Simultaneous induction of immunogenic cell death and PD-L1 downregulation by bufalin-based nanovaccines for potentiate HCC immunotherapy

Abstract

Immunotherapy is a promising treatment in hepatocellular carcinoma (HCC), but with low response rate clinically. Immunogenic cell death (ICD) is considered as a strategy to enhance immunotherapy response. However, chemotherapeutic drugs with ICD make tumor cells upregulating PD-L1 expression to deactivate T-cells via PD-1/PD-L1 pathway. Here, Bufalin (Buf) was validated for the first time as an ICD inducer. Buf triggered reactive oxygen species (ROS) related endoplasmic reticulum (ER) stress to elicit apoptosis and ICD via PERK/eIF2α/ATF-4/CHOP in HCC cells. Additionally, Buf downregulated the expression of PD-L1 to avoid immune escape. Buf can simultaneously activate dendritic cell (DC) maturation and interrupt the PD-1/PD-L1 pathway. To amplify immunotherapy and decrease adverse cardiac reactions of Buf, SP94-modified liposome-coated zeolite imidazolate framework-8 loaded with Buf (Buf-ZIF-lipo-SP94) was designed to effectively increase drug accumulation in tumor by HCC-specific targeting SP94 receptor-mediated endocytosis. Importantly, ZIF-8 with the ability of triggering ROS generation itself synergized Buf to induce stronger ICD effects. Buf-ZIF-lipo-SP94 achieved synergistic effects with anti-PD-L1 for HCC immunotherapy, showing better tumor inhibition rate (>90 %), survival of animals and safety. This study uncovered the potential of Buf in inducing ICD and downregulating PD-L1 expression, developing a Buf-loaded nanovaccine for combination strategy of immunotherapy in HCC.

Graphical abstract

Buf triggered ROS-mediated ER stress to elicit ICD by PERK/eIF2α/ATF-4/CHOP signaling pathway, and reduced PD-L1 expression of HCC. Buf-ZIF-lipo-SP94 increased drug accumulation in tumor with less adverse cardiac reactions, enhanced the activity of ICD, ameliorated immune escape and created favorable conditions for T cell related anti-tumor immune response, which could be used in conjunction with anti PD-L1.

1. Introduction

HCC, a major leading cancer worldwide with poor prognosis, is always diagnosed at an advanced stage with limited therapeutic options [1,2]. Innovative immunotherapies have emerged as a key trend for HCC treatment, such as immune checkpoint inhibitors (ICIs), which ensure long-term survival and manageable toxicity [3,4]. ICI-based combination therapies have been widely recognized as the first-line treatment for the development treatment of HCC [5]. However, only a portion (20–30 %) of patients can benefit from ICIs, due to the heterogeneity and instability of tumors leading to low immunogenicity, as well as the low infiltration of effector T cells in tumors [3,6]. A series of clinical anti-tumor trials suggests that inducing ICD can enhance the response of ICIs to present synergistic effects [7,8].

ICD is a unique immunostimulatory modality of regulated cell death (RCD) that contributes to activate both innate and adaptive immunity by increasing the tumor-specific immunogenicity [9]. Mechanistically, the concomitant activation of ROS production and ER stress activation has been found be indispensable for ICD induction, especially promoting the exposure and emission of DAMPs, such as CALR, ATP and HMGB1 [[10], [11], [12]]. Wherein, eukaryotic translation initiation factor 2-alpha kinase 3 (EIF2AK3, also known as PERK), a main ER stress related proteins, responds to ER stress by inducing eukaryotic translation initiation factor 2 subunit alpha (eIF2α) phosphorylation, which controls a surface exposure of CALR [13,14]. PERK/eIF2α triggers cell death and intranuclear HMGB1 exclusion via regulating transcription factor 4 (ATF-4) and DNA damage inducible transcript 3 (DDIT3, also known as CHOP) downstream. The extracellular DAMPs initiate the uptake and processing of antigens recognized by antigen-presenting cells (APC), like DCs, for subsequent activation of effector T cells, especially cytotoxic T lymphocytes. But clinical trials combining ICD inducers and ICIs are rarely conducted in HCC [9].

Application of current ICD inducers, mainly including chemotherapy drugs, radiotherapy, oncolytic virotherapy, extracorporeal phototherapy, photodynamic therapy, and targeted therapy agents, still face some practical issues [10,15,16]. Moreover, the ICD-inducers generally upregulate tumor cells’ PD-L1 expression and must combine with PD-1/PD-L1 blockade to gain therapeutic efficacy via either addingor in-situ expression of the antibodies [13,[15], [16], [17]]. So we hypothesized that an ICD-inducer that simultaneously down-regulates the PD-L1 expression would eliminate the need for the expensive anti-PD-1/L1 antibodies.

Many studies have indicated that Buf modulated anti-tumor chemoimmunotherapy at a low dose. For example, Buf could increase intratumoral infiltration of T lymphocytes, drive tumor-infiltrating macrophages toward classical activated macrophages (M1) phenotype and enhance the killing efficaciousness of natural killer cells (NK) in HCC treatment [[18], [19], [20], [21]]. However, Buf as ICD inducer for HCC treatment has never been reported. Interestingly, Buf belongs to cardiac glycosides, and can cause intracellular oxidative stress and indirectly lead to ER stress, which are the characteristics possessed by the Na⁺/K⁺-ATP ase [[22], [23], [24]]. Hence, the effect and mechanism of Buf as a potential ICD inducer was confirmed for the first time by vaccine experiments, determination of DAMPs and other validated experimental methods. Unexpectedly, it was discovered that Buf downregulated PD-L1 expression to prevent immune escape, which contributed to anti-tumor immunity of T cells. Dual immune regulatory effect of Buf reformed the immunosuppressive TME to provide possibilities for synergism with ICI. But the potential side effects of Buf including the low in vivo bioavailability and cardiovascular adverse reactions have been reported in preclinical and clinical trials [25,26].

In this study, Buf was confirmed as a potential ICD inducer in HCC and the mechanism was elucidated as that Buf triggered ROS-mediated ER stress to elicit apoptosis and ICD by PERK/eIF2α/ATF-4/CHOP signaling pathway in HCC cells, further increasing maturation of DCs. Furthermore, the potential role of Buf in reducing PD-L1 expression added to the enhancement of immune response. To amplify immunotherapy and decrease adverse cardiac reactions of Buf, nanomedicine was introduced into drug delivery, such as traditional liposomes and metal frame nanomaterials, which provide various functionalities and expand drug application in enhancing tumor therapy [27,28]. SP94-modified liposome-coated ZIF-8 loaded with Buf (Buf-ZIF-lipo-SP94) was designed to effectively increase drug accumulation in tumor by HCC-specific targeting SP94 receptor-mediated endocytosis and pH responsive release in acidic tumor microenvironment (TME). Importantly, Buf-ZIF-lipo-SP94 enhanced ROS-mediated ER stress to induce more significant ICD effects, and reduced PD-L1 expression to ameliorate the immune escape. Ultimately, Buf-ZIF-lipo-SP94 contributed to achieve synergistic effects with anti-PD-L1 in HCC treatment as expected, with a more than 90 % tumor inhibition rate and high survival of animals (Scheme 1).

Scheme 1.

Buf triggered ROS-mediated ER stress to elicit ICD by PERK/eIF2α/ATF-4/CHOP signaling pathway, and reduced PD-L1 expression of HCC. Buf-ZIF-lipo-SP94 increased drug accumulation in tumor with less adverse cardiac reactions, enhanced the activity of ICD, ameliorated immune escape and created favorable conditions for T cell related anti-tumor immune response, which could be used in conjunction with anti PD-L1.

