Nanomedicine-mediated induction of immunogenic cell death and prevention of PD-L1 overexpression for enhanced hepatocellular carcinoma therapy

The present study aims to develop a nanoparticle encapsulating doxorubicin (DOX) and programmed death-ligand 1 (PD-L1) siRNA and evaluate its anti-tumor effects on hepatoma carcinoma (HCC). Nanoparticle encapsulating DOX and PD-L1 siRNA (NPDOX/siPD-L1) was characterized by dynamic light scattering and transmission electron microscopy. Flow cytometry was applied to analyze cell populations, NPDOX/siPD-L1 internalization, and cell apoptosis. Real-Time (RT)-quantitative reverse transcription (qPCR) and western blotting were used to determine the mRNA and protein levels, respectively. Released ATP was determined using ATP determination kit and cytokines were determined using specific ELISAs. A tumor-bearing animal model was established to evaluate the anti-tumor effects of NPDOX/siPD-L1. Treatment of NPDOX/siPD-L1 induced immunogenic cell death (ICD) and PD-L1 overexpression in HCC. In vivo study demonstrated that intravenously injection of NPDOX/siPD-L1 significantly inhibited the tumor volume and PD-L1 expressions of tumor tissue in the H22 tumor-bearing animal model. Besides, the treatment of NPDOX/siPD-L1 also regulated the populations of matured dendritic cells and cytotoxic T cells and the productions of cytokines in the tumor tissues. Taken together, NPDOX/siPD-L1 showed significant anti-tumor effects on HCC by the induction of ICD and inhibition of PD-L1 overexpression.

be associated with the occurrence and development of HCC (Murakami et al. 1998). It is difficult to diagnose HCC at the early stages. There are only 30-40% of patients at HCC early stages who were diagnosed. Once HCC develops into the advanced stages, surgical treatment such as transcatheter arterial chemoembolization becomes less effective, accompanied by poor prognosis and a high recurrence rate (Sia et al. 2017;Bruix et al. 2015). Additionally, surgical treatment is not recommended for patients who had extrahepatic metastases. Thus, chemotherapy is the only option for the treatment of HCC in those patients (Bruix et al. 2015;Li and Wang 2016).
Immunogenic cell death (ICD) is one type of cell death that causes an activation of the immune response (Galluzzi et al. 2017;Garg et al. 2015). When ICD occurs in the tumor microenvironment, calreticulin (CRT) is exposed on the surface of tumor cells, thereby stimulating dendritic cells (DCs) to engulf tumor cells (Galluzzi et al. 2017;Vandenabeele et al. 2016;Pitt et al. 2017). In addition, large amounts of adenosine triphosphate (ATP) are released in the tumor tissues, leading to the recruitment of immune cells including monocytes, macrophages, and DCs against tumor-associated antigens (Vandenabeele et al. 2016;Showalter et al. 2017). After that, activated cytotoxic T cells mediate the anti-cancer immune responses. Additionally, extracellular High mobility group box 1 protein (HMGB1) has also been implicated to be associated with ICD (Vandenabeele et al. 2016;Showalter et al. 2017). HMGB1 is released from the dead cells and is able to bind to toll-like receptor 4, thereby promoting matured DCs to present tumor-associated antigen to the T cells (Vandenabeele et al. 2016;Ladoire et al. 2016). These findings encourage us to discover drugs that are able to induce the ICD for cancer therapy.
Many studies have revealed that chemotherapeutic agents including platinum-based drug and doxorubicin (DOX) not only induce cell apoptosis, also trigger ICD (Wong et al. 2015;Fan et al. 2017;Jessup et al. 2019). However, we found that the treatment of DOX enhanced the expressions of programmed death-ligand 1 (PD-L1) in the tumor cells. We inferred that T cell-mediated anti-cancer immune responses were inhibited due to PD-L1 overexpression. To confirm our hypothesis, we evaluated the effects of DOX on the ICD in the PD-L1 knockdown tumor cells. Firstly, we designed a nanoparticle that can encapsulate DOX and PD-L1 siRNA (siPD-L1). Furthermore, to evaluate the anti-tumor effects of the nanoparticle encapsulating DOX and PD-L1 siRNA (NP DOX/ siPD-L1 ) and its underlying mechanisms, we applied HCC cells and tumor-bearing animal models.

