCC and has been confirmed to achieve prolonged survival and significant benefits in clinical practice. However, incomplete necrosis of HCC, ineffective accumulation and long-time resistance of chemotherapeutic agents and hypoxia in starvation therapy lead to low efficiency of TACE failing to inhibit the recurrence and metastasis of HCC. Current approaches and embolic agents to TACE embolization remain inadequate to ameliorate this deficiency.

Added value of this study

In the present work, we engineered an effective new embolic agent for ROS-enhanced TACE. On the basis of the current findings, it is suggested that Zein-HMnO₂@DOX may achieve inhibition of hepatocellular carcinoma proliferation and metastasis by involvement of oxidative stress, carbohydrate metabolism and HIF-1α/VEGF pathways.

Implications of all the available evidence

In the present study, through facile synthesis and preparation procedure, we developed Zein-based embolic agents Zein-HMnO₂@DOX, which via ROS-enhanced TACE achieved to inhibit HCC progression and metastasis to the lung.

Introduction

Hepatocellular carcinoma (HCC) is one of the worldwide malignant tumors,¹ and show poor response to the management due to hidden symptoms and delayed detection at its early onset stage.^2,3 Akin to other malignant tumors, most patients with HCC have significant resistance to conventional chemotherapy and radiotherapy,4, 5, 6 contributing to the unfavorable outcomes.⁷ Transcatheter arterial chemoembolization (TACE) consisting of starvation therapy and chemotherapy is a recommended therapy for progressive HCC in Barcelona Clinic Liver Cancer (BCLC) guidelines,⁸ and has been confirmed to achieve prolonged survival and significant benefits in clinical practice.^9,10 However, some researches have indicated that TACE fails to achieve complete necrosis of HCC, which may trigger recurrence and metastasis of the tumor.^11,12

Moreover, the ineffective maintenance or accumulation of chemotherapeutic agents used in TACE therapy remains another reason for discounting TACE efficacy, e.g., the clinically-used liquid lipiodol TACE agent is inherently unstable and easily washed away, let alone long-term drug resistance and release at the tumor site.^13,14 Although there are several attempts to modify lipiodol to increase the applicability, the liquid feature at physiological temperature (37 °C) determines that the instability and easy wash-off of lipiodol-based TACE agents remain unresolved.^15,16 In an attempt to develop new intravascular embolic agents, several microspheres including poly (lactic-co-glycolic acid) (PLGA) microspheres and new multifunctional microspheres have been documented to realize embolization therapy.17, 18, 19, 20 Nevertheless, akin to lipiodol-based TACE agents, the risk that rapid blood flow washes away the regular microspheres is increased. Coincidently, the failures in concentration accumulation and long-time resistance of chemotherapeutic drugs loaded in these microspheres at tumor disable TACE.^16,21,22 More significantly, all current transcatheter arterial embolization (TAE) or chemoembolization agents inevitably aggravate hypoxia and stimulate epidermal growth factor receptor (EGFR) upregulation, which will propel the metastasis of post-TACE HCC.23, 24, 25, 26

To tackle these concerns, we developed a distinctive Zein-based TACE agent, wherein Zein served as the main component to encapsulate doxorubicin (DOX)-loaded mesoporous hollow MnO₂ (HMnO₂). Such Zein-based embolic agents enabled highly-efficient and ROS-enhanced TACE of HCC with inhibited recurrence and metastasis (Fig. 1), addressing the efficacy inadequacy of current TACE alone. As a natural material, Zein with high safety that is easily accessible has been widely used in biomedicine,27, 28, 29, 30 and it is a hydrophobic protein but can be dissolved in aqueous ethanol.²⁸ This property determines that injectable Zein will gelate to become solid embolic agent and starve HCC once it touches blood flow, eventually realizing TAE-enabled starvation therapy. This property was also expected to enable Zein to confine HMnO₂@DOX and extend the release profile of DOX for chemotherapy, which integrated with TAE-enabled starvation therapy into TACE therapy.

Fig. 1.

Action principles using such composite embolic agents combining with TACE therapy and CDT therapy. Herein, Zein was implemented as an embolic agent for TACE in rabbit VX2 model. Zein-HMnO₂@DOX composite system was introduced with DOX-loaded HMnO₂, which could generate ROS via the HMnO₂-mediated Fenton-like reaction in the acidic tumor microenvironment. The two components cooperated with each other to reduce vascular density (CD31 and CD34 downregulations), resulting in vessel occlusion, cutting off oxygen and nutrition supply and executing ROS therapy and TACE therapy consisting of ROS therapy and chemotherapy. Moreover, the O₂ release from HMnO₂-mediated CDT inactivates HIF-1α, further down-regulates the expressions of VEGF and GLUT1, which are available for mitigating TACE-aggravated hypoxia and metastasis motivation to inhibit HCC progression and lung metastasis. Created with Biorender.com.

Inspiringly, the high adhesion and cross-linking of Zein with HMnO₂@DOX allow Zein-HMnO₂@DOX to feature high mechanical strength against blood pressure-induced deformation and blood flow-washed shedding, addressing the instability, leakage and wash-off that current embolic agents suffer from. In particular, the hydrophilic component (e.g., polysaccharides) of Zein is also designed to reinforce storage stability of Zein-based embolic agents in water.³¹ Such solid Zein-based TACE agents outweighed clinically-used liquid hydrophobic lipiodol in avoiding drug leakage. Concurrently, they outperformed lipiodol-based TACE agents to induce TAE-unlocked starvation therapy and TACE therapy, and delay HCC progression with decreased intratumoral vascular density (CD34 and CD31 downregulation). Herein, HMnO₂, as the chemodynamic therapy (CDT) agonist, was used to mediate a Fenton-like reaction in response to acidic tumor microenvironment,^32,33 realizing glutathione (GSH) depletion, ROS production and DOX & O₂ release.33, 34, 35, 36 Regarding it, HMnO₂ was expected to directly enhance Zein-based TACE therapy via triggering ROS therapy, overcoming the inadequacy of current TACE. Intriguingly, HMnO₂-unlocked O₂ release in such Zein-based embolic agents was leveraged to remodel the Zein TAE or TACE-aggravated hypoxia microenvironment (HIF-1α) and metastasis impetus (VEGF and GLUT1 downregulations) for further inhibiting cancer metastasis (Fig. 1).³⁷

In short, the marriage of the above action principles allowed such composite Zein-HMnO₂@DOX TACE agents to combine starvation therapy, chemotherapy and ROS therapy. Consequently, a powerful inhibitory effect on HCC progression in the orthotopic rabbit VX2 tumor model was reached, with which the significantly-decreased metastatic lesions in the lung were accompanied. More significantly, sequencing was proceeded to uncover the underlying principles, and oxidative stress and carbohydrate metabolism were identified to be responsible for the significantly-elevated anti-tumor and anti-metastasis outcomes. In summary, this innovative work not only provides a new TACE agent for HCC, but also establishes a new strategy to ameliorate TACE-aggravated hypoxia and metastasis motivation against clinically-common HCC metastasis after TACE operation.

Methods

Detailed information of all used reagents

The names, suppliers and catalog numbers of main employed reagents in this study are given in Table S1.

Detailed information of all used antibodies

The names, suppliers and catalog numbers of employed antibodies for western blotting, immunofluorescence, immunohistochemistry are given in the Tables S2–S4. All the antibodies are validated.

Detailed information of all used cell lines

Human hepatocellular carcinoma cell lines Hep3B, Huh7, and HepG2, and mouse hepatocellular carcinoma cells Hep1-6 were provided by the Institute of Liver Cancer Research, Zhongshan Hospital, Fudan University and the First Affiliated Hospital of Wenzhou Medical University. HUVEC cells were provided by the Laboratory of Ultrasound Department, Shanghai Tenth People's Hospital. Recent STR profiling of all the cell lines has been carried out.

Detailed information of all used animals

C57BL/6 male mice (weighing 18–20 g) at 5–6 weeks of age, male Kunming mice (weighing 20–22 g) at 6–8 weeks of age, and SD male rats (weighing 200–220 g) at 5–6 weeks of age were used in this experiment and were purchased through the Institute of Liver Cancer Research, Fudan University and Shanghai Tenth People's Hospital. All experimental mice were housed in SPF-grade laboratory animal rooms. The mice were fed with sterilized water and chow, which were freely consumed by the experimental mice.

Healthy, male New Zealand white rabbits, conventional grade, weighing about 2.5–3 kg, were purchased from the Animal Center of Zhongshan Hospital, Fudan University and Shanghai Tenth People's Hospital, and the VX2 tumor mass was provided by the First Affiliated Hospital of Wenzhou Medical University. The animals were kept according to the requirements of conventional grade, maintaining a suitable environmental temperature of 20–25 °C, relative humidity of 40–70%, air changes of 10–15 times/hour, air cleanliness of 10,000 level, and free access to food and water.

