Fig 2: The role of NK cells in I/R injury. (A) After I/R treatment on the mouse liver, the proportion of NK cells in the peripheral blood decreased while the proportion of NK cells in the liver tissue increased significantly (n = 6). (B) Multiple immunofluorescences (DAPI + CD161 + CD31). NK cell infiltration increased after reperfusion, especially around the vessels. α‐AsGM1, an inhibitor for NK cells, decreased the infiltration of NK cells. (C) The circulating and hepatic NK cell proportions were significantly reduced using α‐AsGM1. (D) NK cell depletion by α‐AsGM1 can significantly reduce AST and ALT levels (n = 5 ~ 6). **p < 0.01; ***p < 0.001. ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Fig 3: Immunohistochemical analyses of infiltrating immune cells in the stroma of KEAP1-deleted tumors(A–F) Representative images of HE-staining (A, D, G, and J), NQO1-immunostaining (B, E, H, and K), and CD45-immunostaining (C, F, I, and L) of WT (A–C and G–I) and KEAP1-KO (D–F and J–L) 3LL tumors.(G–L) Higher magnification images (G–L) were shown corresponding to the square area in lower magnification images (A–F).(M) Representative images of HE-staining and immunostaining for F4/80, MPO, NK1.1, B220, and CD3 in WT and KEAP1-KO tumors. T, tumor area; S, stroma area; Scale bars, 100 μm.

Fig 4: NRF2 deletion restores the KEAP1-deletion-mediated decrease of immune cell infiltration into tumors(A) Experimental scheme for the bilateral tumor transplantation model of KEAP1-KO and KEAP1-NRF2-DKO cells and the flow cytometry analysis of tumors.(B) Paired plot showing tumor volumes at endpoint in mice injected with both KEAP1-KO and KEAP1-NRF2-DKO cancer cells. Each line connects tumor volumes from the same mouse (n = 8).(C) Percentage of CD45-positive leukocytes in live cells within the KEAP1-KO and KEAP1-NRF2-DKO tumors (n = 4, each).(D) Percentages of NK cell, B cell, macrophage (Mac), neutrophil (Neu), and DC that infiltrated the KEAP1-KO tumors and KEAP1-NRF2-DKO tumors (n = 4, each).(E–P) Representative images of HE-staining (E, H, K, and N), NQO1-immunostaining (F, I, L, and O), and CD45-immunostaining (G, J, M, and P) of KEAP1-KO (E–G and K–M) and DKO (H–J and N–P) tumors. Higher magnification images (K–P) are shown corresponding to the square area in lower magnification images (E–J).(Q) Representative images of HE-staining and immunostaining for F4/80, MPO, NK1.1, B220, and CD3 in KEAP1-KO and DKO tumors. T, Tumor area; S, stroma area; Scale bars, 100 μm. Paired t-test. p < 0.05 was considered statistically significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005. Data are presented as mean, and dots represent individual tumors.

Fig 5: Altered immunophenotype of established liver metastases in TYK2-deficient host mice.(a) Histomorphometric quantification of immunohistochemical GFP staining (as shown in Figure 1c, h) in metastases of TYK2+/+ (+/+), TYK2-/- (-/-), TYK2flox/flox (flox/flox) and TYK2Δhem (Δhem) host mice. The GFP-negative area (stroma and GFP-negative tumor cells) and GFP-positive area (tumor cells) is given in % of the total tumor area. The parameters of the histomorphometry software were adjusted so that tumor areas with weak (tumor weak), medium (tumor medium) and strong (tumor strong) GFP staining intensity could be separated and quantified. All slides were stained in a row and quantified with the same parameters. (b) Histomorphometric quantification of immunohistochemical Ki67 staining of metastases in TYK2+/+ and TYK2-/- host mice, calculated as percentage of positive tumor cells per total tumor cells. (c) Histomorphometric quantification of immunohistochemical Ki67 staining of metastases in TYK2flox/flox and TYK2Δhem host mice, calculated as percentage of positive tumor cells per total tumor cells. (d) Histomorphometric quantification of immunohistochemical BrdU staining of metastases in TYK2+/+ and TYK2-/- host mice, calculated as percentage of positive tumor cells per total tumor cells. (e) Histomorphometric quantification of immunohistochemical BrdU staining of metastases in TYK2flox/flox and TYK2Δhem host mice, calculated as percentage of positive tumor cells per total tumor cells. (f) Immunohistochemical staining for the blood vessel marker endomucin in metastases of TYK2+/+ (+/+) and TYK2-/- (-/-) host mice. Vessels are indicated by arrows. Scale bar = 50 μm. (g) Histomorphometric quantification of blood vessel density in metastases of TYK2+/+ and TYK2-/- host mice (upper bar diagram) as well as TYK2flox/flox and TYK2Δhem host mice (lower bar diagram). (h) Histomorphometric quantification of blood vessel size in metastases of TYK2+/+ and TYK2-/- host mice (upper bar diagram) as well as TYK2flox/flox and TYK2Δhem host mice (lower bar diagram). (i) Immunohistochemical staining for the fibroblast marker PDGFR in metastases of TYK2+/+ (+/+) and TYK2-/- (-) host mice. Fibroblasts are indicated by arrows. Scale bar = 50 μm. (j) Histomorphometric quantification of PDGFR-positive area in metastases of TYK2+/+ and TYK2-/- host mice (upper bar diagram) as well as TYK2flox/flox and TYK2Δhem host mice (lower bar diagram). (k) Immunohistochemical staining of CD8, CD4, F4/80, NK1.1, CD11c and CD19 in metastases of TYK2+/+ (+/+) and TYK2-/- (-/-) host mice. Positive immune cells are indicated by arrows. Scale bar = 50 μm. (l) Histomorphometric quantification of CD8+, CD4+, F4/80+, NK1.1+, CD11c+ and CD19+ immune cells in metastases of TYK2+/+ and TYK2-/- host mice. (m) Histomorphometric quantification of CD8+, CD4+, F4/80+, NK1.1+, CD11c+ and CD19+ immune cells in metastases of TYK2flox/flox and TYK2Δhem host mice. All analyses were performed with mice, 4 weeks after intrasplenic injection of AKP organoids. Bar diagrams represent mean values +/- SEM with each data point representing a mouse. Case Viewer, QuPath and Halo software were used for histomorphometry. Statistical analysis was performed using unpaired Student’s t-test. p values are indicated.