Materials and methods
Construction of nanoparticles encapsulating doxorubicin (DOX) and PD-L1 siRNA (NP DOX/ siPD-L1 ) and its characterization NP DOX/siPD-L1 was constructed according to a previously reported method. In brief, block copolymer PEG-PLA (Sigma-Aldrich, St. Louis, MO, USA), cationic lipid DOTAP (Avanti Polar Lipids, Alabaster, AL, USA), and DOX (Sigma-Aldrich) were combined at a ratio of 10:1:1 according to a single emulsion method (Sun et al. 2015). The formed NP DOX was further loaded with PD-L1 siRNA (siPD-L1, Suzhou Ribo Life Science Co., Ltd., Suzhou, China). After that, the gel retardation assay was used to determine the weight ratios of DOTAP to siPD-L1. Dynamic light scattering (DLS, Malvern Zetasizer Nano ZS90) was used to determine particle size and distribution. Transmission electron microscopy (TEM, JEOL JEM2010 200 kV) was used to observe the size and morphology of NP DOX/siPD-L1 .

Cell lines and animals
HCC cell lines including murine H22 and human HepG2 were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in the complete medium supplied with 10% fetal bovine serum, 100 μg/ml streptomycin, and 100 IU/ml penicillin at 37 °C in the presence of 5% CO 2 .
Mouse bone marrow-derived dendritic cells (BMDCs) and human peripheral blood dendritic cells (PBDCs) were isolated and cultured according to previously reported methods (Madaan et al. 2014;Grievink et al. 2016). For BMDCs isolation and culture, bone marrow progenitors were washed out from shin bone and thigh-bone of the mice, then cultured in the medium containing IL-4 (1 ng/ml) and granulocyte-macrophage colony-stimulating factor (10 ng/ml) (PeproTech, Rocky Hill, NJ). After culturing for 48 h, non-adherent cells were gently washed out. The remaining cell clusters were cultured. Medium was changed every other day. On day 7, the cells were collected for further experiments. For PBDCs isolation and culture, PBDCs were purified from nonadherent cells by an iso-osmotic Percoll density gradient, to yield low-density cells. Then, the low-density cells were plated to remove the monocytes. This treatment was repeated. The cells were cultured in the complete culture medium.

Measurement of ATP
After the cells were treated with DOX or nanoparticles for 24 h, the release of ATP was determined using a chemiluminescent ATP determination kit (Life Technologies, Pleasanton, CA, USA), according to the document of the manufacturers.

Flow cytometry
The surface exposure of calreticulin (CRT) was determined using flow cytometry. H22 and HepG2 cells were treated with Doxorubicin (DOX, 0.25 μM) for 24 h and then labeled with anti-CRT-PerCP-Cy5.5 (antibody online. com. Catalog No. ABIN2486728). The percentage of CRT positive cells was quantified based on Propidium iodide (PI) negative cell populations.
To analyze cellular uptake of DOX and PD-L1 siRNA, H22 cells were incubated with indicated antibodies that were labeled with fluorescence Dye. siRNA was labeled with FITC. FITC signal and DOX signal were detected for determining cellular uptake of PD-L1 siRNA and DOX, respectively.
The populations of matured dendritic cells (DCs) and cytotoxic T cells in the tumor tissues were also detected using flow cytometry (BD FACSCalibur, San Jose, CA, USA) according to a previously reported method (Shang et al. 2018). APC-labeled CD80 (Clone: 16-10A1, BioLegend) and PE-labeled CD86 antibodies were applied to determine the population of DCs. APC-labeled CD4 (Clone: GK1.5, BioLegend) and FITClabeled CD8 (Clone:53-6.7, BioLegend) were applied to determine the population of cytotoxic T cells. The results were analyzed using software FlowJo (LLC, Ashland, Oregon, USA).