2. Materials and methods

2.1. Cell lines and animals

Mouse liver cancer Hepa1-6 cells, human liver cancer HepG2 and Huh7 cells (Cell Bank of the Chinese Academy of Sciences, China) were cultured in MEM or DMEM supplemented with 10 % fetal bovine serum (FBS) and 1 % penicillin/streptomycin at 37 °C in a humidified 5 % CO₂ incubator (Esco Lifesciences Group, Singapore). C57BL/6 (C57) mice (14–16 g) and BALB/c nude mice (14–16 g) were purchased from Shanghai Slac Lab Animal Co., Ltd. (Shanghai, China), with approval file No. PZSHUTCM2302020012 and PZSHUTCM220919014. The animal experiments were conducted under the principles of care and use of laboratory animals.

2.2. Vaccine experiment in vivo

Hepa1-6 cells were pretreated with PBS, Doxorubicin (DOX), and Buf (Anhui Xiqingguo Biotechnology Co., Ltd.) in vitro for 24 h, respectively. And Hepa1-6 cells boiled at 100 °C for 5 min were the dead cells. C57BL/6 mice were randomly divided into 5 groups (n = 6 in every group). The mice were injected under the right armpit with 0.1 mL pretreated Hepa1-6 cell suspension (1 × 10⁶ cells) as vaccination. Among which, injecting Hepa1-6 cell pretreated with PBS was negative group, DOX as positive group, Buf as experimental group and dead cells as control group. One group with no treatment on the right armpit was as blank group. After 7 days, the mice were all rechallenged with untreated Hepa1-6 cells (1 × 10⁶ cells) on the left armpit, and the size of tumors on the left was monitored. The tumor volume was calculated with the following formula:

$$V_{\text{tumor}\text{volume}}=L\times W^2/2$$

Where L is the longest dimension; W is the shortest dimension.

The method of vaccine experiment conducted on nude mice was the same as above.

2.3. In vitro cell cytotoxicity and apoptosis analysis of Buf

The in vitro experiments used human liver cancer cells (HepG2 and Huh7 cells). The cell cytotoxicity of Buf on HepG2 cell and Huh7 cell was detected using CCK-8 assay (Beyotime, China). The cells were seeded in 96-well plates (1 × 10⁴ cells per well), and were respectively incubated at graded increasing concentrations of Buf for 24 h. Then 110 μL medium containing 10 % CCK-8 was added and incubated for an additional 1 h at 37 °C to measure the optical density value (OD values) at 450 nm. The cell viability and half maximal inhibitory concentration (IC₅₀) of Buf were calculated by SPSS. Meanwhile, the apoptosis of tumor cells induced by Buf was observed by Annexin V-FITC/PI apoptosis kit (Beyotime, China). HepG2 cells and Huh7 cells were cultured in 6-well plates and treated with and Buf (0.5 μM, 1 μM) for 24 h, respectively. Then cells were stained with Annexin V-FITC/PI successively for 20 min in the dark at 25 °C. The proportion of apoptosis was finally quantified by flow cytometry analysis.

2.4. In vitro assays of ICD expression

To verify the ICD effect on HepG2 cells and Huh7 cells induced by Buf, the expression level of CALR on the cell surface, the release of HMGB-1 and the secretion of ATP were checked. The tumor cells were incubated with Buf (0.5 μM) and Paclitaxel (PTX, positive control, 20 μM) for 8 h, and then the cells were stained in order with Alexa Fluor 488 anti-Calreticulin antibody (Abcam, UK) for 30 min in the dark. The percentage of CALR positive staining cells was determined by flow cytometry (Beckman, USA). In addition, the ATP concentrations in cells after incubated by different drugs were tested by ATP assay kit (Beyotime, China) to deduce the level of ATP secretion in reverse, because of the instability of ATP in culture medium at 37 °C. The culture medium after incubation also collected and centrifuged to remove impurities were used to quantify the HMGB1 release level by HMGB1 Elisa kit (Elabscience, China).

2.5. Determination of intracellular ROS generation

The intracellular ROS was measured to verify the ability of Buf to induce the generation of ROS in HCC cells. First, HepG2 cells and Huh7 cells were treated by PBS or Buf for 8 h. After removing culture medium, the 2,7-dichlorodihydrofluorescein diacetate (DCFH-DA) probe (5 μM) from ROS assay kit (Beyotime, China) was added to incubate for 20 min at 37 °C in dark to assess ROS production, and then the fluorescence of cells was detected by flow cytometry to evaluate intracellular ROS levels.

2.6. Ultrastructural changes of ER induced by Buf

Ultrastructural changes of ER in tumor cells were visualized under transmission electron microscope (TEM) (SEMC, New York, USA) to determine the ability of Buf to induce ER stress in HepG2 cells and Huh7 cells. Meanwhile, to further determine the correlation between ER stress and ROS, the antioxidant N-acetylcysteine (NAC, 10 mM) (Yuanye, China) was added as the pretreatment for 4 h HepG2 cells and Huh7 cells were treated with PBS, Buf (0.5 μM), NAC + Buf separately. With the corresponding post-processing, the cells were fixed with 2.5 % glutaraldehyde at 4 °C overnight. And then the cells were fixed in 1 % OsO₄ for 1 h, stained with 1 % uranyl acetate, dehydrated through graded acetone solutions, and embedded. Ultimately, the cell areas were blocked mount and cut into slices. The morphological changes of ER lumens were observed by TEM as the direct evidence of ER stress.

2.7. RNA sequencing

HepG2 cells and Huh7 cells were treated with Buf (0.5 μM) for 8 h. Total cellular RNAs were extracted from cells by using TRIzol (Beyotime, China). Transcriptome sequencing was performed by Majorbio (Shanghai, China). The mRNA sequencing was analyzed on the online platform of Majorbio Cloud Platform. Differential expression analysis between Buf treated group and control group was performed by DESeq2 with p < 0.05 and a fold change >1.0 as the threshold for significant differential expression. Data were transformed into heatmap, and KEGG enrichment analysis of differentially expressed genes was implemented by the hierarchical clustering.

2.8. Western blotting and qRT-PCR analysis

HepG2 cells and Huh7 cells were treated with Buf (0.5 μM), NAC + Buf and NAC separately. The total proteins of cells were extracted by RIPA buffer (containing 1 % PMSF) (Beyotime, China). The concentration of protein samples was quantified by a BCA assay kit (Beyotime, China). Equal protein amounts of samples were separated by 10 % SDS-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto polyvinylidene fluoride (PVDF) membranes, and incubated with specific corresponding antibodies, containing p-PERK, p-eIF2α, ATF4, CHOP and PD-L1 (Beyotime, China), overnight at 4 °C followed by corresponding secondary antibody for 1 h. The results were visualized by a chemiluminescence image analysis system.

