Autophagy-amplifying nanoparticles evoke immunogenic cell death combined with anti-PD-1/PD-L1 for residual tumors immunotherapy after RFA

Incomplete radiofrequency ablation (IRFA) triggers mild protective autophagy in residual tumor cells and results in an immunosuppressive microenvironment. This accelerates the recurrence of residual tumors and causes resistance to anti-PD-1/PDL1 therapy, which bringing a great clinical challenge in residual tumors immunotherapy. Mild autophagy activation can promote cancer cell survival while further amplification of autophagy contributes to immunogenic cell death (ICD). To this regard, we constructed active targeting zeolitic imidazolate framework-8 (ZIF-8) nanoparticles (NPs) loaded with STF62247 or both STF62247 and BMS202, namely STF62247@ZIF-8/PEG-FA (SZP) or STF62247-BMS202@ZIF-8/PEG-FA (SBZP) NPs. We found that SZP NPs inhibited proliferation and stimulated apoptosis of residual tumor cells exposed to sublethal heat stress in an autophagy-dependent manner. Further results discovered that SZP NPs could amplify autophagy in residual tumor cells and evoke their ICD, which dramatically boosted the maturation of dendritic cells (DCs). Through vaccination experiments, we found for the first time that vaccination with heat + SZP treatment could efficiently suppress the growth of new tumors and establish long-term immunological memory. Furthermore, SBZP NPs could remarkably promote the ICD of residual tumor cells, obviously activate the anti-tumor immune microenvironment, and significantly inhibit the growth of residual tumors. Thus, amplified autophagy coupled with anti-PD-1/PDL1 therapy is potentially a novel strategy for treating residual tumors after IRFA. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s12951-023-02067-y.


Introduction
Liver cancer is the sixth most common cancer and the third leading cause of cancer-related death worldwide [1,2].Hepatocellular carcinoma (HCC) is the most typical primary malignant liver cancer.Radiofrequency ablation (RFA) is one of the main curative therapies in treating HCC for its safety, shorter hospital stay, and fewer complications [3,4].However, factors such as complex spatial structure, indistinct border of tumors, and failure to access the ablative margin may lead to incomplete radiofrequency ablation (IRFA), resulting in residual tumors and even accelerating their progression [5,6].Research has shown that patients demonstrate an overall recurrence rate as high as 63.5% five years after RFA, with a poorer long-term prognosis than those undergoing surgical resection [7].How to effectively suppress residual tumors is a major clinical challenge.
RFA releases the tumor debris whilst the tumor burden is mitigated, which stimulates the host's anti-tumor immune response [8][9][10].However, the immunogenicity elicited by RFA is too weak to activate sustained antitumor immune response and suppress the recurrence of residual tumors [9].In recent years, RFA combined with immunotherapy has been employed to treat HCC in various clinical studies, demonstrating significant clinical value [11].Immune checkpoint blockade (ICB) is one of the most promising approaches to activating antitumor immunity.It recovers the once-suppressed ability of T cells to recognize tumors by blocking PD-1 from binding with PD-L1, which then activates systemic immunity [12].Nevertheless, the latest research has revealed that increased tumor-infiltrating myeloid cells after IRFA strengthened the immunosuppression microenvironment of residual tumors [13,14], which severely reduced the sensitivity of residual tumors to ICB therapy [15][16][17].
Autophagy plays a paradoxical role in anti-tumor immune response [18,19].Low-level autophagy is cytoprotective and conducive to the survival of tumor cells with restriction on the release of immunogenicity.It leads to the formation of an immunosuppressive microenvironment, favoring the survival of regulatory T cells (Tregs) and the polarization of tumor-associated macrophages (TAMs) [20,21].Contrastively, highlevel autophagy is able to promote cell death in tumor cells and release damage-associated molecular patterns (DAMP) from dead cells, which can remodel the tumor microenvironment (TME) and activate anti-tumor immunity.Studies showed [22][23][24][25] that sublethally heated HCC cells exhibit a low level of autophagy activation after IRFA, serving as self-protection that would induce immune tolerance with resistance to immunotherapy.Multiple preclinical research has confirmed that sufficient intracellular autophagy activation can induce autophagy-dependent cell death (ACD) [26], also known as Type II programmed cell death.It was also reported that enhanced activation of autophagy can release immunogenicity to strengthen anti-tumor immunity while inducing ACD [27].Moreover, sublethally heated residual tumors are pathologically hyper-sensitive to autophagy compared to normal cells [28].Therefore, in contrast to inhibiting protective autophagy, amplified activation of autophagy can not only eliminate low-level autophagy's protection on HCC cells after IRFA [29], but also evoke immunogenic cell death (ICD) in residual tumor cells and reverse the immunosuppressive microenvironment [30,31].Amplified activation of autophagy combined with ICB may have synergistic effects and improve efficacy against residual tumors after IRFA.
We previously reported that the extensive angiogenesis after IRFA led to an augmented permeability and retention (EPR) effect and contributed to the enrichment of nanodrug in residual tumors [32].ZIF-8 is a metal-organic framework synthesized by Zn 2+ and 2-MeIM with good biocompatibility and perfect acidinduced degradability.It has the advantages of variable pores, easily-modified structure, and high drug-loading capacity [33][34][35].Aiming to tackle the clinical challenge of insufficient immunogenicity and decreased microenvironmental sensitivity to immunotherapy after IRFA [14], autophagy inducer STF-62247 (STF) and small molecule inhibitor of the PD-1/PD-L1 interaction BMS202 (BMS) were loaded into the ZIF-8 framework via one-pot synthesis to prepare STF62247-BMS202@ZIF-8/PEG-FA (SBZP) NPs.STF is a small molecule that induces autophagy, which could powerfully transformed protective autophagy into ICD and remodeled the immune microenvironment [36][37][38]; BMS is a small molecule inhibitor of the PD-1/PD-L1 interaction which could bolstered the ability of T cells to fight against tumors [39,40].Overall, we report a novel strategy of amplified activation of autophagy in combination with ICB for residual cancer treatment after IRFA, which may have immense potential in the future.
The preparation of STF@ZIF-8/PEG-FA (SZP), BMS@ZIF-8/ PEG-FA(BZP), and STF-BMS@ZIF-8/PEG-FA (SBZP) Solution A was prepared by dissolving 300 mg of Zn (NO3)2•6H2O in 8 mL of ethanol, while solution B by dissolving 320 mg of 2-MeIM in 8 mL of ethanol.4 mL of STF (5 mg/mL) and/or 4 mL BMS (1 mg/mL) were added dropwise into solution A respectively (stirred at 1200 rpm for 5 min, room temperature).Then, the above solutions were added dropwise to 8 mL of solution B and stirred at 1200 rpm for 15 min at room temperature.The NPs were purified by centrifugation (10,000 rpm) and washed with ethanol 3 times.After vacuum drying, SZ, BZ, and SBZ came into being.50 mg of SZ, BZ, and SBZ were added to 5 mL of deionized water containing 50 mg of PEG-FA, respectively, then ultrasonically oscillated until they were evenly distributed, and stirred at 1200 rpm for 24 h.The NPs were purified by centrifugation (10,000 rpm) and washed with ethanol 3 times, and finally SZP, BZP, and SBZP were prepared after vacuum drying.By UV spectrophotometer, STF showed a characteristic absorption peak near 295 nm, BMS labeled with FITC, FITC-BMS showed a characteristic absorption peak near 512 nm.The drug loading content (DLC) and drug loading efficiency (DLE) of SBZP NPs were calculated according to the previously reported formulas [23].The DLC and DLE of STF in SBZP were 19.75% and 86.93%, while BMS was 11.52% and 86.47%.