Detection of cell apoptosis
In the present study, cell apoptosis was determined using flow cytometry. Annexin V/ PI double staining was applied. H22-OVA cells were treated with nanoparticles for 24 h. Next, the treated H22-OVA cells were co-cultured with CD8 + T cells that were isolated from OT-1 mice in 9 mm petri-dish. Cells were then harvested and suspended in the binding buffer containing FITC-labeled Annexin-V and PI followed by incubation in the dark for 15 min. Then the population of apoptotic cells (Annexin-V + PI + ) was measured by flow cytometry and data were analyzed using software FlowJo.

Quantitative Real-Time reverse transcription(qRT)-PCR
RNA extraction kit was used to isolate RNA from the cells, according to the manufacturer's document. Reverse transcriptase was used in the RT reaction. The Melt curves were used to analyze the accuracy. The expressions of each gene were calculated using 2-△△ Ct values. The mRNA expression values of PD-L1 were normalized to that of GAPDH.

ELISAs
The supernatant was collected from different treatment groups. The productions of the cytokines including high mobility group box 1 protein (HMGB1), transforming growth factor (TGF)-β, IL12p70, and interferon (IFN)-γ were determined using specifics ELISAs according to the manufacturers' instruction (DAKEWE, Beijing, China).

Western blot
The protein was extracted according to the previous methods . In brief, a cold RIPA buffer containing protease inhibitor was used to lyse the cells or tumor tissues. After that, the extraction buffer was centrifuged at 13,000 g for 10 min to remove the cell debris and other insoluble materials. The BCA protein assay kits were applied to qualify the concentrations of extracted proteins.
An equal amount of proteins was loaded and separated using the 10% SDS gel. After that, the gel was transfer to a PVDF membrane, which was blocked with 5% non-fat milk at room temperature for 2 h. Next, a primary antibody against S-HMGB1 (ab18256, Abcam) or PD-L1 (ab205921, Abcam) was used to incubate with the membrane at 4 °C overnight. Appropriated secondary antibodies conjugated with HRP were used and the imaging system was applied to qualify the expressions of each target proteins. In the present study, bovine serum albumin (BSA) was used as an internal control for soluble HMGB1(S-HMGB1). β-actin was used as an internal control for PD-L1.

H22 tumor-bearing animal model
Murine H22 cells (1 × 10 7 cells/per mouse) were subcutaneous injected into the C57BL/6 mice. When tumor volumes reached between 50 and 100 mm 3 , the mice were intravenously injected (i.v.) with PBS, NP DOX , NPsiPD-L1, or NP DOX/siPD-L1 at every 3 days. The nanoparticles contained DOX at a dose of 2.5 mg/kg and siPD-L1 at a dose of 2 mg/kg. Tumor volumes were recorded every 3 days. Volumes were calculated using an equation [tumor volume (mm 3 ) = (tumor length) × (tumor width) 2 × 0.5]. At the end of the experimental period, the mice were sacrificed and tumor tissues were collected. The mRNA and protein levels of PD-L1 were measured using qRT-PCR and western blot, respectively. The experimental protocol was supported by the Ethic Commitment of Hangzhou First People's Hospital.

Histopathological analysis
At the end of the experimental period, the mice were sacrificed and the tumor tissues were collected. After the tumor tissues were fixed in 10% formalin solution, the tissues were embedded in paraffin. Hematoxylin and eosin (H&E) staining was performed and the slides were observed under a microscope.

PCNA immunohistochemistry
To analyze the cell proliferation in the tumor tissues, PCNA immunohistochemistry was applied. Paraffin sections were incubated with mouse primary antibody against PCNA followed by incubation with secondary antibody. Avidin-biotin complex procedure was applied. Five randomly chosen sections PCNA immunostained for each group were viewed at a magnification of × 1000 using an image analyzer system.

Statistical analysis
SPSS 13.0 and GraphPad prism were used in this study. Data were shown as mean ± S.D. One-or two-way analysis of variance with multiple comparisons and Student-Newman-Keuls (SNK) test were performed. A p value that less than 0.05 was thought as a statistical significance between the two groups.