Total cellular RNAs were extracted from cells by TRIpure. And cDNA was obtained with the Prime Script™ RT reagent Kit (TaKaRa, Japan). The qRT-PCR was performed using the Hieff®qPCR SYBR Green PCR Master Mix in Step One Plus Real-Time PCR System. And relative expression levels were calculated by 2 [−ΔΔCt] method and expressed as “fold change”. β-actin was used to normalize mRNA expression. The primer sequences are used as followed:

ATF-4: forward, 5′ CTT ACA ACC TCT TCC CCT TTC 3'

reverse, 5′ GTC TGG CTT CCT ATC TCC TTC 3'

CHOP: forward, 5′ AAC GGC TCA AGC AGG AAA TC 3'

reverse, 5′ TTC ACC ATT CGG TCA ATC AGA 3'

β-actin: forward, 5′ AAG GTG ACA GCA GTC GGT T 3'

reverse, 5′ TGT GTG GAC TTG GGA GAG G 3'

2.9. Synthesis and characterization of Buf delivery system

ZIF-8 was prepared according to the previously report [29]. Zn(NO₃)₂·6H₂O (150 mg) and 2-methylimidazole (300 mg) were suspended in methanol (14 mL) with stirring for 5 min to homogenize the solution, and washed 3 times with methanol. The product was finally collected through centrifugation (10000 rpm, 10 min). For the preparation of Buf-ZIF, Buf (10 mg) was dissolved in 1 mL methanol suspended with ZIF-8 (10 mg) and the solution was stirred overnight. The Buf-ZIF was collected by centrifugation to remove methanol and uncoated Buf. Buf-ZIF-lipo-SP94 were prepared with thin film dispersion-probe ultrasonic method. First, lipid materials DSPC (AVT (Shanghai) Pharmaceutical Tech Co., Ltd. China) and DSPE-mPEG2000-SP94 (Weihua, China) with a molar ratio of 8:2 were dissolved in chloroform and dried by a rotary evaporation at 60 °C (water bath) to form a uniform thin film in a round-bottomed flask. Then the film was hydrated in Buf-ZIF solution at 50 °C for 10 min. Finally, the suspension was sonicated by an ultrasonic processor (Scientz, China) in ice bath to obtain Buf-ZIF-lipo-SP94. Buf-ZIF-lipo was obtained by replacing DSPE-mPEG2000-SP94 with DSPE-mPEG2000 (AVT, China). At last, the nanoparticles obtained were freeze-dried into powder for subsequent use.

The size, polydispersity index (PDI) and Zeta potential of Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 were determined using a Malvern Zetasizer (Malvern, UK). The morphology was observed by TEM. And drug loading (DL) of Buf in nanoparticles was determined with high-performance liquid chromatography (HPLC).

2.10. In vitro drug release study

Buf, Buf-ZIF and Buf-ZIF-lipo-SP94 (containing 0.5 mg Buf) were transferred into the dialysis bags (6000–8000 Da) and placed in 200 mL release medium (1 % tween 80 in 5 % glucose solution, pH 7.4 or pH 5.5) agitating with multifunctional shaker (100 rpm, 37 °C). The drug concentration and solution volume in the dialysis bag were measured at 0, 0.5, 1, 2, 4, 8, 12 and 24 h to calculate the cumulative release rate of different formulations, respectively.

2.11. In vitro cellular uptake

To evaluate the targeting ability of ZIF, phospholipid membrane and SP94, ICG-ZIF, ICG-ZIF-lipo, and ICG-ZIF-lipo-SP94 were prepared through placing Buf with Indocyanine Green (ICG) (Sanen, China) following the previous preparation method. The HepG2 cells were treated with fluorescent nanoparticles with the same concentration (10 μg/mL) of ICG for 8 h. The uptake of ICG in HepG2 cells was investigated by flow cytometry with λ_ex = 638 nm,λ_em = 750–800 nm.

2.12. In vitro cell cytotoxicity, apoptosis analysis and ICD induction of Buf-ZIF-lipo-SP94

The cell cytotoxicity and apoptosis induction of Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 (Buf = 0.5 μM) on HepG2 cells were detected as mentioned above. HepG2 cells were incubated with Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 (Buf = 0.5 μM) for 24 h. Then the cells were incubated by Calcein-AM/PI (Beyotime, China) for 20 min, and observed by confocal laser scanning microscopy (CLSM) (Leica, Germany) to distinguish dead and alive of tumor cells.

Equally, the different ability level of inducing ICD among Buf and other nanoparticles, were checked by examining the expression level of CALR, the release of HMGB-1 and the secretion of ATP as above. Besides, DC activation assay was conducted to measure the efficacy of nanoparticles to induce ICD.

2.13. In vitro hemolysis study of Buf-ZIF-lipo-SP94

Mice blood (4 mL) was added into centrifuge tube containing heparinization, and mixed with twice amount of normal saline. Then centrifuging at 3500 rpm for 10 min to get the blood cell samples. Water (positive control), 5 % glucose aqueous solution (negative control), Buf solution, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 were added to blood cell samples, respectively. The samples were centrifuged at 3500 rpm for 10 min after incubation for 30 min at 37 °C The corresponding UV absorption data of supernatant solution for each group were detected by UV spectrophotometer (540 nm wavelength condition), and the hemolysis degree of each group was calculated. The hemolysis rate (HR%) of the nanoparticle suspension was calculated as follows:

$$\text{HR（}\%）=\left(A_{sample}-A_{negative}/A_{positive}-A_{negative}\right)\times 100\%$$

Where A_sample is the absorbance of different preparation goups; A_negative is the absorbance of 5 % glucose aqueous solution group; and A_positive is the absorbance of Water gpoup.

2.14. In vivo biodistribution of Buf-ZIF-lipo-SP94

The subcutaneous tumor-bearing mouse model was constructed through injection of 0.1 mL Hepa1-6 cell suspension (1 × 10⁶ cells) under the right armpit. When tumor volume reached to about 150 mm³, the mice were randomly assigned to five groups (5 % glucose aqueous solution, ICG, ICG-ZIF, ICG-ZIF-lipo, and ICG-ZIF-lipo-SP94). The preparations were administered via tail vein. The IVIS Lumina XR Imaging System (PerkinElmer, USA) (excitation of 745 nm) was used to view the tumor accumulation at 1, 2, 4, 6, 8, and 24 h. Finally, all the mice were sacrificed. The tumor and main organs were excised, and were also imaged at the aforementioned excitation wavelength.

2.15. In vivo synergistic anti-tumor effects of Buf-ZIF-lipo-SP94 with anti-PD-1 antibodies

The subcutaneous tumor-bearing mouse model was constructed through injection of 0.1 mL Hepa1-6 cell suspension (1 × 10⁶ cells) under the right armpit. After the tumor volume reached to 100 mm³, the mice were stochastically divided into 6 groups including control (5 % glucose solution), Buf, Buf-ZIF, Buf-ZIF-lipo-SP94, Buf-ZIF-lipo-SP94+anti-PD-L1, and anti-PD-L1 (Buf = 0.5 mg/kg, anti-PD-L1 = 6 mg/kg, n = 5 each group). The drugs were injected through tail vein every other day, and the tumor volume and body weight were also monitored for two weeks.

After the end of administration, the tumors obtained from sacrificed mice were weighed and photographed for recording. The tumor growth inhibition rate (IR) of each group is calculated according to the following formula:

$$\text{IR}(\%)=\left({\mathrm{W}}_{\text{blank}}-{\mathrm{W}}_{\text{treat}}\right)/{\mathrm{W}}_{\text{blank}}\times 100\%$$

Where W_blank is the control group, and W_treat is the treatment group.

The tumor tissues were excised and fixed with 4 % paraformaldehyde (PFA), and then were paraffin-embedded and sectioned with a microtome. And tumor tissues were analyzed by Ki67, terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL), and hematoxylin-eosin (H&E) staining assay to evaluate the effect of different treatments on tumor proliferation and apoptosis.

