Targeted IFNγ induction by a genetically engineered Salmonella typhimurium is the key to the liver metastasis inhibition in a mouse model of pancreatic neuroendocrine tumor

Abstract

Background: Liver metastasis is one of the primary causes of death for the patients with pancreatic neuroendocrine tumors (PNETs). However, no curative therapy has been developed so far.

Methods: The anti-tumor efficacy of a genetically engineered tumor-targeting Salmonella typhimurium YB1 was evaluated on a non-functional INR1G9 liver metastasis model. Differential inflammatory factors were screened by Cytometric Bead Array. Antibody depletion assay and liver-targeted AAV2/8 expression vector were used for functional evaluation of the differential inflammatory factors.

Results: We demonstrated that YB1 showed significant anti-tumor efficacy as a monotherapy. Since YB1 cannot infect INR1G9 cells, its anti-tumor effect was possibly due to the modulation of the tumor immune microenvironment. Two inflammatory factors IFNγ and CCL2 were elevated in the liver after YB1 administration, but only IFNγ was found to be responsible for the anti-tumor effect. Liver-targeted expression of IFNγ caused the activation of macrophages and NK cells, and reproduced the therapeutic effect of YB1 on liver metastasis.

Conclusion: We demonstrated that YB1 may exhibit anti-tumor effect mainly based on IFNγ induction. Targeted IFNγ therapy can replace YB1 for treating liver metastasis of PNETs.

Keywords: Salmonella typhimurium; YB1; liver metastasis; pancreatic neuroendocrine tumor; targeted IFNγ therapy.

Figure 1

YB1 inhibited liver metastasis without infecting tumor cells in INR1G9 mouse models. (A) Representative pictures of the livers and spleens from mice inoculated with INR1G9 cells in the spleen and injected with PBS (1G9 + PBS) or YB1 (1G9 + YB1) through the tail veins (n = 10). Scale bar: 500 μm. (B–D) The number of liver metastases (B), size of liver metastases (C) and the length of the spleen (D) was measured in the 1G9 + PBS and 1G9 + YB1 groups. (E–G) H&E staining of the liver sections of the mice inoculated with INR1G9 cells in the spleen and injected with PBS (E) or YB1 (F,G) through the tail veins. (H) Statistical analysis of the proportions of tumor area as represented in (E–G) (n = 10). (I) H&E staining of the liver sections of the mice inoculated with INR1G9 cells in the spleen and injected with YB1 through the tail veins. Red arrows: calcification foci. (J,K) The mice liver sections from the 1G9 + YB1 group were immunostained with anti-salmonella antibodies. Green arrows: INR1G9 metastatic tumors. (L) Quantitative analysis of YB1 foci in the liver or tumor areas as represented in (J,K). Five immunostaining areas were randomly selected for normal liver, non-necrotic tumor and necrotic tumor, respectively. Scale bar in (E–G) and (J,K): 500 μm; scale bar in (I): 50 μm. Data are presented as mean ± SD. **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 2

IFNγ and CCL2 levels were elevated in INR1G9 inoculated mice livers after YB1 treatment with M1 macrophages recruited. (A–F) Mice were injected with PBS or YB1 through the tail vein following intrasplenic injection of culture medium or INR1G9 cells (n = 5). The levels of inflammatory factors including IL12 p70 (A), TNF (B), IFNγ (C), CCL2 (D), IL10 (E), and IL6 (F) in the mice livers were measured using cytometric bead array. (G,H) H&E staining of the liver sections from the mice injected with PBS (G) or YB1 (H) through the tail vein and kept for one week (n = 5). (I) Quantitative analysis of infiltration foci in the livers of PBS or YB1 injected mice as represented in (G,H) (n = 5). (J,K) Immunohistochemistry of the liver sections in (I) with anti-iNOS (J) or anti-CD163 antibodies (K) to show M1 or M2 macrophages, respectively. (L) Quantitative analysis of the percentages of immunostaining positive cells in the infiltrated cells as represented in (J,K) (n = 5). Data are presented as mean ± SD. Ns, non-significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 3

Anti-IFNγ but not anti-CCL2 abrogated the anti-tumoral effect of YB1. (A–F) H&E staining of the liver sections from the control mice (A) or mice with INR1G9 inoculated (B), INR1G9 inoculated and treated with YB1 (C), INR1G9 inoculated, treated with YB1 and anti-IFNγ (D), anti-CCL2 (E) or anti-IFNγ plus anti-CCL2 (F). Scale bar: 500 μm. (G) Statistical analysis of the proportions of tumor area as represented in (A–F) (n = 5). Data are presented as mean ± SD. Ns, non-significant; *p < 0.05; **p < 0.01.

Figure 4

Liver-targeted expression based on AAV2/8 vectors. (A) Schematical diagrams of liver-targeted AAV2/8 vector containing the IFNγ-expressing cassette and packaging plasmids. (B) Green fluorescence detected in the liver, pancreas, stomach, colon and spleen from mice injected with AAV-EGFP or AAV vector for one week. (C) Relative IFNγ levels measured by ELISA in the livers of mice injected with PBS, AAV vector or AAV-IFNr (n = 3). Data are presented as mean ± SD. **p < 0.01.

Figure 5

AAV-IFNr completely inhibited the liver metastases of INR1G9 cells. (A) Representative pictures of the livers from mice inoculated with INR1G9 cells in the spleen and injected with PBS, AAV vector or AAV-IFNr through the tail veins (n = 5). (B) Statistical analysis of the number of liver metastases in the PBS, AAV vector and AAV-IFNr groups. Data are presented as mean ± SD. ****p < 0.0001.

Figure 6

AAV-IFNr activated macrophages and NK cells in the liver. Detection of iNOS (A–C), NK1.1 (E–G) or F4/80 (I–K) by immunofluorescence in the liver of mice (n = 3) injected with PBS (A,E,I), AAV vector (B,F,J) or AAV-IFNr (C,G,K) through tail vein for two weeks. Scale bar: 100 μm. The statistical analyses were presented as mean ± SD for iNOS (D), NK1.1 (H) and F4/80 (L) as indicated. ****p < 0.0001. ns, non-significant.