Structural characterization of NPs
The morphology and size of NPs were examined using JEM-2100F transmission electron microscope under an accelerating voltage of 200 kV.X-ray diffractometer (XRD, D-MAX 2200 VPC Diffractometer, Rigaku, Japan) was employed to measure the crystal structure.The zeta potential of the NPs was explored by Zetasizer Nano ZSE (Malvern, UK).UV-Vis spectrometer (LAMBDA 365, PerkinElmer, USA) recorded the ultraviolet-visible spectroscopy of NPs.The functional groups were observed by Fourier-transform infrared spectroscopy (FTIR, Bruker, Billerica, MA, USA).The chemical composition was identified via X-ray Photoelectron Spectroscopy (XPS, ESCALAB 250, Thermo Fisher, US).Thermogravimetric analyzer (Pyris, PerkinElmer) measured the mass as the temperature changed from 25℃ to 700℃.The procedure was performed in nitrogen atmosphere and at a constant heating rate of 10℃/min.

In-vitro experiment on drug release
5 mg of SBZP were dispersed in 20 mL of PBS (pH 5.5 and pH 7.5).UV-Vis absorbance spectrometer was measured at 295 nm.Standard curve of STF was made based on different concentrations in different pH solutions.Then, the STF concentration of SBZP in the sample solutions was extrapolated from the STF standard curve.To measure the BMS content in SBZP, BMS was labeled with FITC.The same method was applied to measure the BMS concentration of SBZP at 512 nm with a fluorescence spectrophotometer.

Sublethally heated cell model in vitro
Culture dishes sealed with films were heated at different temperatures (43,45,47, and 49 °C) for 15 min and different periods of time (5, 10, 15, 30, and 60 min) at 47 °C [41].Upon being heated, they were transferred to the incubator for 24 h recovery at 37 °C and 5% CO 2 .Cells cultured at 37 °C were selected as the control group.

IRFA subcutaneous tumor model
H22-Luc cells (5 × 10 5 per mouse) were subcutaneously into the right thigh of female BALB/c mice (aged 6 weeks, Guangzhou Yancheng Biotechnology, China).When the tumors grew to 200-400mm 3 , the mice were anesthetized and placed on the VIVA grounding pad.RF electrodes were inserted into the tumor and one-third of it was pierced under ultrasound guidance.IRFA was performed with a power output of 5W for 20 s.D-luciferin potassium salt (PerkinElmer, USA) was intraperitoneally injected for bioluminescent imaging in H22 tumor-bearing mice.

Cell viability and proliferation assays in vitro
HCC cells were seeded in 96-well plates (5,000 cells/ well), incubated with various concentrations of drugs (STF and SZP) and different reagents (PBS, ZIF-8, STF, SZP, SZP + 3MA), and subsequently heated following the method described in Section "Sublethally heated cell model in vitro".After 24 h or 48 h rewarming, the cells were cultured with 10% CCK8 (DOJINDO, Kumamoto, Japan) solutions.A multimode reader (Synergy HTX, Bio-Rad, USA) was used to record the absorbance value at 450 nm, and then, the curve of cell growth was drafted with cell viability and IC50 value calculated.