2.16. Assessment of in vivo ICD induction and effects on T cell, DC and macrophage activation

After the in vivo synergistic anti-tumor study, tumor tissues and immune tissues (paracancerous lymph, spleen, and thymus) were obtained to investigate the changes of anti-tumor immune levels in mice after treatment. A portion of tumor tissue was fixed with 4 % PFA, and then were paraffin-embedded and sliced with a microtome. After paraffin-embedded and sectioned with the same method above, the sliced tissues were stained by CALR, HMGB1 antibody to evaluate ICD induction effect.

Paracancerous lymph, spleen, thymus and remaining tumor tissue were fully ground and passed through the 200-mesh sieve to acquire cell suspensions. The cell suspensions of lymph, thymus and spleen were incubated with antibodies for mature DC staining (CD45-APC-Cy7, CD11c-BV605, MHC-Ⅱ-APC, CD80-PE,CD86-PE-Cy7) (BD, USA). The cell suspensions of tumor and spleen were incubated with antibodies for T cells staining (CD45-APC-Cy7, CD3-Fitc, CD4-BV421, CD8-PE-Cy7) (BD, USA).

2.17. Cardiotoxicity evaluation

The electrocardiogram (ECG) of healthy C57 mice (16–18 g) was used to evaluate the impact of Buf on heart. The mice were stochastically divided into Buf group, Buf-ZIF group, and Buf-ZIF-lipo-SP94 group (Buf = 1 mg/kg). The ECG of the mice was recorded and analyzed by RM6240C biological signal collection and processing system (Chengdu Instrument Factory, China). The mice before administration were deemed as control. The ECG of the mice after administration were recorded for 1 h.

2.18. Systemic side effect and survival assessment

To assess whether the Buf-ZIF-lipo-SP94 had some systemic side effects, the blood samples and the histopathology of major organs (heart, liver, spleen, lung and kidney) were analyzed. The blood analysis was mainly for the liver/kidney functions, including aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood urea nitrogen (UREA) and creatinine (CREA). The organs were determined by H&E staining for pathological examination. Meanwhile, the behavioral status and mortality of mice were monitored within a month. When the rapid weight loss of the mouse exceeds 15 %, severe hemiplegia, inability to drink or feed or tumor volume>1500 mm³, the animal was judged dead, referring to the relevant requirements of experimental animal welfare. Draw Kaplan Meier survival curves and calculate median survival times (MST).

2.19. Statistical analysis

SPSS software and GraphPad Prism were used for statistical analysis of the data. The independent sample t-tests or one-way ANOVA were used for comparison between the two or more groups. ∗p < 0.05 indicates significant statistical differences, and ∗∗p < 0.01 indicates extremely significant statistical differences.

3. Results and discussion

3.1. Vaccine experiments in vivo

Inducing ICD has been emerged as an encouraging approach for establishing a cascade reaction of immunity. The initial stimulation of ICD to tumor cell death is generally immunogenic, and plays a key role in TME and stimulating some host immune response [30]. Common ICD inducers can provoke organismic anti-tumor immune response and memory by releasing DAMPs [31]. Accordingly, a vaccine experiment was performed to investigate whether Buf could improve the antigenicity of tumor cells through the ICD effect in the intact immune system mice. C57BL/6 mice were immunized on the right armpits with Hepa1-6 cells, which were pretreated respectively with PBS, DOX (positive control) and Buf in vitro, and rechallenged on the contralateral side with untreated living Hepa1-6 cells after seven days. The incidence and growth of tumor on the left armpits were monitored for 15 days (Fig. S1A). Unlike the PBS group, the mice vaccinated with the cells pretreated with DOX and Buf effectively blocked rechallenging oncogenic attack, and even if the tumor formed, it clearly grew slowly (Fig. S1B), owing to systemic immune memory response induced by pretreated cells on the right armpits. Meanwhile, other groups were also supplemented including without inoculating any cells and inoculating boiled Hepa1-6 cells on the right armpits, confirming that the nonimmunogenic death had no significance in activating anti-tumor immune response. In ICD-mediated immunotherapy, T cells are the important link in ultimately achieving anti-tumor immunity [32,33]. Therefore, nude mice (lack T cells) were also used for vaccine experiment (Fig. S1C). There was no anti-tumor effect on the left armpits detected among the different groups. The results determined the valuable contribution of T cells in ICD-mediated anti-tumor process and further provided support that Buf stimulated ICD to endow dying tumor cells with immunogenicity, and achieved the elicitation of immunological memory.

3.2. Identification of Buf as a potential inducer of ICD and PD-L1 degradation

Based on the results of in vivo vaccine experiments, the cytotoxicity and apoptosis induced by Buf were investigated which were the preconditions for ICD. As expected, Buf was efficient in inhibiting the growth of human liver cancer cells, with IC₅₀ of 0.82 μM (HepG2) and 0.42 μM (Huh7) in 24 h, respectively (Fig. 1A). Buf induced apoptosis at rates of 31.53 ± 7.04 % (1 μM) and 19.34 ± 7.14 % (0.5 μM) in HepG2, and 19.78 ± 0.92 % (1 μM) and 13.15 ± 1.30 % (0.5 μM) in Huh7, respectively, which were apparently higher than that of control (Fig. 1B). In general, ICD occurs along with the process of RCD, producing related DAMPs. The ability of Buf to induce CALR exposure, ATP secretion and HMGB1 release, which were considered as the necessary conditions for identifying the occurrence of ICD, was investigated [34]. HepG2 and Huh7 cells were treated with Buf and PTX (positive control) for 12 h. The proportion of CALR exposure on cell surface in different groups analyzed by flow cytometry was shown in Fig. 1C. Compared to controls, Buf significantly induced the expression of CALR in both HepG2 (44.38 ± 3.04 %) and Huh7 (37.69 ± 1.33 %) cells. Obvious release of HMGB1 in determining supernatant was detected after being treated with Buf (123.51 ± 17.88 pg/mL in HepG2 cells and 37.24 ± 8.05 pg/mL in Huh7 cells) compared with control (39.43 ± 26.38 pg/mL in HepG2 and 17.39 ± 2.54 pg/mL in Huh7) (Fig. 1D). While ATP was higher in control group than in Buf group (Fig. 1E). From the above results, Buf was proficient in promoting CALR exposure, ATP and HMGB1 release in both HepG2 and Huh7 cells, indicating Buf triggered ICD in HCC cells.

Fig. 1.

The cell death and DAMP emission driven by Buf. (A) Cell viabilities of HepG2 (a) and Huh7 (b) cells after incubation with Buf for 24 h. (B) Representative Annexin V-FITC/PI assay of HepG2 (a) and Huh7 (b) cells after different treatments and the quantification of the result. (C) The assessment of CALR exposure in HepG2 (a) and Huh7 (b) cells after different treatments and the quantification of CALR (CRT) exposure levels (c). (D) The quantification of HMGB1 release levels in HepG2 (a) and Huh7 (b) cells supernatant after different treatments by Elisa. (E) The quantification of ATP levels in HepG2 (a) and Huh7 (b) cells after different treatments by ATP assay kit. (∗: p < 0.05, ∗∗: p < 0.01, n = 3).