To enhance the success rate of establishing animal tumor models and to avoid the influence of estrogens and estrogen receptors on HCC, only male animals were used in the relevant experiments conducted in this study.

Ethics

All animal experiments were performed according to protocols approved by the Laboratory Animal Center of Shanghai Tenth People's Hospital and were in accordance with the policies of the National Ministry of Health (Approval number: SHDSYY-2021-Y3216).

Preparation of HMnO₂@DOX

Silica (SiO₂) nanoparticles were prepared by mixing tetraethyl orthosilicate (TEOS), anhydrous ethanol, deionized water and ammonia. Hollow manganese dioxide nanoparticles were synthesized by using spherical silica as a template with potassium permanganate (KMnO₄) under ultrasonication and etching with sodium carbonate (Na₂CO₃), and the surface of the nanoparticles was modified with PEG. HMnO₂@DOX was obtained by mixing adriamycin hydrochloride with it for 12 h under light-avoidance conditions with stirring.

Preparation of Zein-HMnO₂@DOX composite drug-carrying embolic system (Zein-HMnO₂@DOX)

Zein was dissolved in 80% ethanol solution and stirred for 4 h to form a homogeneous dispersion of Zein in alcohol. Under the action of ultrasound, HMnO₂@DOX nanoparticles resuspended in 80% ethanol were added to the above Zein ethanol solution and stirred for 2 h to prepare Zein-HMnO₂@DOX composite drug-carrying embolic system for spare.

Characterizations

The sample of aqueous solution of HMnO₂@DOX nanoparticles was aspirated and dropped slowly on a transmission electron microscope grid, left to dry, and then the morphology of HMnO₂@DOX nanoparticles was observed and photographed using TEM. A small amount of prepared Zein-HMnO₂@DOX composite drug-carrying embolization system was taken and placed in a copper grid and the morphology and structure were observed using scanning electron microscope. HMnO₂@DOX was taken for sufficient resuspension to make it uniformly dispersed in deionized water. A small amount of resuspension was aspirated and dropped into a nano-laser particle sizer spiking dish, and its particle size and zeta potential were measured, respectively, and the measurements need to be repeated three times. The pore size and nitrogen desorption curves of the prepared nanoparticle powders were determined using a specific surface area and pore structure analyzer.

Drug release test

Adriamycin was prepared and diluted into aqueous solutions of different concentrations in different ratios. The UV absorption spectra with wavelengths in the range of 100–800 nm were determined separately, and a standard curve was established based on its highest absorption peak at 480 nm for the calculation of loading efficiency and loading capacity. PBS with different pH values (pH = 5.5 and pH = 7.4) was selected as the drug release media to simulate the drug release behavior under neutral and acidic conditions, respectively. HMnO₂@DOX was dissolved in the release media of different pH and placed in a dialysis bag, which was then placed in a beaker filled with PBS (pH = 7.4) at a constant temperature of 37 °C with full and slow stirring and oscillation. Samples were taken from the beaker at different time intervals. Immediately after sampling, the beaker was replenished with the equal volume of PBS (pH = 7.4) and then stirring and shaking was continued until the end of sampling at all time points. The absorbance of the samples at different time points at a wavelength of 480 nm was measured using a UV-Vis spectrometer. The amount of DOX released at different time points and pH conditions was calculated from the standard curve.

Phase transition test of Zein

Use a 1 mL syringe to remove 1 mL of the Zein-HMnO₂@DOX composite drug-carrying embolic system prepared as described above, and after inserting the syringe needle below the liquid level of PBS, slowly push the syringe to inject Zein-HMnO₂@DOX into the aqueous phase, and observe the changes and take photographs.

Mechanical study on Zein embolic agent

A 1 mL syringe was used to take 1 mL of the Zein-HMnO₂@DOX composite carrier embolic system, which was injected into the aqueous phase of PBS to make its phase change into a solid state. The Zein-HMnO₂@DOX that had undergone phase change in PBS was picked up with tweezers, and the solid Zein-HMnO₂@DOX was prepared by using a mold (10 mm in diameter and 10 mm in height), and the prepared solid material was used to measure the stress and strain using a thermo-mechanical analyzer.

O₂ release monitoring

The oxygen-generation capacity of HMnO₂ in acidic and H₂O₂-rich tumor microenvironment was measured by monitoring the O₂ production at the various concentrations of HMnO₂ containing H₂O₂ (100 μM) using a portable dissolved oxygen meter (JPBJ-608 portable Dissolved oxygen).

Drug releasing in a composite drug-carrying embolic system (Zein-HMnO₂@DOX)

An imitated vessel model was designed to imitate the normal diameter of the hepatic artery (about 2–5 mm, set at 5 mm), and the length of the vessel was set at about 20 mm. One end of the imitated vessel was blocked with AB adhesive, then saline was injected into the vessel, and 1 mL of Zein-HMnO₂@DOX was injected into the other end of the vessel, and the concentration of adriamycin in the liquid was determined by UV spectrometry after liquid samples were extracted through the outer wall of the imitated vessel at different time points. The drug release efficiency of Zein-HMnO₂@DOX in vitro was calculated by measuring the concentration of adriamycin in the extracted fluid using ultraviolet spectrometer after the fluid samples were extracted from the outer wall of the simulated vessel using a syringe at different time points.

Cellular uptake

Hep1-6, Hep3B, and Huh7 cells were cultured in laser confocal cell culture dishes, and DMEM containing FITC-labeled HMnO₂ was added. The medium containing nanoparticles was aspirated at different time points. Serum-free medium was added to wash three times until the nanoparticles were washed. After the serum-free medium was aspirated, anti-fluorescence quenching blocking solution containing DAPI was added and incubated at 4 °C for 15 min. Then, the cells were placed under a laser confocal microscope for observation and photography. The excitation wavelength of FITC was 495 nm, and the emission wavelength was 520 nm.

In vitro biocompatibility study

After HUVEC and Hep1-6 cells were added to 96-well plate culture with HMnO₂, CCK8 detection kit was added to each well. The temperature condition was kept at 37 °C and left for 1 h. The absorbance wave at 450 nm was detected by multifunctional enzyme marker. Multiple replicate wells were set up for different groups and time points. The blank group was set as the control without cells and only medium was added.

HMnO₂@DOX inhibits the proliferation of hepatocellular carcinoma cells

HepG2, Hep3B, Huh7, and Hep1-6 cells were cultured in 96-well plates. Different concentrations of adriamycin and drug-loaded HMnO₂@DOX nanoparticle suspension were added respectively. After continuing the culture for 24 h, a CCK8 detection reagent was added in each well. The temperature condition was 37 °C and left for 1 h. The absorbance value at 450 nm was detected by multifunctional enzyme marker. Multiple replicate wells were required for different groups and time points. The blank group is the control wells without cells and only medium added.

Reactive oxygen species (ROS) detection by flow cytometry test

Detection of intracellular ROS content was performed by DCFH-DA fluorescent probe labeling method. Hep1-6, Hep3B and Huh7 cells were spread into six-well plates for culture. DMEM, adriamycin solution (100 μg/mL), HMnO₂ nanoparticle suspension (100 μg/mL) and HMnO₂@DOX (100 μg/mL) were added in different subgroups with 2 mL each, and the cells were washed with PBS after being cultured for 8 h in a constant temperature incubator. Diluted DCFH-DA with serum-free culture medium according to 1:1000 to a concentration of 10 μM was proceeded. Subsequently, we removed the cell culture medium and added 1 mL of DCFH-DA (10 μM) to each well. Temperature condition was incubated at 37 °C for 20 min. The cell suspension was collected after trypsin digestion and centrifuged, and the cells were resuspended after discarding the supernatant. The fluorescence intensity at 488 nm excitation wavelength, 525 nm emission wavelength was detected using flow cytometry.

ROS detection by LSCM

The detection principle was the same as that of the flow cytometry assay described above. Hep1-6 mouse hepatocellular carcinoma cell and Hep3B and Huh7 human HCC were spread into laser confocal cell culture dishes for culture. DMEM containing HMnO₂ was added. The concentration of HMnO₂ was 100 μg/mL. At predetermined time points (0, 2, 4, 8 h), the nanoparticle-containing medium was discarded and washed three times with serum-free double-antibody-free medium. Incubation of the probes was performed as in the previous flow cytometry technique. The probe-containing medium was aspirated and washed. Antifluorescence quenching blocking solution containing DAPI was added. The probes were incubated at 4 °C for 15 min, and were observed and photographed under a laser confocal microscope at 488 nm and 525 nm emission wavelengths.

Calcein AM/PI cell viability assay

Prepare the Calcein AM/PI working solution according to the manufacturer's instructions and mix thoroughly. Cells with different treatments were washed with PBS. An appropriate volume of the Calcein AM/PI working solution was added. The cells were then incubated in the dark at 37 °C for 30 min. The staining effect was observed under a fluorescence microscope.

