Cancer Cell Secreted Legumain Promotes Gastric Cancer Resistance to Anti-PD-1 Immunotherapy by Enhancing Macrophage M2 Polarization

Abstract

The interaction between cancer cells and immune cells plays critical roles in gastric cancer (GC) progression and immune evasion. Forced legumain (LGMN) is one of the characteristics correlated with poor prognosis in gastric cancer patients. However, the role of gastric-cancer-secreted LGMN (sLGMN) in modulating the tumor immune microenvironment and the biological effect on the immune evasion of gastric cancer remains unclear. In this study, we found that forced expression of sLGMN in gastric cancer serum correlates with increased M2 macrophage infiltration in GC tissues and predicted resistance to anti-PD-1 immunotherapy. Mechanistically, gastric cancer cells secrete LGMN via binding to cell surface Integrin αvβ3, then activate Integrin αvβ3/PI3K (Phosphatidylinositol-4,5-bisphosphate3-kinase)/AKT (serine/threonine kinase)/mTORC2 (mammalian target of rapamycin complex 2) signaling, promote metabolic reprogramming, and polarize macrophages from the M1 to the M2 phenotype. Either blocking LGMN, Integrin αv, or knocking out Integrin αv expression and abolishing the LGMN/Integrin αvβ3 interaction significantly inhibits metabolic reprogramming and polarizes macrophages from the M1 to the M2 phenotype. This study reveals a critical molecular crosstalk between gastric cancer cells and macrophages through the sLGMN/Integrinαvβ3/PI3K/AKT/mTORC2 axis in promoting gastric cancer immune evasion and resistance to anti-PD-1 immunotherapy, indicating that the sLGMN/Integrinαvβ3/PI3K/AKT/mTORC2 axis may act as a promising therapeutic target.

Keywords: gastric cancer; immunotherapy; legumain; macrophages.

Figure 1

Relationship between LGMN expression levels and M2 macrophage infiltration in gastric cancer. (A) Analysis of TCGA database suggests a correlation between LGMN expression levels in gastric cancer and M2 macrophage polarization. (B) Differential serum LGMN expression in patients sensitive and resistant to anti-PD-1 therapy. (C,D) Immunohistochemical detection of M1 or M2 macrophage infiltration in gastric cancer tissues, and their correlation with serum LGMN expression levels in gastric cancer patients (bar 100 μm). * p < 0.05; ** p < 0.01.

Figure 2

LGMN induces polarization of macrophages from M1 to M2 phenotype. (A) Co-culture of THP-1-derived M1 macrophages with ASG-L and BGC823-L cells followed by flow cytometry to detect expression of M2 macrophage markers CD163 and CD206, and the polarization of M1 macrophages towards M2 macrophages can be blocked by LGMN-neutralizing antibodies. (B) Incubation of THP-1-derived M1 macrophages with rh-LGMN protein for 5 days followed by flow cytometry to detect expression of M2 macrophage markers CD163 and CD206. (C) Stimulation of healthy human PBMCs with IFN-γ followed by flow cytometry to detect expression of CD80 and CD86. (D) Induction of M1 macrophages from PBMCs by culture with rh-LGMN, followed by flow cytometry to detect expression of CD206 and CD163.

Figure 3

LGMN induces polarization of macrophages from M1 to M2 phenotype by forming a complex with Integrin αvβ3. (A) Western blot analysis of the effect of rh-LGMN, rh-LGMN + Integrin αv antibody, and rh-LGMN + Integrin β3 antibody on mTORC2 activity in M1 macrophages. (B) Flow cytometry analysis of the effect of rh-LGMN, rh-LGMN + Integrin αv antibody, and rh-LGMN + Integrin β3 antibody on the polarization of macrophages from M1 to M2 phenotype. (C) Establishment of M1 macrophage cells derived from THP-1 cells with knockdown of Integrin β3 using lentivirus. (D) Flow cytometry analysis of the effect of knocking down Integrin β3 expression on LGMN-mediated polarize macrophages from M1 to M2 phenotype. (E) Stimulation of M1 macrophages derived from PBMCs with rh-LGMN protein for 48 h followed by Western blot analysis to detect changes in mTORC2 signaling pathway activity. (F) Stimulation of M1 macrophages derived from healthy human PBMCs with rh-LGMN for 48 h followed by flow cytometry to detect changes in CD36 expression.

Figure 4

mTOR inhibitor blocks LGMN-induced polarization of macrophages from M1 to M2 phenotype. (A) Flow cytometry analysis of the effect of AZD2014 on LGMN-mediated polarization of macrophages from M1 to M2 phenotype. (B) Western blot analysis of the effect of AZD2014 on upregulation of mTORC2 signaling pathway activity by rh-LGMN in M1 macrophages. (C) Western blot analysis of the effect of rh-LGMN on mTORC1 signaling pathway activity in M1 macrophages.

Figure 5

rh-LGMN protein promotes glycolysis and fatty acid oxidation in M1 macrophages derived from healthy human PBMCs. (A) ECAR assay. (B) Fatty acid oxidation assay. (C) Oxygen consumption assay. (D) Fatty acid uptake assay. (E,F) Low-dose mTOR inhibitor AZD2014 inhibits rh-LGMN-induced glycolysis and oxidative phosphorylation in M1 macrophages. (G,H) Blocking antibodies against Integrin αv but not Integrin β3 inhibits rh-LGMN protein-induced glycolysis (G) and oxygen consumption (H) in M1 macrophages derived from healthy human PBMCs. * p < 0.05; ** p < 0.01; *** p < 0.001 ((G,H): LGMN vs. LGMN + Integrin αv).

Figure 6

sLGMN induces resistance of gastric cancer to anti-PD-1 therapy. (A) Relationship between serum LGMN levels and PFS of gastric cancer to anti-PD-1 therapy. (B) Establishment of 615 mouse gastric cancer model using MCF cells, followed by different therapies to observe the relationship between sLGMN and sensitivity of gastric cancer to PD-1 mAb. (C) Statistical analysis of tumor volume after different treatments in 615 mouse MFC cell gastric cancer model. ** p < 0.01; *** p < 0.001; NS, not significant.