3.3. The mechanisms of Buf in inducing ICD

ROS surge and ER stress activation are established to be the important steps in ICD induction, and ROS production is a part of pathways to trigger ER stress response [[10], [11], [12]]. Previous studies showed that Buf could mobilize the associated protein expressions of ROS, ER stress and apoptosis in several kinds of tumor cells [[35], [36], [37]]. And Buf in promoting ROS production was tested and verified here through flow cytometry (Fig. S2). Meanwhile, the ultrastructural change in the ER was significantly affected by Buf, which mainly manifested as enlargement of ER lumen (a morphological evidence of ER stress) [30,38]. While the morphology of ER was partially normalized because of the N-acetylcysteine (NAC) pretreatment before Buf, which illustrated that ROS production induced by Buf might play an important role in ER stress (Fig. 2A). CALR exposure relies on ER stress which triggers PERK activation and eIF2α phosphorylation to induce ICD [39,40].

Fig. 2.

Buf induced ROS-dependent ER stress in HCC cells. (A) The TME images of HepG2 cells exposed to PBS (a), Buf (b) and NAC + Buf (c) for 8 h, and the TME images of Huh7 cells exposed to PBS (d), Buf (e) and NAC + Buf (f) for 8 h (arrow decorations represent the ER). (B) The pathway analysis of functional enrichment of these different mRNA in KEGG databases (a), and the heatmap generated based on the significant different genes in protein processing in endoplasmic reticulum (b) (HepG2). (C) A schematic diagram showing how Buf induced ER stress via and pathways and excite the emissions of ICD-associated DAMPs. (D) ER stress-related proteins of HepG2 cells and Huh7 cell were analyzed by western bloting (1: Control, 2: Buf, 3: NAC + Buf, 4: NAC). (E) The expression of ER stress-related gene, ATF-4 (a) and CHOP (b) in HepG2 cells detected by qRT-PCR. (F) The expression of ER stress-related gene, ATF-4 (a) and CHOP (b) in Huh7 cells detected by qRT-PCR. (∗: p < 0.05, ∗∗: p < 0.01, n = 3).

Transcriptome sequencing was performed to probe the mechanism of Buf-induced ICD effects in HCC cells. Buf could significantly change genes in the protein processing in ER pathway (Fig. 2Ba and Fig. S3A). The identified genes involved in protein processing in this pathway were illustrated by the heatmap (Fig. 2Bb and Fig. S3B), including EIF2AK3, phosphorylated eIF2α (p-eIF2α), activating transcription factor 3 (ATF-3), ATF-4, DDIT3 and CALR reported as the downstream genes of the ER stress signaling pathway.

The altered expressions of ER stress related protein, including phosphorylated PERK (p-PERK). p-eIF2α, ATF-4, and CHOP, were tested by western blotting. Due to the impact of Buf, phosphorylation levels of PERK and eIF2α were upregulated. Moreover, the expressions of ATF-4 and ATF-4-dependent target protein CHOP were also improved (Fig. 2D and S4). Consistent with protein expression levels, the gene expressions of ATF-4 and CHOP were also verified by qRT-PCR and presented an upward trend with the treatment of Buf (Fig. 2E and F). And the protein and gene expressions in the signaling pathway were all weakened when pretreated with NAC prior to Buf, which explained that the ER stress was mediated by ROS generation.

In a word, Buf could elicit ROS-mediated ER stress. ER stress induced CALR exposure to the cell surface and enhanced the release of ATP and HMGB1 through PERK/eIF2α/ATF-4/CHOP signaling pathway (Fig. 2C). And the apoptosis induced by ER stress and ROS is always observed [41]. So with that, the ability of Buf in inducing apoptosis and ICD was dropped significantly in the context of an initial pretreatment with NAC (Fig. S5).

3.4. Characterization of Buf loaded nano delivery system

To address issues such as poor solubility and low bioavailability of Buf, ZIF was obtained by one-step synthesis method with Zn(NO₃)₂·6H₂O and 2-methylimidazole. And Buf was loaded into ZIF to obtain Buf-ZIF. Buf-ZIF-lipo was prepared by encapsulating Buf-ZIF through the preparation of PEG-modified liposomes. And the final nanoparticles Buf-ZIF-lipo-SP94 was successfully prepared with DSPE-mPEG2000-SP94 modification. The nanoparticles all had uniform distribution, and the TEM image illustrated the polyhedral shape of ZIF, and the spherical structure of Buf-ZIF-lipo-SP94 which wrapped a layer (about 20 nm) of phospholipid outside ZIF (Fig. 3A). The particle size of ZIF and Buf-ZIF obtained was 156.1 ± 3.7 nm and 159.6 ± 5.4 nm, and the particle size of Buf-ZIF-lipo-SP94 were reached to 186.7 ± 4.8 nm after the coating of phospholipid (Fig. 3B and C). Moreover, compared with Buf-ZIF, the Zeta potential of Buf-ZIF-lipo-SP94 changing from 29.9 ± 5.4 mV to −23.6 ± 0.6 mV (Fig. 3B and C), are less easily to be recognized and cleared in the bloodstream [42]. The DL of Buf was estimated by HPLC [43] to be as high as 2.73 ± 0.13 % in Buf-ZIF-lipo-SP94 (Table S1). Further, the result of nitrogen adsorption-desorption in Fig. S6 also shown the surface area and pore size of ZIF significantly decreased after drug loading, indicating that ZIF successfully loaded the Buf.

Fig. 3.

The characterizations of Buf-ZIF-lipo-SP94. (A) The TEM images of ZIF (a), Buf-ZIF (b) and Buf-ZIF-lipo-SP94 (c). (B) The particle size and Zeta potentials of ZIF (1), Buf-ZIF (2) and Buf-ZIF-lipo-SP94 (3). (C) The size and Zeta potential distributions of ZIF, Buf-ZIF and Buf-ZIF-lipo-SP94. (D) In vitro cumulative release curve of Buf in different nanoparticles under pH 7.4 and pH5.5. (E) The stability of different nanoparticles. (F) The fluorescence intensity and quantitative results of cell uptake of ICG-ZIF, ICG-ZIF-lipo and ICG-ZIF-lipo-SP94 analyzed by flow cytometry after 8 h incubation. (∗: p < 0.05, n = 3).

Based on the decomposition of ZIF in acidic condition [44], the in vitro drug cumulative release of Buf from Buf-ZIF-lipo-SP94 in 5 % glucose solution with pH 7.4 and pH 5.5 was evaluated to assess whether the acidic TME is conducive for the release of Buf. Compared with the complete release of free drugs within 1 h, Buf-ZIF and Buf-ZIF-lipo-SP94 exhibited sustained release of 89.34 % and 86.53 % respectively in 8 h (pH 7.4). The pH-responsive release was observed that the cumulative release of Buf-ZIF in pH 7.4 and pH 5.5 were 37.58 % and 67.63 % within 1 h, and the cumulative release of Buf-ZIF-lipo-SP94 in pH 7.4 and pH 5.5 were 29.74 % and 35.71 % within 1 h (Fig. 3D). And the release of Buf from Buf-ZIF-lipo-SP94 was also slower than from Buf-ZIF due to the effect of phospholipid coating.

Freeze-drying is the standard method for long-term storage of nanoparticles [45]. Buf-ZIF-lipo-SP94 rehydrated in 5 % glucose solution after freeze-dried had no significant change in sizes (from 186.7 ± 4.8 nm to 164.0 ± 7.2 nm), while Buf-ZIF changed from 159.6 ± 5.4 nm to 235.5 ± 12.7 nm. And Buf-ZIF-lipo-SP94 was more excellent stability than Buf-ZIF after freeze-dried and in solution for 24 h (Fig. 3E), suggesting the modification of phospholipid membrane additionally stabilizes the nanoparticles.