Cell apoptosis analysis and double staining using live/dead cell viability assay kit
Sublethally heated cells treated by different drugs were stained with Annexin V-FITC Apoptosis Detection Kit (BD Biosciences, USA) and Calcein-AM/PI Dual Staining Kit (Solarbio, Beijing, China).Apoptosis and cell death rates were measured using Cyto-FLEX LX Flow Cytometer (Beckman Coulter, USA).Collected data were analyzed with FlowJo 7.6.5 (Tree Star, USA).The fluorescent signals of live and dead cells were observed with an inverted fluorescence microscope (CLSM IX73, Olympus, USA).

Cellular drug uptake
SMMC7721 cells were seeded in 24-well plates (5 × 10 4 cells/well) and incubated with FITC@ZIF-8 NPs for 2 h, 4 h, 8 h, 12 h, and 24 h, respectively.Subsequently, the cellular uptake ratio and intracellular distribution of ZIF-8 NPs were evaluated by confocal laser scanning microscopy (CLSM) and flow cytometry.

The accumulation and distribution of NPs in tumors
ICG@ZIF-8/PEG-FA NPs and free ICG were injected into the H22 tumor-bearing mice via the tail vein, respectively.Fluorescent images were captured at different time points using IVIS imaging system (IVIS Lumina, Perki-nElmer, USA).The tumor-bearing mice were euthanized to obtain tumors and important organs for fluorescent imaging after 24 h.At the same time, a few tumor tissues and organs were cauterized and nitrified, and subsequently the concentration of [Zn 2+ ] was determined by 5800 ICP-OES (Agilent, Japan).

Transmission electron microscope (TEM)
Sublethally heated cells separately treated with PBS and SZP were collected and fixed in 2.5% glutaraldehyde for 2-4 h.They were rinsed by PBS 3 times, fixed in 1% osmium tetroxide at 4 °C for 2 h, dehydrated with gradient ethanol, transitioned in epoxypropane, and embedded in 812 resin (Spi-Chem) after gradient infiltration.Ultra-thin sections were cut using ultramicrotome (Leica UC7, Vienna, Austria) and double-stained by uranyl acetate and lead citrate.The ultrastructure of tissue was observed by TEM (HT7700, Japan).

In vitro detection of ICD
First, HCC cells and ATG5-knockdown HCC cells were seeded into TC-treated cell slides and treated with drugs and heat stress.Next, the prepared cell slides were incubated with primary antibodies (anti-Calreticulin 1:100, anti-HMGB1 1:100) at 4 °C overnight, which then were incubated with fluorescent secondary antibodies (Alexa Fluor ® 488 1:500, Alexa Fluor ® 555 1:500) and DAPI stain for 1 h.Lastly, the expression of CRT and HMGB1 in cells was examined by CLSM.Additionally, the processed cells and supernatant were obtained.The ATP concentration in the cells and supernatant were measured by an ATP assay kit (Beyotime, China), and the concentration of HMGB1 in the supernatant was examined using an ELISA kit (Solarbio, Beijing).Bone marrow-derived dendritic cells (BMDCs) of mice were cultured with sublethally-heated HCC cells treated with different groups (PBS, ZIF-8, STF, SZP, SZP + shATG5) for 48 h.The collected DCs were stained by FITC anti-mouse CD11c, PE anti-mouse CD86, and APC anti-mouse CD80 antibodies (Biolegend), and subsequently measured via flow cytometry.ELISA kit was employed to identify the concentration of IL-12p70 (Lianke Bio, China EK212/3), IL1-β (Thermo Fisher Scientific 88-7013-88), IL-6 (Thermo Fisher Scientific 88-7064-88), and TNFa (Thermo Fisher Scientific 88-7324-88) in the supernatant of DCs.

In vivo detection of ICD
H22 cells with ATG5 knocked-down and H22 cells were subcutaneously inoculated into the right thigh of BALB/c mice, respectively.Subsequently, IRFA mice model was established following the method described in Section "IRFA subcutaneous tumor model".After four times tail vein injections of NPs, tumor tissues and tumordraining lymph nodes were collected.Next, tumor tissue slides were incubated with primary antibodies (anti-CRT, anti-HMGB1) at 4 °C overnight and observed through CLSM; lymph nodes were examined via flow cytometry, following the method described in Section "In vitro detection of ICD".

In vivo vaccine inoculation trials
Firstly, a whole tumor cell vaccine was prepared.H22 cells were treated in three ways: heat, heat with STF (10 μM), and heat with SZP (10 μM), all of which were performed at 47 °C for 15 min.The cells were harvested after 24 h and trypan blue ensured cell death rate was higher than 80%.Next, 3 × 10 4 dying cells were injected into the left back of BALB/c mice twice (n = 10), 7 d apart.Then, 14 days after the first injection, 5 × 10 5 of H22 cells were subcutaneously injected into the contralateral thigh.On day 15, the spleen of the mice was collected and stained with APC/Fire ™ 750 anti-mouse CD8a, PE anti-mouse CD44, and APC anti-mouse CD62L antibodies, which then were analyzed by flow cytometry.Tumors were measured with a caliper every 2-3 days when they grew visible.The mice were euthanized by inhaling CO 2 on day 24.Meanwhile, ELISA kit was employed to measure the concentration of IFN-γ (Lianke Bio, China EK212/3) and TNF-α in the serum of mice.