Trypan blue staining cell viability assay

After digestion with trypsin, the adherent cells were collected and centrifuged at a speed of 1000–2000g for 1 min. The supernatant was discarded, and the cells were resuspended in an appropriate volume of cell suspension solution. 100 μL of the resuspended cells was taken, and 100 μL of trypan blue staining solution (2×) was added. The mixture was gently mixed and stained for 3 min. A small amount of the stained cells was aspirated, and a hemocytometer was used to count the cells under a microscopy.

Western blotting analysis

Isolated tumor tissues were thoroughly ground and cleaved after addition of RIPA, protease inhibitor, and phosphatase inhibitor. Collect digested cells from different treatment groups for protein extraction. Proteins in the supernatant were collected after centrifugation for BCA protein quantification. After varied formulations and the standard procedures, the PVDF membrane was incubated with the primary antibodies for HIF-1α, VEGF, GLUT1, Bcl-2, and β-actin followed by treated with the secondary antibody for visualization using ECL detection reagents.

RNA isolation and library preparation

Total RNA was extracted using the TRIzol reagent according to the manufacturer's protocol. RNA purity and quantification were evaluated using the NanoDrop 2000 spectrophotometer. RNA integrity was assessed using the Agilent 2100 Bioanalyzer. Then the libraries were constructed using VAHTS Universal V6 RNA-seq Library Prep Kit according to the manufacturer's instructions. The transcriptome sequencing and analysis were conducted by OE Biotech Co., Ltd.

RNA sequencing and differentially expressed genes analysis

The libraries were sequenced on an Illumina Novaseq 6000 platform and 150 bp paired-end reads were generated. Raw reads of fastq format were firstly processed using fastp and the low quality reads were removed to obtain the clean reads. The clean reads were mapped to the reference genome using HISAT2. FPKM of each gene was calculated and the read counts of each gene were obtained by HTSeq-count. PCA analysis was performed using R (v 3.2.0) to evaluate the biological duplication of samples. Differential expression analysis was performed using the DESeq2. Q value < 0.05 and foldchange > 2 or foldchange < 0.5 was set as the threshold for significantly differential expression gene (DEGs). Differentially expressed genes analysis was operated on the online platform of OE Biotech Co., Ltd.

In vivo anti-tumor performance of Zein-HMnO₂@DOX in xenografts model

Injections of saline containing Hep1-6 cells (1 × 10⁸ cells per milliliter) subcutaneously into the backs of C57BL/6 male mice were established to produce Hep1-6 tumor xenografts, which were continuously maintained for nearly 2 weeks until the tumor volume was nearly 100 cubic millimeters. Then, these mice carrying Hep1-6 tumors were randomly divided into five groups: control group (saline injection), Zein group, Zein@DOX group, Zein-HMnO₂ group and Zein-HMnO₂@DOX group. These mice were injected intratumorally with different formulations on the first day of treatment. The body weights of the tested mice and the length and width of the tumors were measured every 3 days. Measurements were taken for 21 consecutive days. Tumor tissues from each group were collected and histopathological studies such as H&E, HIF-1α, and VEGF antibody staining were performed. The instructions of the Regional Ethics Committee for Animal Experiments Animal Care were followed throughout the animal experiments and the protocol was approved by the Ethics Committee.

Metabolism, distribution and biocompatibility of HMnO₂ in vivo

HMnO₂ (concentration: 100 μg/mL, volume: 100 μL) was injected through the tail veins of Kunming mice and SD rats, and venous blood was collected at different time points (0 min, 5 min, 10 min, 20 min, 30 min, 1 h, 2 h, 4 h, 6 h, 12 h, 24 h, and 48 h) after injecting the nanomaterials. The manganese content of the collected venous blood samples was determined. Hep1-6 tumor xenografts were established by subcutaneous injection of saline containing Hep1-6 cells (1 × 10⁸ cells per milliliter) into the backs of C57BL/6 male mice. HMnO₂ labeled by rhodamine B (concentration: 100 μg/mL) was injected through the tail vein. At different time points after injection of nanomaterials (0, 1, 2, 4, 6, 8, 10, 12, and 24 h), the emission wavelength of 520 nm channel was selected, and fluorescence images were captured by a small animal in vivo imager to observe the distribution of nanomaterials labeled with rhodamine B dye. After tail vein injection, mice at different time points (0 d, 1 d, 7 d, 14 d, and 28 d) were examined for the biosafety of blood routine, liver and kidney function examination by collecting venous blood, and pathologic detection of critical organs by H&E staining.

In vivo anti-tumor performance of Zein-HMnO₂@DOX on rabbit VX2 tumor model

All rabbit interventional procedures were performed by experienced interventional radiologists in interventional therapy. New Zealand white rabbits (weighing 2.5–3.0 kg) were used to construct rabbit VX2 tumor models. Briefly, after 12 h of fast, the rabbits were abdominally dissected, and a VX2 tumor mass was implanted in the left lobe of the liver, and the abdominal cuts were sterilized after being sutured layer by layer. Postoperatively, the experimental rabbits were observed for general status, diet, and surgical incision. About 10–14 days after the implantation of VX2 mass, the size and morphology of liver tumor were observed by enhanced CT (designated as Day 0). A single tumor nodule in the left lobe of the liver of the rabbits with uniform density and clear boundary was confirmed by enhanced CT, which was confined to the liver parenchyma, and the rest of liver had uniform texture and no metastatic lesions were found.

36 male rabbits (weighing 2.5–3.0 kg) bearing VX2 tumors were randomly divided into six groups (n = 6), respectively, Groups A-F. Group A: Control (saline); Group B: conventional TACE, i.e., Lipiodol-DOX, which was delivered supra-selectively into the tumor-supplying arteries, and then the arteries were blocked with gelatin sponges to prevent leakage out of the embolized arteries; Group C: Lipiodol-HMnO₂@DOX; Group D: Zein; Group E: Zein-DOX; Group F: Zein-HMnO₂@DOX. Supra-selective operations in Groups C–F were performed as those in group B. TACE was performed under digital subtraction angiography (DSA) using a microcatheter through a superselective catheter to localize tumor tissue and its vascular system. The intraoperative angiography of TACE showed that the common hepatic artery runs into the proper hepatic artery, which was divided into left and right branches to supply the left and right lobes after entering the liver, respectively. All VX2 tumors were located in the left lobe of liver and were supplied by branches of left hepatic artery.

After ascertaining the left hepatic artery, embolic agents in different groups were injected into the left hepatic artery according to the administration regimens mentioned above. The endpoint of embolization was set as the absence of tumor staining or the presence of Lipiodol deposition within the small branches of the portal vein after injecting above agents. In contrast, in the Lipiodol-free groups, 0.5 mL of iophorol as contrast agents of DSA was added to the mixture and injected (to facilitate imaging) until contrast reflux was observed. All rabbits were examined by thoracic and abdominal enhanced CT approximately 14 days after the TACE therapy procedure. Enhanced CT detection method: the marginal ear venous channel was connected to a high-pressure syringe, and the total amount of iophorol contrast agent was 7 mL + 3 mL of saline was drawn under negative pressure. The high-pressure syringe was set at a rate of 0.3 mL/s. A tracking system was applied to monitor the abdominal aorta at the diaphragmatic level, and the arterial phase scan was started at the time of peak contrast density (about 130 Hu) (about 26 s), and the venous phase scan was performed about 50 s and the delayed phase scan was performed about 65 s after the completion of the scan. The scanning parameters were set at the fixed parameters: tube voltage of 120 KV, tube current of 250 mA, scanning field of view (FOV) of 200 mm × 200 mm, layer thickness of 2.5 mm, and all the images were reconstructed with a layer thickness of 1 mm. The size of the tumor diameter was measured, and the presence of metastatic foci in the lungs was observed, and the tumor volume was calculated according to the formula: V (cm³) = the largest cross-sectional diameter of tumor on CT scanning plane² (cm²) × the number of layers of tumor area on CT scanning × the thickness of CT scanning slice (cm)/2.

Blood specimens of rabbits were collected for blood routine, liver and kidney function, and myocardial enzyme spectrum laboratory tests. Rabbit liver, heart, lung, kidney, spleen, lung and tumor tissues were collected for H&E staining. Terminal deoxynucleotidyl transferase mediated dUTP biotin nick end labeling (TUNEL), ki67 immunofluorescence and intratumor DOX fluorescence imaging were performed on tumor tissues to detect tumor proliferation, apoptosis and DOX intratumor distribution. The embolization effect was evaluated by immunohistochemistry of CD31, CD34, CD41 on tumor tissues. Tumor tissues were evaluated by western blot, immunohistochemistry and immunofluorescence to detect changes in the expression of HIF-1α, VEGF and GLUT1.