Treatment of doxorubicin (DOX) induced immunogenic cell death (ICD) and overexpression of PD-L1 in HCC
We explored the effects of DOX on ICD and the expressions of PD-L1 in H22 and HepG2 cells. As shown in Fig. 1a, b, we investigated the effects of DOX on the expressions of CRT, which is responsible for triggering immune cells to kill cancer cells. The results demonstrated that the percentage of CRT positive cells were significantly increased after the cells were treated with DOX. In addition, the results showed that the treatment of DOX significantly increased the release of HMGB1 and ATP when compared with those in the control group (Fig. 1c-e). These results supported that the treatment of DOX induced ICD.
In addition, we co-cultured DOX-treated HCC cells (H22 or HepG2 cells) with DCs (BMDCs or PBDCs). The results showed that the percentage of matured DCs were significantly increased when the cells were treated with DOX, indicating the treatment of DOX promoted the mature of DCs (Fig. 1f, g). Furthermore, the frequency of PD-L1 positive cells and the mRNA and protein levels of PD-L1 in HCC cells were significantly increased in the DOX-treated groups (Fig. 1h-j).

Construction and characterization of DOX and PD-L1 siRNA encapsulated nanoparticles (NP DOX/siPD-L1 )
A single emulsion method was used to construct NP DOX/siPD-L1 . As shown in Fig. 2a, nanoparticles that were encapsulated with DOX, polyethylene glycol-poly (d,l-lactide) (PEG 5K -PLA 8K ), and cationic lipid DOTAP were mixed at a ratio of 10:1:1 using the emulsion method. Next, the formulated nanoparticles were loaded with PD-L1 siRNA. As shown in Fig. 2b, the gel retardation assay was applied to determine different weight Besides, e released ATP was measured using ATP determination kit. As shown in f, DOX-treated HCC cells were co-incubated with DCs (BMDCs or PBDCs). After 48 h, g flow cytometry was used to determine matured BMDCs (CD45 + CD11c + cells). Additionally, h-j the expressions of PD-L1 in HCC were determined using flow cytometry, RT-qPCR, and western blot, respectively. Data were represented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. All the experiments were repeated three times with similar results ratios of DOTAP and siPD-L1. The NP DOX/siPD-L1 displayed a uniformed size distribution and a round morphology and its average diameter were equal to 120 nm (Fig. 2c). The zeta potentials of NPDOX and NPDOX/siRNA were about 5.0 and 35.0 mV, respectively (Fig. 2d). The selection of the ratio of PEG-PLA, cationic lipid DOTAP, and DOX was determined according to the encapsulation efficacy of DOX and the surface charge of the obtained nanoparticle. In this ratio, the encapsulation efficacy of DOX could reach 80% and the zeta potential of NPDOX was about 35.0 mV, making it possible to adsorb negative siRNA.