In vivo anti-tumor treatment and immune activation
Antitumor studies were carried out in H22 IRFA subcutaneous tumor model.As shown in Fig. 8A, 25 mice were categorized into five groups: control, STF, BZP, SZP, and SBZP.All drugs were injected 2 days before IRFA six times with interval of 2 days (STF: 8 mg/Kg; BMS: 4 mg/Kg).Since the 10th day after IRFA, the tumors were measured by caliper every 2 days, and the mice were weighed.On day 24, the tumor tissues were harvested after the final experiment.At the same time, ELISA kit was employed to measure the concentration of IFN-γ, TGF-β (Lianke Bio EK981-96), IL-6, and TNF-α in the serum of mice from different groups.

Immunocytochemistry and immunofluorescence staining
Prepared tumor tissue sections were stained with Ki-67, TUNEL (One Step TUNEL Apoptosis Assay Kit, Beyotime), LC3B, P62, CD4, and CD8, respectively.Antibody concentrations used were all 1:100 for primary antibodies and 1:500 for secondary antibodies.Next, immunofluorescence and immunohistochemical analysis were carried out.

Statistical analysis
Measurement data were expressed as mean ± standard deviation.Kaplan-Meier survival curves were compared using the log-rank Mantel-Cox test.Data were analyzed by GraphPad Prism 7.0 (GraphPad Software, San Diego, CA).P < 0.05 was considered significant.

IRFA promotes autophagy in HCC cells and the emergence of an immunosuppressive microenvironment
Previous research has revealed that IRFA can induce HCC cells' resistance to heat treatment through protective autophagy [24,42].To explore the effects of thermal ablation on the autophagy of HCC cells, we assessed in vitro the autophagy profile of SMMC7721 and Huh7 cells exposed to heat stress at different temperatures (43,45,47, and 49 °C) for 15 min and at 47 °C for various periods of time (5, 10, 15, 30, and 60 min).The western blot results demonstrated that the autophagy level of SMMC7721 and Huh7 cells was positively correlated to temperature and the duration of heat treatment, featured by an increased conversion of LC3-I to LC3-II and a decreased expression of P62.Autophagy activation was most prominent in HCC cells heated at 47 °C for 15 min (Fig. 1A, B). "Autophagic phenotype" of "vacuoles" in the cytoplasm was found in sublethally heated SMMC7721 and Huh7 cells via microscope (Additional file 1: Fig. S1).Therefore, the condition of HCC cells being heated at 47 °C for 15 min was determined as the optimal method to activate autophagy in subsequent experiments, which was consistent with other results [43,44].Then we transfected transiently SMMC7721 and Huh7 cells with mRFP-GFP-LC3 plasmid to observe the changes in autophagic flux after sublethal heat stress.When autophagosomes fused with lysosomes, GFP in mRFP-GFP-LC3 was quenched in the acidic environment of autolysosome, with the red fluorescence emitted from mRFP left.Thus, autolysosomes appeared as red dots (only mRFP) while autophagosomes as yellow ones (merged of mRFP and GFP).CLSM showed a significantly increased number of red and yellow spots after sublethal heat stress (Fig. 1C and Additional file 1: Fig. S8), with red spots outnumbering yellow ones.This suggested that sublethal heat stress could induce more autophagosomes in HCC cells without impeding autophagic flux.Furthermore, we examined how IRFA influenced the autophagy levels in H22-bearing mice via western blot (Fig. 1E).The results showed that the autophagy level was prominently elevated 3 days and 8 days after IRFA compared with that before IRFA.Immunofluorescent staining and Immunohistochemical results also exhibited the same changes (Fig. 1F and Additional file 1: Fig. S2).To further investigate the changes in the tumor immune microenvironment after IRFA, we analyzed the proportion of tumor-infiltrating CD4 + and CD8 + T cells by flow cytometry (Fig. 1G), immunohistochemistry (Additional file 1: Fig. S5), and immunofluorescence (Fig. 1F).It turned out that the infiltration of CD4 + and CD8 + T cells in the tumor tissue increased 3 days after IRFA, but dropped 8 days after IRFA (Fig. 1H).At the same time, the tumor-infiltrating CD11C + DCs evaluated by flow cytometry suggested trends consistent with that of T cells.On top of that, we discovered that 3 and 8 days after IRFA, the infiltration rate of CD11b + F4/80 + TAMs continuously grew to 1.92 ± 0.50 and 4.12 ± 0.64 folds of pre-IRFA levels, respectively.Similar changes could also be identified from the ratio of immune cells in the spleen of the mice (Additional file 1: Fig. S3, S4).According to ELISA, the levels of tumor-killing cytokines (TNF-α, IFN-γ) shot up on the third day after IRFA, but plummeted 8 days after IRFA; meanwhile, the levels of tumor-promoting cytokines (TGF-β, IL-6) rose progressively (Fig. 1I).These results suggested that the immune response was mildly activated 3 days after IRFA, which could be linked to inflammatory cell infiltration caused by IRFA.Yet such immune activation was transient, which was followed by the emergence of a prominently suppressive immune microenvironment 8 days after IRFA.