SP94 is a specific targeting peptide (amino acid sequence: SFSIHTPILP), whose targeting ability arises from its high affinity binding to the asialoglycoprotein receptor (ASGPR) overexpressed on the surface of liver cancer cells. To explore the uptake of the Buf-ZIF-lipo-SP94 by HCC cells, the fluorescence labeled nanoparticles (ICG-ZIF, ICG-ZIF-lipo and ICG-ZIF-lipo-SP94) were synthesized by replacing Buf with ICG. As shown in Fig. 3F, the flow cytometry results demonstrated that ICG-ZIF (54.21 ± 6.99 %) was rapidly take up by HepG2 cells after 8 h incubation, and the modification of ordinary phosphorus layer had little effect on cellular uptake of ICG-ZIF-lipo (58.65 ± 4.06 %). While, the SP94 peptide mediated more efficient endocytic uptake of ICG-ZIF-lipo-SP94 (74.90 ± 2.58 %) in HepG2 cells due to the HCC-specific targeting ability [46].

3.5. In vitro cell cytotoxicity and effect on ICD and PD-L1 expression of Buf-ZIF-lipo-SP94

Buf-ZIF-lipo-SP94 promoted drug uptake in HCC cells, and its cytotoxicity was further studied by CCK-8 assay. It was noteworthy that Buf had a good inhibitory effect and the IC₅₀ of Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 were 0.48 ± 0.09 μM, 0.33 ± 0.18 μM, and 0.27 ± 0.08 μM (calculated in Buf), respectively (Fig. 4A). Buf-ZIF-lipo-SP94 increased the cytotoxicity of Buf by about 3 times, due to effective delivery of the nanoparticles into tumor cells. The fluorescence images of live and dead cells were also verified that Buf-ZIF-lipo-SP94 induced the most cell death (Fig. 4C). And the apoptosis rates of cells receiving different treatments were analyzed. By comparison, the total apoptosis rates of Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 were 11.58 ± 1.23 %, 12.80 ± 0.11 %, 15.92 ± 0.31 % and 23.24 ± 0.22 %, respectively (Fig. 4B and D). As seen above, Buf-ZIF-lipo-SP94 had the best antiproliferative and apoptotic effects on HepG2 cells, which might owe to the adequate cell endocytosis by SP94 and the additive effect of ZIF.

Fig. 4.

The in vitro cytotoxicity and ICD induction of Buf-ZIF-lipo-SP94. (A) The cell viability of HepG2 cells treated with different concentrations of Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94. (B) The quantitative analysis of Annexin V-FITC/PI assay. (C) The representative image of live and dead cells incubated with Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 after 24 h (Concentration of Buf = 0.5 μM). (C) The representative Annexin V-FITC/PI assay of HepG2 cells after treated with Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94 after 24 h (Concentration of Buf = 0.5 μM). (E) The quantification of CALR (CRT) exposure levels in HepG2 cells after the treatments of Buf. (F) The quantification of ATP levels in HepG2 after the treatments of Buf. (G) The quantification of HMGB1 release levels in cells supernatant after the treatments of Buf. (H) The representative flow cytometric plots of mDC co-incubated with HepG2 cells after different treatments. (I) The quantitative analysis of mDC co-incubated with HepG2 cells after different treatments. (J) The PD-L1 expression level in HepG2 cells treated with different concentrations of Buf, Buf-ZIF, Buf-ZIF-lipo and Buf-ZIF-lipo-SP94. (K) The quantification of PD-L1 levels in cells supernatant after the treatments. (∗: p < 0.05, ∗∗: p < 0.01, ∗∗∗: p < 0.001, n = 3).

As expected, ZIF induced more ROS production upon detecting the ROS concentration in HepG2 cells (Fig. S7), which was speculated to have an additive effect of Buf on ICD. It turned out to be the case that Buf-ZIF and Buf-ZIF-lipo-SP94 was better at triggering ICD in HepG2 cells, compared with Buf. Buf elicited the CALR exposure in HepG2 cells (39.37 ± 1.00 %), and a pronounced level of CALR exposure was induced after Buf being loaded into ZIF (43.68 ± 2.26 %) and ZIF-lipo-SP94 (58.43 ± 0.55 %) (Fig. 4E and Fig. S8). As shown in Fig. 4F and G, Buf-ZIF-lipo-SP94 elevated HMGB1 extracellular release and ATP secretion compared with other treatments. Further, the different degrees of ICD-induced DAMP release could eventually influent the discrepant ratio of mDCs (Fig. 4H and I). The result of the proportion of mDCs was control < Buf < Buf-ZIF < Buf-ZIF-lipo < Buf-ZIF-lipo-SP94. The trend of inducing ICD by different nanoparticles was consistent with that of cytotoxicity. Additionally, the PD-L1 expression of HepG2 cells has been found to be regulated to varying degrees, with an order from high to low being control > Buf > Buf-ZIF > Buf-ZIF-lipo > Buf-ZIF-lipo-SP94 (Fig. 4J and K), which showed the ability of Buf-ZIF-lipo-SP94 to improve PD-L1-related tumor immune escape and was more conducive to synergistic efficiency with ICI.

3.6. Safety evaluation and tumor specific targeting of Buf-ZIF-lipo-SP94 in vivo

Before the tail vein administration of nanoparticles, the safety of Buf-ZIF-lipo-SP94 was demonstrated through in vitro hemolysis experiments (Fig. S9), which was below 5 % hemolysis rate, and was considered as a nontoxic level. Because of the potential cardiotoxicity of Buf, the effect of Buf on heart rate in mice was recorded by electrocardiogram examination in vivo [47]. The average heart rates of mice after injected with Buf, Buf-ZIF and Buf-ZIF-lipo-SP94 (Buf = 1 mg/kg) were analyzed at different time points (Fig. 5A and Fig. S10). Due to the use of anesthetics, the heart rate of mice was relatively low. The heart rates were increased by approximately 1.3 (Buf-ZIF) and 1.2 (Buf-ZIF-lipo-SP94) times, respectively at 10 min after treatment. Differently, the heart rate changed sometimes fast and sometimes slow after treated with Buf, which increased to 1.36 times and also decreased to 1.5 time, manifested as arrhythmia. More importantly, the recovery of heart rate in Buf-ZIF and Buf-ZIF-lipo-SP94 groups was also quicker. On the whole, the negative impact of Buf on heart was attenuated remarkably by the versatile nanoparticles [48]. The functionalization of Buf-ZIF-lipo-SP94 increased selectivity of Buf on tumor cells, and little release of Buf to reduce its toxic attack on normal tissue.

Fig. 5.

The electrocardiograms of mice after different treatment and the targeting ability in vivo and tissue biodistribution of Buf-ZIF-lipo-SP94. (A) The electrocardiograms of mice at 10 min after administered with Buf, Buf-ZIF, and Buf-ZIF-lipo-SP94. (B) The fluorescence images of tumor-bearing mice at different times after intravenous injection of control (1), ICG-ZIF-lipo-SP94 (2), ICG-ZIF-lipo (3), ICG-ZIF (4) and ICG (5), respectively. (C) The fluorescence quantitation of tumor in tumor-bearing mice at different time points. (D) The fluorescence quantitation of tissues after tumor-bearing mice sacrificed at 24 h. (E) The fluorescence images of tumor, heart, liver, spleen, lung and kidneys collected after mice sacrificed at 24 h (1: Control, 2: ICG-ZIF-lipo-SP94, 3: ICG-ZIF-lipo, 4: ICG-ZIF, 5: ICG). (∗∗: p < 0.01, ∗∗∗: p < 0.001, n = 3).