NP DOX/siPD-L1 induced ICD of H22 cells and inhibited overexpression of PD-L1
We then determined the effects of NP DOX/siPD-L1 on the ICD and the expressions of PD-L1 in the H22 cells. First, we examined the internalization of NP DOX/siPD-L1 of the cells. The results demonstrated that both DOX and siPD-L1 were internalized into the cells, indicating the NP DOX/siPD-L1 were successfully delivered into the cells (Fig. 3a-c). Next, we determined the effects of NP DOX/siPD-L1 on the ICD by evaluation of the frequency of CRT positive cells, and the release of HMGB1 and ATP. The results demonstrated that treatment of NP DOX/siPD-L1 significantly increased the frequency of CRT positive cells and the release of HMGB1 and ATP when compared with those in the NP siPD-L1 -treated groups (Fig. 3d-f ), indicating that the treatment of NP DOX/siPD-L1 significantly induced the ICD of H22 cells. The weight ratios of DOTAP to siPD-L1 were determined using gel retardation. Besides, c, d TEM image, size distribution, and surface potential of NP DOX/siPD-L1 were also determined. All the experiments were repeated three times with similar results Moreover, to determine the efficiency of PD-L1 siRNA, we measured the mRNA and protein levels of PD-L1 in H22 cells. The results showed that NP siPD-L1 -and NP DOX/siPD-L1 -treated group significantly decreased the mRNA and protein levels of PD-L1 (Fig. 3g), indicating that PD-L1 siRNA were successfully designed and encapsulated. As shown in Fig. 3h, H22-OVA cells were co-cultured with CD8 + T cells that were isolated from OT-1 mice. And then we determined the apoptosis of H22 cells. The results showed that treatment of NP DOX/siPD-L1 promoted the apoptotic cell FITC was used to label siRNA and DOX signal was used to label DOX. Flow cytometry was used to analyze internalization of NP DOX/siPD-L1 internalization. d CRT positive cells were analyzed using Flow cytometry. e, f Released HMGB1 and ATP were determined after the H22 cells were treated with NP DOX/siPD-L1 . g The mRNA and protein levels of PD-L1 were determined using RT-qPCR and western blot, respectively. As shown in the h, treated H22-OVA cell were co-incubated with CD8 + T cells, and the apoptosis of H22 cells was determined. i Annexin V/PI double staining were applied and flow cytometry was used to determine cell apoptosis. Data were represented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. All the experiments were repeated three times with similar results population when compared with those in the NP siPD-L1 -or NP DOX/siNC -treated group (Fig. 3i).

NP DOX/siPD-L1 existed anti-tumor effects on a H22 tumor-bearing animal model
Before we applied NP DOX/siPD-L1 into in vivo study, we firstly determined the stability of NP DOX/siPD-L1 in serum. NP DOX/siPD-L1 was incubated with fetal bovine serum (10%, v/v)containing PBS for 72 h and the size of nanoparticles was detected by DLS. As shown in Additional file 1: Fig. S1, a negligible change was observed within 72 h To confirm the effects of NP DOX/siPD-L1 on in vitro study, an H22 tumor-bearing animal model was established to determine the in vivo anti-tumor effects of NP DOX/siPD-L1 . Figure 4a displays the experimental schematic H22 tumor-bearing animal model and NP DOX/siPD-L1 administration. After subcutaneous injection of H22 cells for 1 week, the nanoparticles were intravenously administrated (i.v.) every 3 days. The results showed that tumor volumes in NP DOX/siPD-L1 treated groups were significant decreased when compared with those in the NP siPD-L1 -or NP DOX/siNC -treated groups (Fig. 4b). Next, we determined the mRNA and protein levels of PD-L1 in tumor tissues. The results showed that the mRNA and protein levels of PD-L1 were significantly decreased in the NP siPD-L1or NP DOX/siPD-L1 treated group (Fig. 4c, d). Furthermore, histology examination and PCNA were performed. The results showed that the treatment of NP DOX/siPD-L1 inhibited cell proliferation in the tumor tissues (Fig. 4e).
Moreover, the present study also investigated the in vivo distribution of NP DOX/siPD-L1 . H22 tumor-bearing mice were intravenously administrated with NP DOX/siPD-L1 (the Fig. 4 NP DOX/siPD-L1 existed anti-tumor effects on an H22 tumor-bearing animal model. a, b C57BL/6 mice were subcutaneously injected with H22 cells. After 1 week, tumor-bearing mice (n = 8) were intravenously injected with PBS, NP DOX , NPsiPD-L1 or NP DOX/siPD-L1 , respectively. Tumor volumes were measured every 3 days (n = 8). At the end of the experiment, c, d the mRNA and protein levels of PD-L1 in tumor tissues were determined using RT-qPCR and western blot, respectively. e Histology examination of tumor tissue was performed using H&E staining. Besides, PCNA analyses were performed to determine the cell proliferation in tumor tissues. Scale bar was 10 μm injection dose of 2.5 mg/kg), mice were sacrificed 24 h later, main organs and tumor tissues were collected and examined using fluorescence in vivo imaging system (IVIS). We found that NP DOX/siPD-L1 efficiently accumulated in tumor tissues, and unavoidably, nanoparticles were also enriched in the spleen, liver, and kidney (Additional file 1: Fig.  S2A, B).