Cellular uptake, biodistribution and cytotoxicity of ZIF-8/ PEG-FA NPs
To explore HCC cells' uptake of the NPs, we successfully synthesized FITC-labeled ZIF-8.Immunofluorescence demonstrated that SMMC7721 cells could rapidly take up FITC@ZIF-8 NPs and the fluorescence intensity gradually magnified (Fig. 3A, B).Flow cytometry analysis demonstrated the increased accretion of SMMC7721 cells that had taken up FITC-labeled NPs over time (Fig. 3C) and 100% cellular uptake of NPs was detected 12 h after incubation.
Then we synthesized active targeting ZIF-8/PEG-FA NPs labeled by ICG and identified their biodistribution in H22-bearing mice compared with free ICG.The IVIS images (Fig. 3D) illustrated that free ICG quickly gathered in the liver and tumor tissues.Twelve to twenty-four hours later, the fluorescent intensity of free ICG reduced, and it was excreted from the liver rapidly.In comparison, ICG@ZIF-8@PEG-FA NPs accumulated in the tumor tissues longer.Their fluorescence signal in tumors peaked at 4 h and lasted at least 24 h (Fig. 3E).The quantitative fluorescence trend chart suggested that ICG@ZIF-8@PEG-FA NPs may possess better cycling stability and therefore increase drug accumulation in local tumors.Additionally, according to the fluorescence imaging of major organs and tumors ex vivo 24 h after injection, ICG@ZIF-8@ PEG-FA NPs mainly accumulated in tumors and fractionally in the livers (Fig. 3F), while no obvious fluorescence signals were spotted in other organs (Fig. 3H).At the same time, the examination of Zn 2+ accumulation in the tissues through ICP-OE also revealed consistent results (Fig. 3I).That ICG@ZIF-8@PEG-FA NPs could effectively accumulate in residual tumor tissues after IRFA may be ascribed to the enhanced EPR effect we previously reported [32] and the active targeting modification.Meanwhile, the CCK8 experiment discovered no apparent cytotoxicity of serial concentration of ZIF-8 NPs.The cell viability was above 90% even at a concentration of 250 µm (Fig. 3J).

SZP enhances autophagy of sublethally heated HCC cells
The process of autophagy can be divided into three relatively independent steps (Fig. 4A): the formation of autophagosome, the fusion of autophagosome and lysosome, namely autolysosome, and the degradation.Since sublethal heat stress induced protective autophagy of HCC cells, we further amplified autophagy via SZP which was loaded with STF to induce ACD [43][44][45] in residual tumor cells.To confirm the effect, we incubated sublethally heated SMMC7721 and Huh7 cells with SZP for 4 h.TEM detection revealed that SZP led to a marked increase of autophagosome-like double membrane vesicles in cytoplasm (Fig. 4B).Since autophagic flux is a dynamic process, we further verified whether the enhanced autophagy was attributed to the increased formation of autophagosome or the inhibition of its fusion with lysosome.As we know, 3-methyladenine (3-MA) could inhibit the formation of autophagosomes while Chloroquine (CQ) could inhibit the fusion of autolysosome, both resulting in autophagy inhibition.Western blot demonstrated that compared to the control group, SZP significantly promoted the transformation of LC3B-I to LC3B-II in SMMC7721 and Huh7 cells.It demonstrated that SZP further amplified autophagy induced by sublethal heat.At the same time, SZP failed to elicit such remarkable changes when the cells were pretreated with 3-MA (Fig. 4C).However, the expression of both LC3B and P62 was further elevated when the autophagic flux of SZP-treated cells was blocked by CQ (Fig. 4D).These results suggested that the SZP-induced increase in autophagic flux resulted from the incremented formation of autophagosomes.At the same time, we transfected cells with mRFP-GFP-LC3 plasmid to trace the changes in autophagic flux, and the same conclusion was reached.Immunofluorescent results visually revealed that after cells were treated with SZP, the number of autophagosome and autolysosomes multiplied more than 20 times compared with that of the control group (Fig. 4E, F).When CQ was added, the number of autolysosomes was reduced because of inhibited fusion of autolysosome, but autophagosomes continued to increase (Fig. 4G).In summary, the enhanced autophagy induced by SZP NPs in sublethally heated cells was ascribed to the formation of autophagosomes instead of the inhibition of autolysosome degradation.This ultimately led to the potent activation of autophagic flux, which was necessary for inducing ACD [48].