Quantitative analysis for immunohistochemistry and immunofluorescence

The prepared samples were observed under a fluorescence microscope, and images were captured for analysis. All tested tissue slides underwent manual white balance, and photographs were taken of randomly selected areas under identical environmental conditions. The immunohistochemical and immunofluorescence results of the tissue slides were quantitatively analyzed using Image Pro Plus 6 image analysis software (Media Cybernetics Inc., Bethesda, MD) through Integral Optical Density (IOD) assessment. Firstly, open the image for analysis and perform intensity calibration. Secondly, select positive areas by clicking on Measure-Count/Size. Thirdly, select the measurement indicators by clicking on Measure-Select Measurements and choose “area” and “IOD”. Fourthly, click on Count to perform the counting and View-Statistics to obtain the statistical results. In the results, “area” and “IOD” represented the size of the positive areas and the fluorescence intensity, respectively.

Statistical analysis

All the experiments were performed in triplicate. The obtained data were expressed as the mean value ± standard deviation (SD) and the statistical significance between two groups was analyzed by unpaired Student's t test through SPSS 19.0 (SPSS, Chicago, IL, USA). For continuous variables that follow a normal distribution across multiple groups, a one-way ANOVA is initially employed to compare between groups. Depending on the different hypotheses and variables of interest, Tukey's multiple comparisons test is then utilized to conduct pairwise comparisons. Single, double, triple and four asterisks represent p ≤ 0.05, 0.01, 0.001 and 0.0001, respectively, wherein ∗p < 0.05 was considered statistically significant and ∗∗∗∗p < 0.0001 was extremely significant.

Role of funders

The funders had no role in the study design, data collection, data analysis, interpretation or writing of this report.

Results

Synthesis of Zein-based TACE agent (Zein@HMnO₂@DOX)

Zein-based TACE agent (Zein@HMnO₂@DOX) was obtained via dispersing HMnO₂@DOX in Zein, wherein mesoporous HMnO₂ nanoparticles were firstly synthesized via the classic etching method,³² followed by DOX co-entrapment (Fig. 2a). Spherical HMnO₂ nanoparticles with 100 nm in diameter are successfully fabricated (Fig. 2b and Figure S1), and Zeta potential tracking shows the surface potential variation of HMnO₂ and its immediate products (Figure S2), verifying the successful synthesis of HMnO₂. Apparently, hollow structure and mesoporous channels with large pore diameter and surface area are clearly observed (Fig. 2b–d). This determines that HMnO₂ can serve as vehicles to encompass massive chemotherapeutic drug (DOX) (Fig. 2e), which are favorable for ensuring the stability, avoiding the leakage and attaining the continuous release of DOX. The loading ratio of DOX in HMnO₂ is obtained via referring to the UV absorbance-concentration standard curve of DOX (Figure S3). The loading ratio is tunable as a function of the feeding ratio of DOX:HMnO₂ (Figure S4), and the ideal feeding ratio of DOX:HMnO₂ should be set as 3:1 under which the approximate saturation loading ratio (20%–25%) is obtained.

Fig. 2.

Preparation, physiochemical parameters and functional tests of Zein-HMnO₂@DOX. (a) Schematic illustrations of the synthetic process of HMnO₂@DOX and Zein-HMnO₂@DOX. (b) Transmission electron microscopy (TEM) image of HMnO₂ nanoparticles, HAADF-STEM and element mapping images of Mn, O of HMnO₂ nanoparticles (Scale bar = 50 nm). (c, d) N₂ adsorption and desorption isotherms (c) and pore-size distribution profile (d) of HMnO₂ nanoparticles. (e) UV spectrum of HMnO₂, HMnO₂@DOX (50 μg/mL) and HMnO₂@DOX (100 μg/mL) (f) ESR spectra in different groups (i.e., DOX, HMnO₂, and HMnO₂@DOX). for detecting •OH using DMPO as capturing agent. (g) CLSM images of Hep1-6 cells after different treatments (i.e., Control: PBS; Positive: Rosup in DCFH-DA kits; HMnO₂ and HMnO₂@DOX) and subsequent ROS indicator (i.e., DCFH-DA) staining (Scale bar = 50 μm). (h) Flow cytometry (FCM) patterns of DCFH-DA-stained Hep1-6 cells for tracking ROS levels after different treatments. (i) Dissolved O₂ level in the mixture of HMnO₂ (0 and 20 μg/mL) and H₂O₂ (100 μM) at pH = 6.0, wherein portable dissolved oxygen meter (JPBJ-608 portable Dissolved oxygen) was used to track them. (j) Phase change tests of Zein, wherein ① represents the injection of Zein solution into PBS; and ② represents PBS injection into Zein solution. (k) Morphology of Zein-HMnO₂@DOX, wherein ① and ② represent digital photographs of Zein solution and Zein-HMnO₂@DOX solution, respectively; and ③, ④ and ⑤show the scanning electron microscope (SEM) images of Zein-HMnO₂@DOX solution. (l) SEM image of Zein-HMnO₂@DOX solid state, wherein ① and ② represent images of the surface of Zein-HMnO₂@DOX mass (Scale bar = 1 μm); and ③ (Scale bar = 50 μm) and ④(Scale bar = 20 μm) represent images of the cut-cross section of Zein-HMnO₂@DOX mass. (m) In vitro vascular simulation embolization of Zein-HMnO₂@DOX, wherein ① represents the image of cut-cross section of simulated vessel with the embolism of Zein-HMnO₂@DOX (Scale bar = 1 mm); and ② represents part of the image of ① (Scale bar = 200 μm). The high-power field of SEM images of Zein-HMnO₂@DOX embolic agents show in image ③ (Scale bar = 100 μm) and ④ (Scale bar = 20 μm). (n) Release profiles of DOX from Zein-HMnO₂@DOX at pH = 6.0. (o) Strain–stress curves of Zein and Zein-HMnO₂@DOX, respectively. (p, q) Force-displacement curves of Zein (p) and Zein-HMnO₂@DOX (q), respectively, as determined by rheological analysis. Data are expressed as mean ± standard deviation (SD) (n = 3).

Besides serving as carriers, HMnO₂ nanoparticles can trigger the Fenton-like reaction under acidic condition, during which hydroxyl radicals (•OH) are expected to burgeon.³⁸ To verify it, electron spin resonance (ESR) spectra of different samples in 5,5-dimethyl-1-pyrrolidine oxide (DMPO)-rich solution were monitored.³⁹ Compared to DOX, HMnO₂ and HMnO₂@DOX are evidenced to give birth to abundant •OH via the Fenton-like reaction with H₂O₂ at low pH since remarkable characteristic peaks of •OH emerge (Fig. 2f). Subsequently, cellular-level ROS birth via the HMnO₂-mediated CDT was inspected using the classic ROS indicator (e.g., DCFH-DA),40, 41, 42 and HMnO₂ nanoparticles show a high engulfment by HCC cells (i.e., Hep1-6, Hep3B and Huh7) (Figures S5–S7) for ensuring the reliability of HMnO₂-mediated ROS production assay. HMnO₂-contained groups indeed induce stronger green fluorescence than Control (Fig. 2g), suggesting ROS production by HMnO₂-mediated H₂O₂ decomposition. In particular, DOX itself fails to trigger ROS production, but may stimulate HMnO₂-mediated CDT process to produce more ROS (Fig. 2f and g). Furthermore, flow cytometry (FCM) analysis also directly reveals the ROS production by HMnO₂@DOX when comparing to Control (Fig. 2h).43, 44, 45 Notably, the annihilator, i.e., GSH, was pre-introduced to remove the produced ROS (Fig. 2h), indirectly reflecting HMnO₂-mediated ROS birth. Apart from •OH, O₂ release from HMnO₂-containing H₂O₂ solution whose pH is set as 6.0 is accessible since O₂ is another product of HMnO₂-mediated CDT at low pH condition (Fig. 2i). The results adequately suggested that HMnO₂ nanoparticles indeed participated in Fenton-like reaction and disintegrated into Mn²⁺ under H⁺ condition, enabling ROS therapy and continuous chemotherapy against tumor since tumor microenvironment is acidic and can stimulate DOX release (Figure S8).