NP DOX/siPD-L1 regulated the populations of matured dendritic cells and cytotoxic T cells and cytokine expression patterns in the tumor tissues
Finally, we investigated the underlying mechanisms of NP DOX/siPD-L1 on its anti-tumor effects. First, we analyzed the percentage of matured dendritic cells and cytotoxic T cells in tumor tissues. The results showed that treatment of NP DOX/siNC or NP DOX/siPD-L1 significantly increased the percentage of matured dendritic cells in tumor tissues. However, the treatment of NP siPD-L1 did not significantly affect the percentage of matured dendritic cells (Fig. 5a, b). Additionally, NP siPD-L1 -treated groups demonstrated a significant increase of the percentage of cytotoxic T cells, and NP DOX/siPD-L1 -treated groups Fig. 5 NP DOX/siPD-L1 regulated the populations of matured dendritic cells and cytotoxic T cells and cytokine expression patterns in the tumor tissues. a-d Flow cytometry was used to analyze CD45 + CD11b + CD11c + DCs and CD8 + cytotoxic T cells. e-g The expressions of TGF-β, IL12p70, and IFN-γ in tumor tissues were determined using ELISA. Data were represented as mean ± SD (n = 8). *p < 0.05, **p < 0.01, ***p < 0.001 displayed an even higher percentage of cytotoxic T cells in tumor tissues as compared to the NP DOX/siPD-L1 -treated groups (Fig. 5c, d).
Moreover, we determined the productions of cytokines including TGF-β, IL12p70, and IFN-γ. The results demonstrated that the productions of TGF-β were significantly decreased in the NP DOX/siNC , NP siPD-L1 , or NP DOX/siPD-L1 -treated groups in comparison to the control group (Fig. 5e). The NP DOX/siPD-L1 -treated groups displayed strong inhibitory effects on the productions of TGF-β. Additionally, the treatment of NP DOX/siPD-L1 also showed significant promotion effects on the productions of IL12p70, and IFN-γ (Fig. 5f,  g). Taken together, these results supported that the treatment of NP DOX/siPD-L1 regulated the populations of matured dendritic cells and cytotoxic T cells and the expressions of cytokines including TGF-β, IL12p70, and IFN-γ in the tumor tissues.