SZP mediates ACD
Next, we investigated whether SZP could induce ACD in sublethally heated HCC cells.HCC cells and sublethally heated cells were treated with different concentrations of STF.According to the cell viability curve, the 24 h half maximal inhibitory concentration (IC50) of STF at 37 °C was 2.41 (SMMC7721) and 2.12 times (Huh7) of that at 47 °C (Fig. 5A).Similar results were acquired for the comparison of 48 h IC50: STF's IC50 at 37 °C was 2.78 (SMMC7721) and 2.27 times (Huh7) of that at 47 °C (Additional file 1: Fig. S6).The results suggested that STF substantially enlarged the effect of sublethal heat stress on killing HCC cells.We also compared the impact of STF and SZP on the cell viability of sublethally heated HCC cells.It turned out that the 24 h IC50 of STF in SMMC7721 and Huh7 cells was 1.92 and 1.83 times that of SZP (Fig. 5B), and the comparison of 48 h IC50 yielded similar results (Additional file 1: Fig. S7), indicating that SZP outperformed STF in suppressing the survival of sublethally heated HCC cells.In summary, sublethal heat stress activated low-level autophagy and ACD could be induced by further amplified autophagy activation.Importantly, sublethally heated HCC cells exhibited higher sensitivity to this strategy than normal ones.The SZP NPs we prepared outmatched STF in inducing ACD.
We continued to verify whether the SZP-mediated cell death in residual tumors relied on autophagy.SMMC7721 and Huh7 cells were incubated with ZIF-8, STF, SZP, and SZP + 3MA for 4 h, respectively.Then, they were exposed to sublethal heat stress at 47 °C.After 24 h, CCK8 assay (Fig. 5D), live/dead cell staining (Fig. 5C), apoptosis experiment (Fig. 5E, G), and colony formation assay (Additional file 1: Fig. S11, S12) were carried out.The results pointed out the ability of SZP to inhibit cell proliferation, promote cell death, and induce cell apoptosis.We also found that the killing effect induced by SZP on cells could be obviously attenuated by 3-MA pretreatment.It has been widely recognized that mTOR signaling pathway participated in the suppression of autophagy.In this research, western blot uncovered that both STF and SZP could significantly suppress the expression of p-mTOR protein of sublethally heated SMMC7721 and Huh7 cells in an autophagy-dependent manner, which is proved by the co-treatment of SZP and 3-MA (Fig. 5H).In conclusion, SZP NPs suppressed the proliferation of tumor cells as well as enhanced ACD by inhibiting mTOR signaling pathway.

SZP induces autophagy-dependent ICD in vitro
Previous studies have suggested that autophagy plays a pivotal role in ICD for its involvement in the release of DAMP from dead cells [49].Calreticulin (CRT) exposure, HMGB1 release, and ATP secretion signified the occurrence of ICD (Fig. 6A).Therefore, we examined whether SZP could evoke the ICD of sublethally heated HCC cells via autophagy.ATG5 is recognized as a key protein in autophagy, and the knockdown of ATG5 causes STF unable to induce autophagy [43].To analyze the role of SZPinduced autophagy in ICD, we first established cell lines (SMMC7721, Huh7, and H22 cells) with the knockdown of ATG5.Western blot confirmed that the expression of ATG5 was diminished in the cells transfected with ATG5 shRNA (Fig. 6B).CRT exposure was an initial event in ICD cascade that served as an activator of immune cells, HMGB1 mainly acted as an immunogenic mediator in the tumor, and extracellular release of ATP triggered DCs activation.Firstly, we used immunofluorescence staining (Fig. 6F and Additional file 2: Fig. S14) and flow cytometry (Additional file 2: Fig. S13) to explore the CRT exposure levels in sublethally heated SMMC7721 and Huh7 cells that were treated with SZP.The results demonstrated that SZP significantly reinforced CRT exposure.HMGB1 was a highly abundant chromatin-binding protein.According to the immunofluorescent results, HMGB1, stimulated by SZP-induced autophagy, translocated from nucleus to cytosol (Fig. 6F).ELISA (Fig. 6E) and Western blot (Fig. 6C) were employed to further quantify the levels of HMGB1 released to the supernatant of sublethally heated cells treated with SZP.The results indicated that SZP prominently potentiated the release of HMGB1.Next, the intracellular and extracellular concentrations of ATP were evaluated.We noticed that ATP secretion rose by 3.7-fold after sublethally heated cells underwent treatment of SZP, while the intracellular concentration of ATP declined visibly (Fig. 6D).However, a considerable drop in the expression and release of immunogenic molecules was detected after ATG5 knockdown in the above situation.It was concluded that SZP NPs could evoke the ICD of sublethally heated HCC cells in an autophagy-dependent manner.
The DAMP generated from ICD held the key to the maturation of DCs.Accordingly, we evaluated the effect of SZP in boosting the maturation of DCs.BMDCs were co-cultured with sublethally heated SMMC7721 and Huh7 cells treated with different groups (Fig. 6G).The results of flow cytometry revealed that the ratio of mature DCs (CD11c + CD80 + CD86 + ) significantly increased in the group treated with SZP to 2.3 times that of the control group, indicating that SZP boost the maturation of DCs.However, the group with ATG5 knockdown reversed the maturation of DCs induced by SZP (Fig. 6H, I).Additionally, elevated secretion of cytokines (IL-12p70, IL-10, TGF-β, and IL-6) that embody DCs activation (Fig. 6J) was detected by ELISA.In summary, SZP NPs could evoke ICD of sublethally heated HCC cells in an autophagy-dependent manner and boosted the maturation of DCs.