Zein is hydrophobic in water, but is soluble in ethanol,28, 29, 30 meaning that Zein solution in ethanol is injectable and can uniformly mix with HMnO₂@DOX. As expected, Zein is dissolved sufficiently in 80% ethanol solution, and can be injected into aqueous PBS, consequently shaping into solid blocks (image ① in Fig. 2j), enabling the TAE and TAE-unlocked starvation therapy. Although the inverted sequence, i.e., injecting PBS into Zein solution in ethanol, fails to create blocks (image ② in Fig. 2j), this inverted sequence is inapplicable in actual clinical TACE. Comparing to Zein alone in ethanol (image ① in Fig. 2k), HMnO₂@DOX nanoparticles are evenly dispersed in Zein ethanol solution (images ② and ③ in Fig. 2k). The obtained solid Zein-HMnO₂@DOX after touching water shows micron-scale pore channels (images ④ and ⑤ in Fig. 2k), wherein Zein shows a reticular cross-linked structure and adhesively wrap HMnO₂@DOX nanoparticles (image ⑤ in Fig. 2k). SEM images of the obtained solid Zein-HMnO₂@DOX are illustrated. It is found that the surface presents a granular cross-linking structure (images ① and ② in Fig. 2l), and the cross-section of the solid Zein-HMnO₂@DOX displays a honeycomb-like pore structure after cross-linking (images ③ and ④ in Fig. 2l), which thus provide large capacity for loading MnO₂@DOX. Besides, an adherence test with pig skin to simulate human tissue reveals that Zein displays adhesiveness on tissues, and both HMnO₂ and HMnO₂@DOX loading with Zein fail to influence the tissue adhesion strength of Zein network (Figure S9). To confirm the embolic effect of Zein-HMnO₂@DOX on vascular occlusion, a set of simulated vessel models was employed in vitro. Since the vessel diameter of the intrahepatic artery (hepatic left and right hepatic arteries) is about 2–3 mm, the simulated vessel models with the same internal diameter were employed. The saline, which simulates blood flow (liquid phase), was infused into the simulated blood vessels to replicate the vascular environment in the physiological state. After Zein-HMnO₂@DOX was injected into the above simulated blood vessels, a liquid-to-solid phase transition that occludes the vessel occurs to Zein-HMnO₂@DOX, and the injection pressure increases significantly until no more embolic agents are allowed to be injected into the model. SEM inspection is proceeded to observe and trace the occlusion process and outcome (① and ② in Fig. 2m). Therein, Zein-HMnO₂@DOX narrows and further occludes the majority of the lumen space, which significantly cut off saline flow in the simulated blood vessel and thus satisfy the demand of tumor vascular embolism. The magnified SEM images reveal that Zein-HMnO₂@DOX can shape into a dense solid structure in the vessels, and the solid embolic agents are evolved into a honeycomb appearance (③ and ④ in Fig. 2m). By virtue of cross-linking and wrapping by Zein, the obtained solid Zein-HMnO₂@DOX along with HMnO₂ disintegration in response to acidic tumor microenvironment dictate the continuous DOX release-mediated chemotherapy (Fig. 2n and Figure S8), accounting for the distinctive TACE against HCC.

Because the embolic material will be subjected to the blood flow and blood pressure during vascular embolization, the new composite embolic material requires a certain mechanical strength to resist blood flow erosion in order to maintain the original shape. To figure it out, stress–strain analysis was firstly carried out, and results show that such Zein-HMnO₂@DOX TACE agents produce larger elastic moduli than Zein alone (Fig. 2o). Additionally, rheological analysis was proceeded to analyze the mechanical properties, and the deformation degree of Zein-HMnO₂@DOX is inferior to Zein alone under the same pressure (Fig. 2p, q). All of the above confirm that the crosslinked structure of Zein after phase transformation can provide sufficient mechanical strength to maintain a sustained embolic state in blood vessels.

Anti-tumor evaluations based on HMnO₂-mediated ROS therapy and DOX chemotherapy

Cellular-level TAE-mediated starvation therapy of Zein is impracticable due to the absence of blood vessels. Herein, HMnO₂-mediated ROS therapy and DOX therapy were assessed since both contributions are also the pivotal components of such enhanced TACE treatment of HCC. Prior to it, we tested the biosafety of HMnO₂ carriers, and HMnO₂ carriers are found to have no cytotoxicity on normal cells (i.e., HUVEC) (Figure S10) since the cell viability remains above 85% even at the concentration of 100 μg/mL. In the safety experiments in vivo, after 48 h post-intravenous injection, the plasma concentration of manganese continued to decline to a low level (Figure S11), suggesting that HMnO₂ exhibits a rapid metabolism trend, guaranteeing the in vivo safety. The rapid metabolism and specific ROS production only in tumor allow such HMnO₂ carriers to induce ROS therapy at tumor, and will not do damages to other normal organ tissues in the long-term evaluations (Figure S12), with which no evident variations of blood, liver and kidney indices are accompanied (Figure S13). HMnO₂ has a long-term biological safety in vivo and has no effects on blood system, liver and kidney function. Pharmacokinetic and biosafety studies in normal rats also receive the same findings (Figures S14–S16).

As described in above in vitro experiments that HMnO₂ is more likely to participate in Fenton-like reactions under acidic conditions, releasing encapsulated chemotherapeutic drugs and Mn²⁺. In light of the fact that the pH value in tumor is low,^46,47 which is applicable for HMnO₂@DOX disintegration to release DOX and ROS therapy to kill tumor cells. In the CCK8 assay, DOX alone is found to have an inhibitory effect on the proliferation of HepG2, and Hep3B, Huh7 human hepatoma cell lines and Hep1-6 mouse hepatoma cell line in a concentration-dependent manner (Fig. 3a and Figure S17). Remarkably, after combination with HMnO₂-mediated ROS therapy, more cell deaths including HepG2, Hep3B and Huh7 and Hep1-6 cells are obtained in HMnO₂@DOX-treated groups. Confocal microscopy observation after calcein-AM/propidiumiodide (PI) co-staining was implemented to explore the anti-tumor ability. Results reveal that HMnO₂ alone indeed elicits Hep1-6 deaths via the Fenton-like ROS birth, and DOX alone induces more cell deaths. The combination of DOX chemotherapy and HMnO₂-enabled ROS therapy allows HMnO₂@DOX to exert the most powerful anti-tumor activity and bring about almost all Hep1-6 deaths, as represented by red fluorescence (Fig. 3b). Consistent conclusions were obtained in the human hepatocellular carcinoma cell lines Hep3B and Huh7 (Figures S18 and S19). Besides, Trypan Blue Staining Cell Viability Assay to perform live/dead cell assay on Hep3B and Huh7 cells with different treatment also yielded the same conclusion (Figures S20 and S21).

Fig. 3.

Anti-tumor evaluations of ROS therapy and DOX-mediated chemotherapy. (a) Relative viabilities of HepG2 cells and Hep1-6 cells after incubation with different treatments (i.e., DOX and HMnO₂@DOX). (b) CLSM images of Hep1-6 cells co-stained with calcein-AM/PI after different treatments for 12 h (Scale bar = 50 μm). (c) Time axis of in vivo anti-tumor experiments and operation details. Created with Biorender.com. (d) Detailed tumor–growth curves of each mouse per group after different treatments including Control, Zein, Zein@DOX, Zein-HMnO₂ and Zein-HMnO₂@DOX (Scale bar = 50 μm). (e, f) Digital photographs (e) and tumor volumes (f) of dissected tumors isolated from Hep1-6 tumor-bearing mice that experienced corresponding treatments in different groups. (g) Time-dependent average body weights of Hep1-6 tumor-bearing mice that experienced different treatments in corresponding groups. (h) Immunohistochemical images of tumor slices isolated from Hep1-6 tumor-bearing mice that experienced corresponding treatments in different treatment groups, including HE, HIF-1α, and VEGF. Data are expressed as mean ± SD (n = 4). One-way ANOVA followed by Tukey's multiple comparisons test were used to test the difference significance between the groups, and ∗p < 0.05 vs Control group.

Beyond the cellular-level anti-tumor evaluations, the in vivo anti-tumor outcomes of HMnO₂-mediated ROS therapy and DOX chemotherapy were carried out. The detailed experiment procedure is shown in Fig. 3c, wherein subcutaneous Hep1-6 xenografted tumor model on mice was employed. In light of the fact that femoral artery puncture for injecting embolic agents into subcutaneous tumor is impracticable, such a subcutaneous tumor model is inappropriate for evaluating Zein-mediated TAE and TAE-derived starvation therapy. Instead, intratumoral injection manner was used to assess the influences of HMnO₂-mediated ROS therapy and DOX chemotherapy on subcutaneous HCC progression. Herein, Zein solution as cargo reservoir was introduced to store DOX and HMnO₂@DOX and enable the continuous DOX release and ROS birth. As expected, no significant inhibition against tumor progression is observed in Zein group compared to Control. By contrast, Zein-enabled continuous DOX release permits Zein@DOX to tremendously shrink tumor compared to Control through monitoring the tumor growth profiles (Fig. 3d). Identical results are obtained through appraising the volume of collected tumor at the end of experimental period (Fig. 3e and f), and DOX-contained groups receive the considerably-elevated inhibition impact on subcutaneous HCC progression. Notably, the considerably-inhibited tumor growth is found in Zein-HMnO₂ compared to Control, which implies that anti-tumor contribution of HMnO₂-mediated ROS therapy in Zein-enabled continuous HMnO₂ release is achieved. Intriguingly, the anti-tumor outcome of continuous DOX chemotherapy is strong enough to cover up the anti-tumor contribution of HMnO₂-mediated ROS therapy, as evidenced by the comparison between Zein-HMnO₂/Zein-HMnO₂@DOX and Zein@DOX (Fig. 3d–f). Notably, no fluctuation of mouse body weight is observed, unveiling the safety of this treatment process (Fig. 3g).