Discussion
In this study, we found that DOX induced ICD in the HCC cells including H22 and HepG2. ICD is characterized by the surface expose of CRT, the release of extracellular HMGB1 and ATP . Our results demonstrated that the treatment of DOX significantly increased the percentage of CRT positive cells, and the release of extracellular HMGB1 and ATP. Additionally, the release of HMGB1 is a late event of ICD, leading to matured DCs to present tumor-associated antigen to the T cells Zhang et al. 2019). We then co-cultured DOX-treated HCC cells with DCs, and the results demonstrated that higher amount of matured DCs in the DOX-treated treated group as compared with those in the control group. These results supported that the treatment of DOX induced ICD in the HCC, which are consistent with a previous finding, in which DOX not only induces cell apoptosis, also triggers ICD (Kawano et al. 2016). However, the treatment of DOX enhanced the expressions of PD-L1 in the tumor cells. It is known that PD-L1 binds to programmed cell death protein 1 (PD-1), a coinhibitor expressed on T cell surface that regulates T cell-mediated immune response (Zou et al. 2016;Roemer et al. 2016). Overexpression of PD-L1 is an obstacle for the treatment of DOX. Therefore, the present study aims to explore the effects of DOX on the ICD when it comes to that PD-L1 was inhibited.
Pegylated liposomal doxorubicin has been approved by the Food and Drug Administration (FDA) in 1995 (Hadjidemetriou et al. 2016). Since then, nanoparticle-based delivery systems for the DOX have drawn much attention in recent years (Jain et al. 2015;Malinovskaya et al. 2017). In the present study, we developed NP DOX/siPD-L1 in two steps. First, we mixed nanoparticles that were encapsulated with DOX, polyethylene glycolpoly (d,l-lactide) (PEG 5K -PLA 8K ), and cationic lipid DOTAP at a ratio of 10: 1: 1 using the emulsion method. Second, we loaded siPD-L1 into the formulated nanoparticles. Our results demonstrated that the NP DOX/siPD-L1 displayed a uniformed size distribution and a round morphology. Additionally, both DOX and siPD-L1 were internalized into the cells, indicating NP DOX/siPD-L1 successfully being delivered the cells.
Overexpression of PD-L1 on the cancer cells promote cancer cell escape from T cellmediated immune response (Tang et al. 2018). The previous study has demonstrated that the blockage of PD-L1 is an effectively adjunctive strategy to boost ICD (Tang et al. 2018). In this study, we observed that the knockdown of PD-L1 did not impact the ICD events including the expressions of CRT, the release of extracellular HMGB1 and ATP.
Interestingly, when we combined DOX and siPD-L1, the results showed that treatment of NP DOX/siPD-L1 significantly increased the frequency of CRT positive cells and the release of HMGB1 and ATP when compared with those in the NP siPD-L1 -treated groups. These results supported the adjunctive role of blocking PD-L1 in the augment of ICD. Furthermore, we co-cultured treated H22-OVA cells with CD8 + T cells and evaluated the effects of NP siPD-L1 and NP DOX/siPD-L1 on the cell apoptosis. CD8 + T cells are also known as cytotoxic T cells, which have been indicated to play a crucial role in cancer cell surveillance. In the present study, we found that the treatment of NP siPD-L1 or NP DOX/siPD-L1 promoted HCC cell apoptosis.
To confirm the anti-tumor effects of NP DOX/siPD-L1 in vitro, an H22 tumor-bearing animal model was established. First, the results showed that the expressions of PD-L1 were significantly decreased in the NP siPD-L1 -or NP DOX/siPD-L1 treated group, indicating that the DOX and siPD-L1 were successfully delivered by nanoparticles. Surprisingly, the results showed a significant decrease in tumor volume in NP DOX/siPD-L1 treated groups. These results are consistent with PCNA analysis, in which showed that cell proliferation in tumor tissues was inhibited in the NP DOX/siPD-L1 treated group. We then explored whether the anti-tumor effects of NP DOX/siPD-L1 were associated with the regulation of the population of immune cells. Two types of cells including matured DCs and CD8 + T cells were analyzed in the tumor tissues. Matured DCs are known to present antigen to the T cells, whilst CD8 + T cells are known to kill cancer cells (Gardner and Ruffell 2016;Wang et al. 2019). Our results showed that NP DOX/siPD-L1 significantly increased the percentage of matured dendritic cells and CD8 + T cells in tumor tissues, indicating NP DOX/ siPD-L1 regulated the populations of matured DCs and CD8 + T cells. Additionally, we also evaluated the productions of classic cytokines including TGF-β, IL12p70, and IFN-γ. It is known that CD8 + T cells secrete cytokines including IFN-γ which have anti-tumor functions Gao et al. 2017). IL12p70, a cytokine produced by DCs, has been reported to trigger the activation of CD8 + T cells by the production of chemokines. (Lu 2017) It has been reported that TGF-β regulates T cell differentiation (Colak and Ten Dijke 2017). The results demonstrated that NP DOX/siPD-L1 -treated groups inhibited the productions of TGF-β and promoted the productions of IL12p70 and IFN-γ. In summary, these results supported that the treatment of NP DOX/siPD-L1 existed antitumor effects by the regulation of the populations of matured DCs and CD8 + T cells and the expressions of cytokines including TGF-β, IL12p70, and IFN-γ.

Conclusion
In the present study, for the first time, we successfully developed a NP DOX/siPD-L1 that is able to simultaneously deliver DOX and PD-L1 siRNA into HCC cells and tumor tissues. Our results demonstrated that the treatment of NP DOX/siPD-L1 significantly induced ICD in the HCC cells. Additionally, in vivo study supported that treatment of NP DOX/siPD-L1 significantly inhibited tumor growth in part by the regulation the populations of matured DCs and cytotoxic T cells and the expressions of cytokines including TGF-β, IL12p70, and IFN-γ.