Autophagy-dependent ICD induced by SZP and immune vaccine experiments in vivo
We established the IRFA subcutaneous H22 model to further evaluate whether SZP could induce ICD in vivo.SZP NPs (concentration of STF: 4 mg/kg) were injected through the tail vein every 2 days, 4 consecutive times (Fig. 7A).The release of DAMP which characterized ICD was investigated.Immunofluorescent analysis showed that SZP increased CRT exposure, HMGB1 release as well as LC3 expression in tumor tissues (Fig. 7C).However, in ATG5-knockdown mice tumors, SZP-induced DAMP release was significantly reduced (Fig. 7E, F, G).Meanwhile, we collected tumor-draining lymph nodes (TDLN) of the mice for evaluating DCs maturation.It was found that SZP treatment significantly promoted the maturation of DCs, which was 2.5-fold higher than that of the control group.On the contrary, SZP treatment failed to induce DC maturation in ATG5-knockdown mice tumors (Fig. 7B, D).These results illustrated that SZP NPs triggered ICD in residual tumors after IRFA and then contributed to the maturation of DCs in an autophagy-dependence manner.
To investigate whether SZP could activate an effective anti-tumor immune response after IRFA, we carried out a vaccination experiment, which is the gold standard for evaluating ICD in vivo.Firstly, H22 cells were exposed to different treatments (sublethal heat, sublethal heat + STF, sublethal heat + SZP) for 24 h as a tumor cell vaccine and injected into the left back of BALB/c mice twice with an interval of 7 days (Fig. 7H).Subsequently, the contralateral side of the mice was challenged with live H22 cells.It turned out that 12 days after tumor transplantation, new tumors could be visibly noticed in the control group and the heat group, but tumors in the latter group grew slower than those in the former.Such difference implied that heat-treated cell vaccine was a weak ICD-inducer that could not prominently suppress the growth of new tumors.On day 24, vaccination with heat + STF-treated cells and heat + SZP-treated cells significantly suppressed the growth of new tumors with a 30% (3/10) and 40% (4/10) tumor-free ratio, respectively (Fig. 7I).This suggested that vaccination derived from heat + STF and heat + SZP played a strong role in impeding tumor initiation.Moreover, the tumor grew much more slowly in the 6 tumor-developed mice of the heat + SZP group (Fig. 7J).Compared to the control group, the inhibition rates of tumor growth in the heat + SZP group and heat + STF group were 81.55 ± 5.96% and 61.55 ± 5.96% accordingly (Fig. 7K).It implied that SZP may serve as a more potent ICD-inducer of sublethally heated cells.Meanwhile, immunohistochemical results revealed that SZP could activate the systemic anti-tumor immune response, which was verified by the substantial increase in CD4 + and CD8 + T cell infiltration in tumor tissues (Fig. 7L).A significant rise in the secretion of IFN-γ and TNF-α (Fig. 7O) was also captured, the former of which was a typical anti-tumor cytokine secreted from activated T cells, and the latter was a cytokine from macrophage that could directly kill tumors and mediate immunity [50].Furthermore, we collected the spleens of mice after the injection of H22 cells and analyzed the effector memory T cell (TEM) in the spleen via flow cytometry.The results demonstrated that there was an obviously higher percentage of TEM in the spleen undergoing heat + STF-treated and heat + SZP-treated cells vaccine (Fig. 7M), which was 4.58 and 3.16 times as that of the control group (Fig. 7N).In conclusion, it was considered that vaccination with heat + SZP-treated cells could act as in situ vaccine to suppress the growth of new tumors, activate a systemic anti-tumor immune response, and establish long-term immunological memory.

SBZP inhibits the growth of residual tumors after IRFA
Encouraged by the results of activating ICD and establishing immune memory in vivo, we further evaluated whether SZP could reverse the immunosuppressive microenvironment after IRFA and thus solve the clinical difficulty of poor sensitivity to anti-PD-1/PD-L1 therapy.BMS is a novel small-molecule inhibitor, characterized by high oral bioavailability, strong permeability into solid tumors, and low cost compared with anti-PD-1/PD-L1 antibodies.However, features such as aggregation and hydrophobicity in aqueous media limit its efficacy, which can be overcome by the encapsulation of nanocarrier in the present study [51].We tested the inhibitory effects of different NPs, the specific procedures were illustrated in Fig. 8A.H22-bearing mice undergoing IRFA were divided randomly into five groups (PBS, STF, BZP, SZP, and SBZP) and administered with NPs (concentration of 4 mg/kg STF and 2 mg/kg BMS) via the tail vein every 2 days for 6 consecutive times.Western blot and immunofluorescence were used to verify the activation of autophagy induced by NPs.The results demonstrated that the level of LC3B in SZP and SBZP treatment groups obviously increased (Fig. 8B and Additional file 2: Fig. S15).Next, the curve of tumor volume and bioluminescent imaging (Fig. 8C, E, H) revealed that the BZP group exhibited a mild inhibitory effect on the residual tumor.This undesirable outcome indicated that the residual tumors after IRFA had poor sensitivity to anti-PD-1/PD-L1 therapy.In contrast, the SZP and SBZP groups displayed a significant inhibitory effect on residual tumors.Both SZP and SBZP initially performed well in inhibiting tumors.However, the residual tumors regrew slowly 22 days after the treatment of SZP.Therefore, using SZP alone was incapable of achieving lasting outcomes.Contrastively, the sustained suppression of residual tumors in the SBZP group indicated that STF could enhance the sensitivity of residual tumors to anti-PD-1/PD-L1 therapy.On day 24, all mice were euthanized, and tumors were extracted and weighed (Fig. 8D).The results showed that in comparison with the control group, the tumor inhibition rates of the BZP, SZP, and SBZP groups were 29.62%, 71.00%, and 86.76%, respectively (Fig. 8F).In addition, we evaluated the influence of different NPs on tumor proliferation and apoptosis.It turned out that the proportion of Ki-67 + decreased and TUNEL + increased in the SBZP group (Fig. 8G), illustrating that SBZP was able to considerably inhibit tumor proliferation and contribute to tumor apoptosis.These data suggest that SBZP NPs can exert a strong inhibitory effect on the growth of residual tumors and combination therapy has a more significant outcome.
There was no significant change in the body weight of the mice in each group (Fig. 8I), which indicated that NPs (SZP, BZP, and SBZP) induced few side effects.H&E staining of major organs revealed no obvious histopathological changes (Additional file 2: Fig. S17).Biochemical blood examination (Additional file 1: Fig. S9) and blood routine (Additional file 1: Fig. S10) also indicated that SBZP NPs were biocompatible.