After pathological examinations, apoptosis is discerned to contribute to the HCC inhibition (Fig. 3h and Figure S22). Zein-enabled continuous HMnO₂ and DOX release in Zein-HMnO₂@DOX promotes increased apoptosis compared to Control. As the hypoxia marker, HIF-1α and its downstream gene VEGF play a synergistic role in counteracting antiproliferative and proapoptotic actions, thereby significantly promoting tumor growth and metastasis.^48,49 The intratumoral injection of Zein and Zein-DOX still worsens hypoxia microenvironment (HIF-1α upregulation),⁵⁰ and upregulates the metastasis driving factor (i.e., VEGF), while the intratumoral injection of Zein-HMnO₂ effectively alleviates hypoxia TME and further downregulates the VEGF. Astonishingly, although the anti-tumor contribution of HMnO₂-mediated ROS therapy in the Zein-HMnO₂@DOX group is covered up by DOX chemotherapy (Fig. 3d–f), O₂ release from HMnO₂-mediated CDT remits the hypoxic microenvironment and attenuates the risk of tumor metastasis, as validated by the downregulations of HIF-1α and VEGF compared to Zein@DOX group (Fig. 3h and Figure S23).

In vivo anti-tumor evaluation of Zein composites-unlocked TACE therapy and ROS therapy on rabbit VX2 liver tumor model

In such Zein-based TACE agents, TAE-mediated starvation therapy and DOX-mediated chemotherapy cooperatively constitute the TACE therapy, which will unite with HMnO₂-mediated chemotherapy to resist orthotopic HCC. To validate it, orthotopic VX2 liver tumor model that is the most widely used animal model in the study of TACE treatment against HCC was established; and then femoral artery puncture was performed for injecting such Zein-based TACE agent in a supra-selected vessel. The experiment procedures of such Zein-based TACE agents against HCC are depicted in Fig. 4a, and the orthotopic VX2 models by laparotomy and VX2 tumor mass implantation in the livers of rabbits are successfully established (Fig. 4b). Additionally, the catheter also successfully retains in the right femoral artery of the rabbit (Fig. 4c). The successful implementations of these technologies provide reliable and forceful guarantee for evaluating TACE-based anti-tumor consequences using such Zein-based TACE agents.

Fig. 4.

Evaluations on the in vivo anti-tumor efficacy using such Zein-HMnO₂@DOX TACE agent on orthotopic VX2 tumor models on rabbit. (a) Experimental flows of TACE on the rabbits. The tumor-bearing rabbits underwent enhanced CT scan after being transplanted VX2 tumor mass to confirm the successful modeling, then received TACE treatment, and followed up for 2 weeks to determine the efficacy of intervention. Created with Biorender.com. (b) Successful establishment of VX2 models by laparotomy and VX2 tumor mass implantation in the livers of rabbits. (c) Successful retention of the catheter in the right femoral artery of the rabbit. (d) Schematic diagram of the TACE treatment process. Herein, microcatheter was slowly introduced from the femoral artery into the target vessel, followed by injections of different embolic systems into the left hepatic artery to complete TACE treatment. Created with Biorender.com. (e) DSA images of rabbits VX2 models during TACE process with different treatments at pre-injection and post-injection. The red cycle regions represent the vessels supplying the tumor. (f) Representative enhanced CT images of rabbits in different groups before and after 14 days post-TACE treatment. The red cycle regions represent the VX2 tumor areas. (g) Representative digital images and macroscopic views of dissected VX2 tumors isolated from rabbit livers of each group after 14 days post-TACE treatment. (h, i) Tumor weights (h) and volumes (i) of dissected VX2 tumors isolated from the rabbit livers that experienced corresponding treatments in different groups. (j) Immunofluorescence images of tumor slices isolated from VX2 tumor-bearing rabbits that experienced corresponding treatments in different groups, including TUNEL, Ki67, and self-labeled DOX (Day14 are the same samples with the same treatment with DOX row of Figure S41) (Scale bar = 100 μm). (k) Representative H&E staining images of collected livers and tumors after 2-weeks post-TACE treatment in different groups (Scale bar = 50 μm). Data are expressed as mean ± SD (n = 6). One-way ANOVA followed by Tukey's multiple comparisons test were used to test the difference significance between the groups, and ∗∗∗∗p < 0.0001 vs Control group.

The detailed operation schematic of TACE therapy is shown in Fig. 4d. There are no statistical differences in body weight (Figure S24) and tumor volumes (Figure S25) between groups (p ＞ 0.05 (one-way ANOVA)), which can be used to reliably compare the efficacy of different treatment groups after TACE treatment. Digital subtraction angiography (DSA) is used to monitor the operation process, and the branches of left hepatic artery are increased and disordered, running at the edge of tumor before TACE operation, deciding that branches are small and unclear (Fig. 4e). In contrast, postoperative angiography shows blood flow stagnation and branch disappearance in left hepatic artery due to the dense aggregation embolism in tumor focus in lipiodol- or Zein-containing related groups, wherein lipiodol itself and iophorol served as the contrast agents in lipiodol-contained groups and Zein-contained groups, respectively. The precise and local TACE ensures that the trunk and branch in other hepatic arteries remain intact without reflux and deposition of lipiodol or Zein embolic agents.

Subsequently, contrast-enhanced CT scanning was enforced. Before TACE operation, enhanced tumor margins between nodules and normal liver parenchyma were clear, and only single HCC nodule is detected in left liver lobe without any metastases (Fig. 4f, Day 0). After 14 days post-TACE (Fig. 4f, Day 14), enhanced CT images of upper abdomen show that the VX2 tumor volume in the control group (saline injection) ascends significantly, and the density decreased significantly compared with that before operation. In contrast, Lipiodol@DOX and Lipiodol-HMnO₂@DOX groups cause dense lipiodol deposition in the left lobe of liver, and obvious liquefaction and necrosis areas are observed in some tumor areas. Although the tumor volume increases in comparison to that before operation, the incremental magnitude is inferior to that in Control, suggesting that lipiodol-based TACE therapy delayed HCC progression. Intriguingly, Zein-containing groups including Zein, Zein@DOX and Zein-HMnO₂@DOX groups only exhibit a slight increase in tumor volume, and smaller incremental magnitude than lipiodol-contained groups is obtained. After the post-surgery CT scans, livers and tumors are collected completely and photographed. The tumor tissues in all groups are soft in texture, and give birth to large areas of liquefaction and necrosis (Fig. 4g). Compared with the control group, the tumor weights (Fig. 4h) and volumes (Fig. 4g, i) in the other five groups significantly decline compared to those in the control group (p < 0.05 (one-way ANOVA followed by Tukey's multiple comparisons test)). Through comparing either the two corresponding groups, Zein@DOX vs Lipiodol@DOX or Zein-HMnO₂@DOX vs Lipiodol-HMnO₂@DOX, lower tumor weights and volumes are received in the Zein-based groups (Fig. 4g–i). This phenomenon reveals that Zein-based TACE agents outperform lipiodol-based TACE agents in inhibiting tumor progression, which can be attributed to enhanced stability, drug leakage inhibition and continuous drug release. The continuous DOX release and HMnO₂-mediated ROS therapy arm Zein-HMnO₂@DOX with the most powerful ability to inhibit HCC progression with the lowest tumor weight (Fig. 4i).

After pathological examinations, DOX is observed in tumor area in all DOX-containing groups (Fig. 4j and Figure S26), indicating the presence of chemotherapy. After that, terminal deoxynucleotidyl transferase mediated dUTP biotin nick end labeling (TUNEL) and immunofluorescence detection of Ki67 antigen were performed. Compared with control group, TUNEL positive rate (apoptosis) is significantly increased in the other five groups, while Ki67 positive rate (proliferation) is decreased, as indicated in Fig. 4j and Figure S26. Inspiringly, H&E staining indicates Zein-HMnO₂@DOX group induce more obvious necrosis areas that highlight cytoplasmic changes (e.g., lysis, sparse state, matrix disintegration) and nuclei damages (e.g., nuclei pyknosis and karyolysis) (Fig. 4k). These results indicate that Zein can occlude blood vessels after phase transition, unite with chemotherapy and DOX therapy to promote VX2 cell apoptosis and inhibit their proliferation, thus achieving starvation therapy for tumors and inhibiting rabbit VX2 liver tumors.