The tumor targeting performances in vivo and tissue biodistribution of Buf-ZIF-lipo-SP94 were further evaluated in tumor-bearing mice. Free ICG, ICG-ZIF, ICG-ZIF-lipo and ICG-ZIF-lipo-SP94 were prepared for fluorescence images in vivo. As depicted in Fig. 5B, the IVIS images showed that ICG quickly spread throughout the body, which mostly gathered in the liver and rarely concentrated at the tumor site, and the fluorescence intensity was eliminated rapidly and basically disappeared after 8 h. In comparison, the ICG accumulation of ICG-ZIF and ICG-ZIF-lipo at the tumor site increased and the fluorescence disappeared slower, thanks to the advantages of long circulation in body and pH response in TME of nanocarriers. While ICG-ZIF-lipo-SP94 accumulated more and durably in the tumor tissues, and the fluorescence signal in tumors lasted at least 24 h. The fluorescence quantification intuitively suggested that ICG-ZIF-lipo-SP94 increased drug accumulation in local tumors (Fig. 5C), which effectively demonstrated the advantages of SP94 modification and the better cycling stability of ICG-ZIF-lipo-SP94. Moreover, the fluorescence of major organs and tumors were imaged after 24 h. Finally, the fluorescence intensity of ICG-ZIF, ICG-ZIF-lipo and ICG-ZIF-lipo-SP94 mainly gathered in tumors and livers (Fig. 6D and E), and there was no fluorescence in ICG group. The fluorescence intensity in tumors was ICG-ZIF-lipo-SP94 > ICG-ZIF-lipo > ICG-ZIF, which was as same as in vivo.

Fig. 6.

The in vivo anti-tumor effect with systemic administration of different formulations in Hepa1-6 tumor-bearing mice. (A) The tumor volume growth curves due to in vivo anti-tumor effects of different formulations. (B) The photographs of the stripped tumors from the sacrificed mouse after the last administration (1: Control; 2: Buf; 3: Buf-ZIF; 4: Buf-ZIF-lipo-SP94; 5: Buf-ZIF-lipo-SP94+PD-L1; 6: PD-L1). (C) The tumor weights of the different treatment groups were presented. (D) The body weight of mice in different treatment groups. (E) Survival curve of tumor-bearing mice in different groups. (F) The H&E, Ki67, and TUNEL staining in tumor tissues harvested from different groups (scale bar: 200 μm) (1: Control; 2: Buf; 3: Buf-ZIF; 4: Buf-ZIF-lipo-SP94; 5: Buf-ZIF-lipo-SP94+PD-L1; 6: PD-L1). (G) The quantitate analysis of H&E (a), Ki67 (b), and TUNEL (c) expression in tumor tissues harvested from different groups. (H) The expression of p-PERK, p-eIF2α, ATF-4, CHOP and PD-L1 in tumor tissues harvested from different groups (1: Control; 2: PD-L1; 3: Buf; 4: Buf-ZIF; 5: Buf-ZIF-lipo-SP94; 6: Buf-ZIF-lipo-SP94+PD-L1). (∗: p < 0.05, ∗∗: p < 0.01, ∗∗∗: p < 0.001, n = 3, 5 or 8).

3.7. In vivo anti-tumor efficacy of Buf-ZIF-lipo-SP94

Encouraged by the superior effects of ICD induction and PD-L1 downregulation of Buf-ZIF-lipo-SP94 in vitro, it was hypothesized that Buf-ZIF-lipo-SP94 could induce anti-tumor immunity and might effectively assist immunotherapy in vivo. The potential anti-tumor efficacy of Buf-ZIF-lipo-SP94 and the combination therapy of Buf-ZIF-lipo-SP94 and anti-PD-L1 in vivo were studied simultaneously. The Hepa1-6 tumor-bearing mice were stochastically divided into six groups (5 % glucose solution, Buf, Buf-ZIF, Buf-ZIF-lipo-SP94, Buf-ZIF-lipo-SP94+PD-L1 and PD-L1) and intravenously administrated with 0.5 mg/kg Buf and 6 mg/kg anti-PD-L1 every 2 days for half a month.

The growth trend of tumor volume, tumor morphology, and the tumor weight of each group were evaluated and compared (Fig. 6A–C). Compared with control group, other groups exhibited varying degrees of anti-tumor effects. The effect of anti-PD-L1 was weak, suggesting the low response of anti-PD-L1 in vivo. The slowest growth of tumors was Buf-ZIF-lipo-SP94+PD-L1. As witnessed by the results, tumor weights were consistent with the change in tumor volume, with the tumor inhibition rates of 59.4 % (Buf), 69.7 % (Buf-ZIF), 73.8 % (Buf-ZIF-lipo-SP94) and 93.9 % (Buf-ZIF-lipo-SP94+PD-L1) and 27.9 % (PD-L1), respectively. Definitely, Buf-ZIF and Buf-ZIF-lipo-SP94 potentiated the efficacy of Buf, with the functionalization of accentuating ROS release and specific targeting, based on the analysis of previous experimental results. Anti-PD-L1 alone and Buf-ZIF-lipo-SP94 alone inhibited tumor growth to varying degrees, while their combination elicited most effective and significant delay in tumor growth. The result wrapped up and concluded the synergistic effect and potential advantages of Buf-ZIF-lipo-SP94 in enhancing anti-PD-L1 immune response, worth further in-depth exploration.

The H&E, Ki67, and TUNEL staining of tumor tissue were conducted and showed a similar trend with the anti-tumor effect (Fig. 6F and G). The results showed the diminished level of proliferation and an upgraded level of apoptosis of tumor cells were observed in the Buf-ZIF-lipo-SP94 and Buf-ZIF-lipo-SP94+PD-L1 group, and Ki67 positivity tumor sections in Buf-ZIF-lipo-SP94+PD-L1 group was more outstanding, which indicated a remarkable anti-tumor effect of Buf-ZIF-lipo-SP94 and the combination with anti-PD-L1. Indubitably, Buf-ZIF-lipo-SP94 played a superior effect in anti-tumor and the combination therapy of Buf-ZIF-lipo-SP94 and anti-PD-1 achieved synergistic effect as expected.

The survival rate was also monitored within a month (Fig. 6E). The mice in control group had the MST of 17 days. And the MST of PD-L1 group was 20 days. In contrast, the MST results of other groups related to Buf were all longer than 30 days (including 30 day). Wherein, the survival rates of mice in Buf-ZIF-lipo-SP94 group and Buf-ZIF-lipo-SP94+PD-L1 groups were 69.64 % and 84.41 % on the 30th day, showing effectively anti-tumor of Buf-ZIF-lipo-SP94 and the combination with anti-PD-L1 in HCC to achieve the substantial survival benefit. Additionally, there were no conspicuous fluctuations of body weight in each group (Fig. 6D). H&E staining results of major organs revealed that there was no obvious pathological lesion after treatments (Fig. S11). It also showed in serum enzyme detection that the liver function and renal function indexes including AST, ALT, UREA and CREA were all within the normal ranges (Fig. S12), indicating negligible side effects of Buf-ZIF-lipo-SP94 and the combination with anti-PD-L1. The results aforementioned indicated that Buf-ZIF-lipo-SP94 had remarkable impacts on anti-tumor compared with free Buf and had no obvious systemic toxicity, which established the optimal combination of chemo-immunotherapy to boost anticancer immunity and extend survival in HCC.

