Fig 2: Cyclin G2 knockout in macrophages attenuates the inhibitory effects of IFN-γ on colon cancer cell growth. A–C MC38 cells were mixed with BMDMs from WT and Ccng2−/− C57BL/6 mice at a ratio of 5:1 and injected subcutaneously into C57BL/6 mice, which were then treated with IFN-γ at specific times. Gross tumors (A), tumor weights (B), and tumor volumes (C) were measured at the endpoint. Data were analyzed with the unpaired Student’s t-test. Data are presented as the mean ± SEM (n = 5). D A schematic model depicting the role of cyclin G2 in macrophages after IFN-γ treatment. Upregulated cyclin G2 after IFN-γ treatment inhibited the interaction between PP2Ac and STAT1, thereby increasing the nuclear import of STAT1 and promoting CXCL9 transcription. Increased CXCL9 secretion can promote CTL chemotaxis and inhibit vascular endothelial cell angiogenesis, ultimately inhibiting tumor progression. **p < 0.01; ****p < 0.0001

Fig 3: Cyclin G2 in macrophages regulates CTL chemotaxis and vascular endothelial cell tube formation via CXCL9. A CXCL9 levels in the supernatants of BMDMs from WT and Ccng2−/− C57BL/6 mice treated with IFN-γ were determined by ELISA (representing 3 independent experiments). B, C CXCL9 levels in the supernatants of THP-1 stable cell lines (Nonsense, shcyclin G2#1, and shcyclin G2#2 and Vector and Flag-cyclin G2) treated with IFN-γ were determined by ELISA (representing 3 independent experiments). D, E CTL chemotaxis analyzed by treating conditioned medium from BMDMs isolated from WT and Ccng2−/− C57BL/6 mice treated with or without recombinant CXCL9. Scale bar = 200 μm (representing 3 independent experiments). F Tube formation experiments showed the tube formation ability of SVEC4–10 cells treated with conditioned medium from BMDMs isolated from Ccng2−/− C57BL/6 mice. The recombinant CXCL9 was added or not added to the conditioned medium. Scale bar = 500 μm (representing 3 independent experiments). G Tube formation experiments showed the tube formation ability of HUVECs treated with conditioned medium from a THP-1 stable cell line (shcyclin G2#1). The recombinant CXCL9 was added or not added to the conditioned medium. Scale bar = 200 μm (representing 3 independent experiments). (A–C, E–G) Data were analyzed with the unpaired Student’s t-test. Data are presented as the mean ± SD **p < 0.01; ***p < 0.001; ns, not significant

Fig 4: The CXCL9-CXCR3 axis is crucial for PEx-mediated protection against surgical stress-promoted CRLM(A) Schematic diagram illustrating the surgical stress-promoted CRLM models in Sed and PEx Cxcl9 KO mice.(B and C) Representative images of liver metastatic tumors (B) and quantification of the tumor burdens in Sed and PEx Cxcl9 KO mice (n = 6/group) (C).(D) Schematic diagram illustrating the surgical stress-promoted CRLM models in Sed and PEx Cxcr3 KO mice.(E and F) Representative images of liver metastatic tumors (E) and quantification of the tumor burdens (F) in Sed and PEx Cxcr3 KO mice (n = 6/group).(G) Schematic diagram showing the adoptive transfer of PEx WT KCs and Cxcl9 KO KCs into DT-administered Sed or PEx KC-DTR mice, which then underwent the surgical stress-promoted CRLM model.(H and I) Representative images of liver metastatic tumors (H), and tumor burdens (I) in recipient Sed or PEx KC-DTR mice that received WT or Cxcl9 KO KCs (n = 5/group).(J) Schematic diagram showing surgical stress-promoted CRLM model in adoptive transfer of Sed WT or Cxcr3 KO CD8+ T cells into CD8+ T cell-depleted Sed or PEx WT mice.(K and L) Representative images of liver metastatic tumors (K), and tumor burdens (L) in Sed or PEx WT recipient mice that received WT or Cxcr3 KO CD8+ T cells (n = 5/group).(M) Schematic diagram of surgical stress-promoted CRLM models in Sed and PEx WT mice treated with IgG isotype control or anti-CXCL9 antibody (Ab).(N and O) Representative images of liver metastatic tumors (N), and tumor burdens (O) in Sed and PEx WT mice treated with anti-CXCL9 Ab or IgG isotype control (n = 5/group).(P) Schematic diagram of the surgical stress-promoted CRLM model in Sed and PEx WT mice treated with normal control (1% DMSO in saline) and CXCR3 antagonist (AMG487) after surgery until 3 weeks.(Q and R) Representative images of liver metastatic tumors (Q), and tumor burdens (R) in Sed and PEx WT mice treated with normal control and AMG487 (n = 5/group). All experiments: error bars represent mean ± SEM, statistical analysis by Student’s t test. ns, not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. See also Figures S4–S6.

Fig 5: CXCL9 enhances anti-tumor activity of CXCR3+ CD8+ T cells(A and B) Hepatic CD8+ T cells from WT mice were stimulated with serum from Sed or PEx WT mice for 24 h (A). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (B) (n = 5/group).(C and D) Hepatic CD8+ T cells from WT mice were stimulated with serum from Sed and PEx Cxcl9 KO mice for 24 h (C). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (D) (n = 5/group).(E and F) Cxcr3 KO hepatic CD8+ T cells were stimulated with serum from Sed and PEx WT mice for 24 h (E). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (F) (n = 5/group).(G and H) Tumor CD8+ T cells from WT mice were stimulated with serum from Sed and PEx WT mice for 24 h (G). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (H) (n = 5/group).(I and J) Tumor CD8+ T cells from WT mice were stimulated with serum from Sed and PEx Cxcl9 KO mice for 24 h (I). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (J) (n = 5/group).(K and L) Tumor Cxcr3 KO CD8+ T cells were stimulated with serum from Sed and PEx WT mice for 24 h (K). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (L) (n = 5/group).(M and N) Peripheral CD8+ T cells from healthy volunteers were stimulated with serum from the same CRLM patient before and after exercise for 24 h (M). ELISA analysis of Granzyme B, IFN-γ, and TNF-α levels in the medium (N) (n = 5/group). All experiments: error bars represent mean ± SEM, statistical analysis by Student’s t test. ns, not significant, ∗p < 0.05, ∗∗p < 0.01. See also Figure S7.