Lung metastasis inhibition in rabbit VX2 liver tumor model

To comprehensively understand the Zein-based embolic agents-unlocked starvation therapy, vascular density markers, i.e., CD31 and CD34,⁵⁰ were examined, and the considerably-decreased expression in the experimental groups are obtained, suggesting the successful vascular occlusion by Zein-based or lipiodol-based embolic agents (Fig. 5a and Figure S27). This result denotes that TAE and TACE-unlocked starvation treatment can inhibit CD31-labeled angiogenesis and reduce CD34-labeled vessel density and encourage blood platelets adhesion to generate thrombus and cut off nutrition and oxygen supply, which thereby combines with ROS therapy and chemotherapy to contribute to excellent therapeutic outcomes. To verify the vascular occlusion-arised thrombus, its biomarker, i.e., platelet membrane glycoprotein CD41 that navigated physiological hemostasis and pathologic thrombosis was traced.⁵¹ The expression of CD41 is significantly increased in the five experimental groups compared to Control (Fig. 5a and Figure S27). In particular, Zein-based embolic agents perform much better than lipiodol-based ones in occluding blood flow and initiating starvation therapy since the much higher CD41 expressions in Zein-based groups than lipiodol-based groups are found.

Fig. 5.

Exploration of the potential ability to inhibit lung metastasis using such composite TACE agents in the rabbit VX2 model. (a) Immunohistochemical images of VX2 tumor slices isolated from rabbit livers that experienced different treatments, including CD31, CD34, and CD41 to assess the efficacy of the starvation therapy by angiogenesis and platelet aggregation (Scale bar = 50 μm). (b) Immunohistochemical and immunofluorescence images of VX2 tumor slices isolated from rabbit livers that experienced different treatments, including HIF-1α, GLUT1 (Scale bar = 50 μm), and VEGF (Scale bar = 100 μm) to evaluate the influence of the Zein-HMnO₂@DOX on tumor oxidative stress, angiogenesis and glucose metabolism. (c) WB bands of various proteins collected from VX2 tumor implanted on tumor-bearing rabbits that experienced corresponding treatments in different treatment groups, including β-actin, VEGF, and GLUT1 to further evaluate the influence of the Zein-HMnO₂@DOX on tumor angiogenesis and glucose metabolism. (d) Representative enhanced CT images of rabbits in different groups before and 14 days after TACE treatment in both mediastinum windows and lung windows to evaluate the lung metastasis. (e) Representative images of histological analysis of lungs of VX2 tumor-bearing rabbits 14 days after TACE treatment by H&E staining for different groups (Scale bar = 1 mm).

Hypoxia inducible factor (HIF-1α) and its downstream including VEGF are the hallmarks of high cancer aggression and metastasis, and the activations of GLUT1 and VEGF in cancer cells can support increased glucose uptake/glycolysis and permit endothelial cell proliferation.52, 53, 54 It is not difficult to understand why high expressions of HIF-1α and VEGF spur metastasis.23, 24, 25, 26 A large number of clinical practices have witnessed the increase of HIF-1α and VEGF in the blood of patients after TACE, worsening prognosis and expediting metastasis and recurrence of HCC.^38,55 Additionally, glucose transporter protein 1 (GLUT1) that is a marker of poor prognosis in a variety of tumors is also identified as the downstream gene of HIF-1α that can be adaptively activated by overexpressed HIF-1α.^54,56,57 To address this concern, the introduction of HMnO₂ not only serves as carriers to load and enable continuous DOX release. More significantly, it is also designed to trigger ROS production to execute ROS therapy, and simultaneously release O₂ to alleviate or even reverse the TACE-aggravated hypoxia against the upregulations of HIF-1α, VEGF and GLUT1, which will benefit the metastasis and progression inhibition of HCC. As expected, Zein-based and lipiodol-based TAE or TACE indeed tremendously propel HIF-1α expression and aggravate hypoxia compared to Control (Fig. 5b and Figure S28). Inspiringly, HMnO₂-containing groups arrest the hypoxia escalation caused by TACE, as indicated by the two comparisons, i.e., Lipiodol@DOX vs Lipiodol-HMnO₂@DOX, and Zein@DOX vs Zein-HMnO₂@DOX.

Similar variation trends occur to the downstream genes of HIF-1α including VEGF and GLUT1, wherein TACE therapy using lipiodol-based and Zein-based embolic agents drives the upregulations of VEGF and GLUT1 (Fig. 5b and Figure S28) compared to Control group. However, HMnO₂-mediated O₂ release suppress their expressions to some extent via sequestering HIF-1α when comparing Lipiodol@DOX with Lipiodol-HMnO₂@DOX, or comparing Zein@DOX with Zein-HMnO₂@DOX. Identical conclusions are obtained using western blotting (WB) technology (Fig. 5c and Figure S29), determining that HMnO₂-mediated O₂ release via the CDT process can relieve hypoxia to repress the spontaneous HCC metastasis even though TACE therapy favors hypoxia aggravation and metastasis motivation. Appealingly, the much higher expressions of HIF-1α, VEGF and GLUT1 emerge in Zein-based groups than those in lipiodol-based groups, validating the stronger vascular occlusion ability and starvation therapy property of Zein TACE agents over commercial lipiodol-based TACE agents. This phenomenon also explains why Zein-based embolic agents did better jobs in inducing thrombus and inhibiting HCC growth than lipiodol-based embolic agents. Besides, WB analysis of Bcl2 proteins collected from VX2 tumor implanted on tumor-bearing rabbits that experienced corresponding treatments in different treatment groups was enforced to evaluate the influence of the Zein-HMnO₂@DOX on tumor apoptosis. The Zein-HMnO₂@DOX group brings about the powerful pro-apoptotic effects compared to Control group (p < 0.05 (one-way ANOVA followed by Tukey's multiple comparisons test)). (Figure S30).

To verify the abilities of HMnO₂-originated ROS release to attenuate TACE-aggravated hypoxia and restrain metastasis impetus, further investigation on the lung metastases from liver tumor was monitored since the spontaneous metastasis of VX2 tumor from liver to lung is readily observable. After 14 days post-TACE treatment, the rabbits still were raised for almost another 1 month so as to leave sufficient time to support HCC metastasis. Preoperative lung and postoperative pulmonary metastases in rabbits were evaluated by chest-enhanced CT. CT images show scattered multiple metastatic lesions in both lungs of rabbits in the Control, Lipiodol@DOX, Zein, and Zein@DOX groups, and the number of metastatic nodules is more than that in other two HMnO₂-containing groups (Fig. 5d). After collecting lung tissues for H&E staining, the embedded tissue profiles show obvious lung tissue destruction and multiple metastatic foci in HMnO₂-free groups (e.g., Control, Lipiodol@DOX, Zein and Zein@DOX) (Fig. 5e). In contrast, the number of nodules in both Lipiodol-HMnO₂@DOX group and Zein-HMnO₂@DOX group is much lower than that in above four HMnO₂-free groups (Figure S31). Based on the above experimental results, it is concluded that HMnO₂ in such Zein-HMnO₂@DOX TACE agents ameliorate TACE-aggravated hypoxia microenvironment (HIF-1α decline), and then down-regulated the expression of GLUT1 and VEGF to sequester the metastasis motivation, and eventually achieved the inhibitory effect on lung metastasis.

The glycolysis pathway, oxidative stress signaling pathway, and HIF-1α/VEGF signaling pathway in vitro were assayed and analyzed with HepG2, Hep3B and Huh7 cells. The changes in protein expressions of HIF-1α, VEGF and GLUT1 via WB, and ROS generation and mitochondrial membrane potential alterations in various treatment groups are consistent with results in the VX2 rabbit models (Figures S32–S39). The above outcomes in vitro suggest that HMnO₂ can ameliorate hypoxia by generating ROS, reduce the expression of HIF-1α, and thus down-regulate the expression of VEGF and GLUT1.

The biosafety of Zein-HMnO₂@DOX was assessed since excellent biosafety is the precondition of clinical translation. Herein, as the natural component, Zein has been accepted to be biocompatible. HMnO₂@DOX potentially poses threat to normal organs since HMnO₂@DOX suffer from the leakage after TACE. After 24 h post-intravenous injection, HMnO₂@DOX can be metabolized by liver and kidneys and excreted in urine since red fluorescence representing rhodamine B-labeled HMnO₂@DOX almost vanishes (Figure S40). This phenomenon denotes the safety of HMnO₂@DOX even though they escape from Zein embolic agents into blood circulation. Through comparing the concentrations of DOX in the tumor within 1 week and after 14 days (Figures S41 and S42), it is obtained that after passing through intrahepatic arterial injection, DOX could be released via Zein-HMnO₂ into the liver tumor region for wide distribution. The concentration of DOX in the tumor region decreases gradually by metabolism over time. Additionally, pathological examinations of normal organs show no obvious pathological injuries and severe inflammatory reaction (Figure S43). Furthermore, no statistical differences in blood routine, liver function, renal function and myocardial zymogram tests are obtained (Figure S44). By comparing 7 days and 14 days after TACE, there is no significant difference in the liver and kidney functions of rabbits among the treatment groups (Figure S45). These results validate the high short-term or long-term biological safety of Zein and HMnO₂.