3.8. Buf-ZIF-lipo-SP94 induced ICD and regulated PD-L1 expression to activate immune responses in vivo

Encouraged by the satisfactory antitumor results, the mechanism of Buf-ZIF-lipo-SP94 with anti-PD-L1 in anti-tumor immune in vivo was assessed as well by WB, immunohistochemistry and flow cytometer. The altered expressions of ER stress related protein (p-PERK, p-eIF2α, ATF-4 and CHOP) in tumor tissues showed that Buf impacted the ER stress to provide possibilities for inducing ICD in vivo and the effect was enhanced by Buf-ZIF-lipo-SP94 (Fig. 6H and Fig. S13). The ability of Buf-ZIF-lipo-SP94 to downregulate PD-L1 expression in tumor was also validated. These results provided a basis for Buf to achieve synergistic anti-tumor immune effects with ICI in vivo, by inducing ICD and improving PD-L1-related immune escape.

Furthermore, the immunohistochemistry results of CALR exposure and HMGB1 release in tumor tissues were shown in Fig. 7A and B. Buf related groups significantly stimulated the DAMPs excretion in vivo, and Buf-ZIF-lipo-SP94 (53.95 ± 2.00 % of CALR and 55.68 ± 5.93 % of HMGB1) prominently boosted the positive proportion further, compared with Buf (37.03 ± 4.83 of CALR and 35.24 ± 1.35 % of HMGB1) and control (19.55 ± 1.73 % of CALR and 16.70 ± 0.90 % of HMGB1). DC maturation is a crucial link to initiate regulation and maintain adaptive immune response [49,50]. Buf and Buf nanoparticles induced ICD in vivo, and the immune activation performances on DC maturation was evaluated by collecting single-cell suspensions from paracancerous lymph, thymus and spleen. The activated mDCs can further stimulate effector T lymphocytes to improve adaptive immunity in tumor. The major effector T cells (T helper cells, CD3⁺CD4+T cells; cytotoxic T lymphocytes, CD3⁺CD8+T cells) in tumor tissues and spleens of different groups were also stained with corresponding fluorescently labeled antibodies and detected by flow cytometry to assess their tumor infiltration and functions in TME and body.

Fig. 7.

The in vivo ICD induction and regulation of anti-tumor immunity after dealing with different formulations. (A) The immunofluorescence of CALR (CRT) and HMGB1 in tumor tissues harvested from different groups (scale bar: 200 μm). (B) The quantitate analysis of CALR (CRT) (a) and HMGB1 (b) in tumor tissues. (C) The percentages of mDCs in the lymph (a: CD11c + CD80+DC; b: CD11c + CD86+DC). (D) The percentages of mDCs in the thymus (a: CD11c + CD80+DC; b: CD11c + CD86+DC). (E) The percentages of mDCs in the spleen (a: CD11c + CD80+DC; b: CD11c + CD86+DC). (F) The quantitative analyses of the percentages of activated CD8⁺ T cells (a), CD4⁺ T cells (b) in the in the tumor of mice with different administration groups. (G) The quantitative analyses of the percentages of activated CD8⁺ T cells (a), CD4⁺ T cells (b) in the in the spleen of mice with different administration groups. (∗: p < 0.05, ∗∗: p < 0.01, n = 3).

Buf-ZIF-lipo-SP94 promoted the DCs maturation with advantage, and the combination with anti-PD-L1 could further enhance this effect because of the enhanced immune activation which was consistent with the ability of inducing ICD. The percentage of mature DCs in paracancerous lymph distinctively increased after the treatment with Buf-ZIF-lipo-SP94 (CD80 with 0.38 ± 0.11 % and CD86 with 0.47 ± 0.08 %) owing to its strong ICD effect, compared with that in other groups. And on this basis, the maturation of DCs in Buf-ZIF-lipo-SP94+PD-L1 had a level of 1.57 times (CD80⁺) and 2.75 times (CD86⁺) that of the control group (Fig. 7C and Fig. S14A–B). Corresponding to the above results, the prominent enhancement of CD4+T cells and CD8+T cells was recruited into tumor tissues in Buf-ZIF-lipo-SP94 (12.68 ± 0.43 % and 48.79 ± 0.81 %) and Buf-ZIF-lipo-SP94+PD-L1 (13.34 ± 1.90 % and 57.38 ± 4.41 %) groups compared to the control (8.09 ± 0.39 % and 38.59 ± 0.19 %) (Fig. 7F and Fig. S15A–B). Given the finding that the Buf could effectively induce ICD of tumor cells in vivo, the results of activating ICD and establishing immunological effect of Buf-ZIF-lipo-SP94 were more excellent, benefit from the collaboration of ZIF and the HCC specific targeting of SP94 in vivo, which increased effect of anti-PD-1 therapy in Buf-ZIF-lipo-SP94+PD-L1 groups.

Besides, the percentage of mDCs in thymus and spleens were significantly elevated from groups of Buf-ZIF-lipo-SP94 and Buf-ZIF-lipo-SP94+PD-L1, laying the groundwork for the activation of the systemic immunity (Fig. 7D–E and Fig. S14C–F). Correspondingly, the enhancement of CD4+T cells, CD8+T cells in the spleens had the similar trend, which proved an improvement in systemic immunity by Buf nanoparticles and the combination of Buf-ZIF-lipo-SP94 with anti-PD-L1 (Fig. 7G and S15C-D).

In general, Buf-ZIF-lipo-SP94 reversed the immunosuppressive TME and enhancing systemic immunity by driving ICD recruiting and activating DC to recruit cytotoxic CD8⁺ T cells, and finally amplifying anti-tumor immunity response of anti-PD-L1. It was a promising synergism of Buf-ZIF-lipo-SP94, an ICD inducer, and anti-PD-L1 for anti-tumor immunotherapy in HCC.

4. Conclusion

In summary, it was confirmed that Buf was an ICD inducer, which triggered ROS-related ER stress to elicit apoptosis and ICD by PERK/eIF2α/ATF-4/CHOP signaling pathway in the anti-tumor immunity of HCC. Buf promoted the exposure and emission of DAMPs in HCC cells to initiate the uptake and processing of antigens recognized by DCs for subsequent anti-tumor immune responses. The ability to downregulate PD-L1 expression in HCC was also validated to improve PD-L1-related immune escape. The satisfactory effect that Buf enabled opened up a new way of HCC treatment. Considering some side effects of Buf, we developed HCC-specific targeting nanovaccine Buf-ZIF-lipo-SP94, which was safer and more effective to combate HCC. As expected, pH-responsive Buf-ZIF-lipo-SP94 had good tumor accumulation and released drugs in acidic TME. The activity of Buf-mediated ICD to ameliorate anti-tumor immunity and elevate anti-PD-L1 response in TME was effectively enhanced, which achieved synergistic anti-tumor immunity effects of ICD inducer and ICI in HCC treatment. Taken together, this study demonstrated that Buf as a promising ICD inducer could offer an efficient method to enhance tumor immunogenicity for optimizing immunotherapeutic regimens, providing a versatile platform in HCC treatment in the future.

CRediT authorship contribution statement

Zhe Li: Writing – review & editing, Writing – original draft, Methodology. Lixia Chen: Writing – review & editing, Writing – original draft. Ruifeng Zeng: Software, Methodology. Yi Shen: Software. Jinshuai Lan: Funding acquisition. Tong Zhang: Project administration, Funding acquisition. Yue Ding: Project administration, Funding acquisition.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is/are the supplementary data to this article.

Data availability

The data that has been used is confidential.