Mechanism exploration of Zoledronic acid combined with PD-1 in the treatment of hepatocellular carcinoma

Abstract

How to increase the response of immune checkpoint inhibitors (ICIs) is a challenge. In clinical, we found that Zoledronic acid (ZA) may increase the anti-tumor effect of immunotherapy for hepatocellular carcinoma (HCC). To explore the underlying mechanism, we established a mouse model of HCC by subcutaneously injecting Hepa1-6 cell line. The result showed that the tumor volume in the ZA plus anti-PD-1 monocloning antibody (anti-PD-1 mAb) treatment groups was significantly smaller than that of control group, and the onset time of tumor inhibition was even shorter than that of the anti-PD-1 mAb group. Using flow cytometry (FC) to detect the proportion of major immune cell subsets in tumor tissues of each group of mice, we found that the synergistic anti-tumor effect of ZA and anti-PD-1 mAb may be related to ZA-induced polarization of macrophages toward the M1 phenotype. Next, we performed bulk RNA sequencing on tumor samples from different groups to obtain differentially expressed genes (DEGs), which were then input DEGs into pathway enrichment analysis. Data indicated that ZA participated in the M1-type polarization via ferroptosis-related pathways. Our results revealed how ZA involves in the anti-tumor effect of PD-1 monoclonal antibody and provided a potential therapeutic candidate for patients with HCC.

Keywords: Hepatocellular carcinoma; Macrophage; PD-1 antibody; Zoledronic acid.

Fig. 1

The case of combined ZA and anti-PD-1 in the treatment of bone metastasis from liver cancer. A MRI of the primary lesion of the liver cancer before treatment. B The process of treatment. (Antiviral therapy was given within the first month after the diagnosis of primary liver cancer.) The patient received two TACE treatments in the first and second month. At the 6th month, radioknife treatment was used for the patient. Bone metastasis was found at the 10th month. The anti-PD-1 treatments combined with ZA were applied in the 10th, 11th and 12th months, respectively. C MRI images of the liver and left scapula after the patient was found to have bone metastasis. At the 11th month, the left scapular lesion was significantly reduced on the 7th day after treatment. At the 12th month, the lesion in the right lobe of the liver was stable, and the metastasis in the left scapula was significantly reduced. At the 14th month, there was no progression of the intrahepatic lesion, and the extrahepatic metastasis almost disappeared. D, E High AFP and PIVKA were detected and subsequently reaching maximum values when bone metastasis was observed. Both of them decreased gradually after anti-PD-1 combined ZA treatment. F The patient had a high level of CA19-9 at diagnosis. From the beginning to the tenth month, the level of CA19-9 gradually decreased but remained above normal. And the rate of decrease gradually slowed down or even slightly increased. After the third administration of anti-PD-1 + ZA, CA19-9 decreased to the normal range (< 37mAU/ml). G With the application of the combined treatment, CTC and CTM were gradually reduced to 0

Fig. 2

Evaluating the effect of tumor immune infiltration and tumor purity on the drug sensitivity of ZA based on patients with HCC in TCGA and ICGC datasets. A The flowchart of data research. TCGA-LIHC (n = 373), ICGC-LICA-JP (n = 231), and ICGC-LICA-FR (n = 161) were included into analyses. The ESTIMATE score and tumor purity were first calculated. B Boxplots showed the patient with HCC who had higher degree of immune infiltration had significantly higher ESTIMATE score and lower tumor purity in TCGA-LIHC cohort(****, p < 0.0001). C The boxplot showed prediction of the drug sensitivity in TCGA-LIHC cohort. The value of IC50 were lower in stage I and stage II than stage III (*, p < 0.05). D Boxplots showed the values of IC50 were lower in HCC patients with high immunogenicity (TCGA-LIHC ****, p < 0.0001; ICGC-JP ***, p < 0.001; ICGC-FR *p < 0.05)

Fig. 3

The combination of ZA and ICI has synergistic anti-tumor efficacy in vitro. A Schematic diagram of HCC modeling in mice. Mice were defined as ZA-1 and ZA-2 based on different administration frequencies. B The animal experiments were grouped into combined application of ZA and anti-PD-1, only use anti-PD-1 or ZA, and control group. More than ten mice were included into each group. C, D Tumor volumes of mice in each group were monitored (labeled as *, p < 0.05; ***, p < 0.001). E Representative immunohistochemical images of tumor tissues stained with Ki67 from each group. The quantification of Ki67 + cells number per area was counted and calculated (labeled as *, p < 0.05; ***, p < 0.001). F Representative HE staining images of lung, liver, and kidney tissues were taken from each group. G, H Tumor weight and body weight were recorded and compared between each group during the progression of tumor (labeled as *, p < 0.05; **, p < 0.01)

Fig. 4

Changes in the proportion of immune cells analyzed by flow cytometry. A Representative flow cytometric images showed gating strategy and immune cell clustering. B Bar chart showed the proportion of M1 macrophages in CD45 + cells of each group. (labeled as ***, p < 0.001; ****, p < 0.0001). C Bar chart showed the cell proportion of M2 macrophages in CD45 + cells of each group(labeled as **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). D Bar chart showed M1/M2 ratio in each group(labeled as *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). E–H Bar chart showed the cell proportion of T cells, CD4 + T cells, CD8 + T cells, and neutrophils in CD45 + cells (labeled as *, p < 0.05; ***, p < 0.001). I, J Bar chart showed the plasma concentration of IL-18 and IFN-γ in each group (*, p < 0.05)

Fig. 5

Ferroptosis-related pathway leading to the conversion of macrophages Bubble plots of KEGG and GO enrichment analyses. A, B DEGs between the application of ZA group and control group were input into KEGG and GO pathway datasets. Bubble plots showed DEGs were enriched in ferroptosis-related pathway. C, D DEGs between anti-PD-1 group and the combination of ZA and anti-PD-1 group were input into KEGG and GO pathway datasets