RNA sequencing for resolving mechanism and signaling pathway

In order to further investigate the mechanism of Zein-HMnO₂@DOX inhibiting tumor in vivo, transcriptomics was used to explore the biological function and signal pathway. Under the screening conditions of p-adjust < 0.05 and Fold change ≥ 2.0, 3859 differential genes are obtained, among which 2345 genes were down-regulated and 1514 genes were up-regulated after Zein-HMnO₂@DOX treatment (Fig. 6a and b). GO and KEGG enrichment analysis were performed for further analysis of the differential genes. GO enrichment analysis shows that differential genes are mainly enriched in metabolism-related biological processes and molecular functions with a focus on the glucose homeostasis-related biological processes and glucose metabolism-associated molecular functions (Fig. 6c), e.g., such as Fructose-2,6-bisphosphate 2-phosphatase activity and l-lactate dehydrogenase. KEGG enrichment analysis shows that the differential genes are mainly enriched in HIF-1 signaling pathway, glycolysis/gluconeogenesis, Fructose and mannose metabolism, insulin signaling pathway and Alcoholism-related signaling pathway (Fig. 6d).

Fig. 6.

RNA sequencing of tumor tissues for analyzing the potential tumor inhibition pathways in Zein-HMnO₂@DOX. (a, b) Volcano map (a) and heat map (b) of tumor tissues after TACE in the Control group and Zein-HMnO₂@DOX group (p < 0.05, |fold change| ≥ 2). (c) GO analysis of differential gene expression profiles based on RNAseq after the Zein-HMnO₂@DOX treatment. (d) KEGG analysis of differential gene expression profiles based on RNAseq after the Zein-HMnO₂@DOX treatment, which is available for revealing the affected pathways by the above differential genes. (e) Percentage plot of different metabolism-related pathways in the total metabolism-related pathways screened by KEGG analysis. (f) Bubble plot of the enrichment scores in metabolic pathways related to glucose metabolism. (g) Chord diagrams of the differential potential genes in the carbohydrate metabolism pathway. Genes labeled by red rectangular frames are related to glycolysis or gluconeogenesis.

Through transcriptomic KEGG analysis, we further determine the percentages of different metabolism-related pathways in the total metabolism-related pathways, and find that gene differential expression in the glucose metabolism-related pathways accounted for the highest percentage (Fig. 6e). After that, metabolic pathways associated with glucose metabolism are concentrated to create bubble plots (Fig. 6f), wherein glycolysis/gluconeogenesis, fructose and mannose metabolisms are remarkably enriched. To further screen differential genes in the carbohydrate metabolism pathway, we depict chord diagrams and identify related genes that can encode corresponding proteins to regulate carbohydrate metabolism, e.g., ALDH9A1, ENO2, HKDC1, PGAM1, LOC100351519, LOC100358185, LOC103351328, LOC127486711 and LOC127488764 (Fig. 6g). Taken all above sequencing together, oxidative stress and carbohydrate metabolism (glucose metabolism) pathways are underscored in such a Zein-based Zein-HMnO₂@DOX-mediated TACE and ROS therapy, which contributed to the repressed progression and metastasis of HCC.

Discussion

HCC is a common malignant tumor in the world.¹ Current outcomes for HCC remain unsatisfactory due to late diagnosis and low response to conventional therapies.^3,5,6 Although patients receiving TACE could receive benefits,⁹ they still suffered from incomplete HCC necrosis after treatment.¹¹ More significantly, hypoxic TME exacerbation by starvation therapy and low-maintenance distribution of chemotherapeutic agents in the tumor territory during TACE treatment caused tumor recurrence and metastasis.^13,14 In an attempt to surmount these deficiencies, it is urgent to design and develop a new high-performance TACE embolization chemotherapy system.

Herein, we designed a ROS-enhanced Zein-based new TACE embolic agents applied to HCC therapy. Specifically, distinct components in this system fulfill their respective responsibilities and achieve each other. Firstly, DOX chemotherapy has a direct inhibitory effect on HCC. HMnO₂, as a vehicle of DOX, could effectively deliver DOX to HCC and enhance the combined treatment efficacy. Furthermore, HMnO₂ actively participates in the Fenton-like reaction, thus exhibiting its remarkable TME-responsive property and consequently curative efficacy. HMnO₂ is an ideal TME-responsive nanomaterial with GSH-consuming capacity, which catalytically produces ROS and oxygen from hydrogen ions (H⁺)/H₂O₂ to further remodel the hypoxic TME,³⁸ consequently boosting the therapeutic efficiency of TACE. In this regard, HMnO₂ has also been evidenced to be engaged in the oxidative stress pathway through ROS birth, which participated in the expression of GLUT1 and VEGF. The high expression of VEGF and GLUT1 has been proven to correlate with tumor metastasis, downstreaming HIF-1α.^56,57 All above design principles have not been found in current microspheres-based embolic agents.

Such Zein-based embolic agents featured excellent mechanical strength and stably withstand blood pressure squeezing without deformation, which avoided the washing off by blood flow. Moreover, the elevated adhesion and cross-linking of Zein with HMnO₂ carriers controlled the release of HMnO₂@DOX, and especially united with HMnO₂ decomposition-determined delayed DOX release in response to acidic tumor microenvironment to enable continuous DOX release for chemotherapy. Zein gelation once touching water successfully occluded intratumoral blood vessels, elicited thrombus, cut off nutrition and oxygen supply and eventually realized starvation therapy, which combined therapy chemotherapy to constitute Zein-based TACE. Zein-based TACE outperformed clinical lipiodol-based TACE in delaying orthotopic HCC progression. Although TACE aggravated hypoxia represented by HIF-1α upregulation and encouraged metastasis represented by VEGF and GLUT1 overexpression, the introduced HMnO₂ not only produced ROS via the CDT process to magnify TACE-based anti-tumor outcome, but also released O₂ to attenuate TACE-aggravated hypoxia and metastasis motivation, which repressed the metastasis of orthotopic rabbit VX2 tumor to lung and reduced the number of metastatic nodules. Moreover, sequencing analysis revealed oxidative stress and glucose metabolism interference by such biocompatible Zein-based embolic agents-unlocked TACE and ROS therapy is responsible for the excellent inhibitory effect on the progression and metastasis of HCC. More significantly, Zein-HMnO₂@DOX accommodates high tissue adhesion with non-spherical structure against blood flow-rushing, which reduces the occurrence rates of post-embolization syndrome (e.g., off-target vascular embolization in normal organs). As well, compared to microspheres, such TACE agents are anticipated to avoid ischemia-induced neoangiogenesis,⁵⁸ and temporarily protect normal liver parenchyma during TACE.⁵⁹

This study has some limitations. Firstly, in this study, we only investigated the role of hollow MnO₂ nanoparticles as ROS generators and have not yet explored the effects of other morphologies of MnO₂ or other ROS generators. In future research, we could further explore the possibilities of alternative approaches. Secondly, in addition to activating the HIF-1α-VEGF/GLUT1 signaling pathway, Zein-HMnO₂@DOX may also be involved in other signaling pathways that remain to be explored. Thirdly, whether other signaling pathways or targets are involved in the effects of Zein-HMnO₂@DOX against HCC as well as their complex interrelated mechanisms remain to be explored.

Collectively, the inhibitory effect of Zein-HMnO₂@DOX in TACE on HCC proliferation and metastasis is promising, and Zein-HMnO₂@DOX provides new perspectives and possibilities to ameliorate the inferior therapeutic efficacy of TACE caused by the hypoxic microenvironment. In the future development of TACE embolic agents, it is probable that an increasing variety of microenvironment-responding nanoplatforms would be introduced into the embolization system. Moreover, we believe that the efficacy of Zein-HMnO₂@DOX in TACE requires further preclinical investigations.

Contributors

K. Zhang designed the project, and conceived and proposed the novelty and paper structure. L. Song, C. Zhu, Q. Shi, Y. Xia, X. Liang, W. Qin, T. Ye and B. Yang performed the experiments. K. Zhang, L. Song and C. Zhu analyzed the data. K. Zhang and L. Song wrote this manuscript, and K. Zhang revised and organized this manuscript. K. Zhang, J. Xia and X. Cao supported this project. K. Zhang and J. Xia have verified the underlying data. K. Zhang supervised the project and all authors commented on this manuscript. All authors read and approved the final version of the manuscript.

Data sharing statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. For RNA Sequencing analysis, original data have been deposited into the Sequence Read Archive database under the accession number PRJNA1168652 on National Center for Biotechnology Information.

Declaration of interests

The authors declare no conflict of interest.