SBZP remodels the immune microenvironment of residual tumors after IRFA
The modulatory effect on the tumor immune microenvironment was examined to explore its potential antitumor mechanism of SBZP.We found that SBZP treatment boost CRT exposure and HMGB1 release in tumor tissues (Fig. 9A, B), and induced the maturation of DCs in the TDLN with a level 3.64 times and 5.10 times that of the control group (Fig. 9D, E).In addition, we discovered that SBZP activated systemic T cell anti-tumor immune responses.Immunofluorescence staining (Fig. 9A), immunohistochemistry (Additional file 2: Fig. S16), and flow cytometry (Fig. 9F) detected prominent enhancement in the tumor infiltration of CD4 + T cells, CD8 + T cells, and CD11C + DCs cells in the SBZP group.According to flow cytometry, the proportion of TAMs infiltration after SBZP treatment was obviously reduced, lower by 71.5% and 81.1%, respectively (Fig. 9G) in comparison with the control group.Furthermore, an obvious increase in tumor-killing cytokine (IFN-γ, TNF-α) and a decrease in tumor-promoting cytokines (TGF-β, IL-6) in the serum of SBZP-treated mice were found by ELISA (Fig. 9C).In conclusion, SBZP NPs can activate ICD, reverse the immunosuppressive microenvironment, and activate anti-tumor immune response after IRFA.They can serve as promising immunotherapy for inhibiting residual tumors after IRFA.

Conclusion
In summary, we confirmed that IRFA induced protective autophagy and aggravated the immunosuppressive microenvironment.To this end, we successfully constructed SBZP NPs, which prominently inhibited the progression of residual tumors after IRFA.Our study provided a new therapeutic modality that combining the amplification of autophagy with ICB.Through this strategy, SBZP could convert protective autophagy after IRFA to ACD by enhancing autophagy, triggering the release of immunogenicity and evoking ICD.Hence, SBZP shows great performance in remodeling the immune microenvironment and strengthening immune surveillance.It can be concluded that this new modality is endowed with great potential in treating residual tumors after IRFA.

Fig. 5 Fig. 6
Fig. 5 SZP mediates ACD.A The dose-response curve for CCK8 assay of IC50 of STF62247 in different concentrations treating cells at 37 °C and 47 °C for 24 h (n = 3).B The dose-response curve for CCK8 assay of IC50 of STF62247 and SZP NPs in different concentrations treating HCC cells at 47 °C for 24 h (n = 3).C, F Live/dead cells double staining and quantitative analysis of sublethally heated SMMC7721 and Huh7 cells exposed to different groups (n = 4).Scale: 1000 µm.D Analysis of the proliferation of sublethally heated SMMC7721 and Huh7 cells under influence of different groups by CCK8 (n = 3).E, G Representative flow cytometry plots and quantitative analysis of apoptosis in sublethally heated SMMC7721 and Huh7 cells under influence of different conditions (n = 3).H Immunoblotting of p-mTOR, t-mTOR, P62, and LC3B in sublethally heated SMMC7721 and Huh7 cells under influence of different groups.The results were expressed as mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.ACD, Autophagy-dependent cell death; IC50, Half maximal inhibitory concentration

Fig. 7
Fig. 7 Autophagy-dependent ICD induced by SZP and immune vaccine experiments in vivo.A Flow chart of in vivo detection of ICD in tumor-bearing mice.B, D Detection and quantitative analysis of mature DCs (of CD11C + gate) in tumor-draining lymph nodes via flow cytometry (n = 3).C, E-G Expression and quantitative analysis of CRT, HMGB1, and LC3B in tumors via CLSM (n = 3).H Flow chart of in vivo inoculation with immune vaccine.I: Kaplan-Meier plot of survival rate in different groups of tumor-free mice (n = 10).J, K Comparison of tumor growth curves and tumor weights in mice of different groups (n = 6).L Representative IHC images showing the expression of CD4 and CD8 in mice tumor of different groups.M, N Representative flow cytometry plots and quantitative analysis of TEM (of CD8 + gate) in the spleen of mice (n = 3).O The levels of IFN-γ and TNF-α in mice serum from inoculated with different immune vaccines by ELISA (n = 4).Results were expressed as mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.IHC, Immunohistochemistry; TEM, effector memory T cells

Fig. 8
Fig. 8 SBZP inhibits the growth of residual tumors after IRFA.A Flow chart of in vivo IRFA for subcutaneous tumors with the treatment of NPs.B Immunoblotting of LC3B in tumor tissues treated with different NPs.C, E IVIS bioluminescent imaging and quantitative analysis of mice tumors before and 24 days after IRFA (n = 5).D, F, H, I Photos, weights, and volumes of the tumors in mice were treated with different NPs (n = 5).G Representative CLSM graphs of TUNEL, Ki-67 fluorescent staining, and H&E of tumor tissues treated with different NPs.Scale: 100 µm.Results were expressed as mean ± SD, *pp < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.IVIS, In vivo Imaging System