Enhanced protection of C57 BL/6 vs Balb/c mice to melanoma liver metastasis is mediated by NK cells

Abstract

The B16F10 murine melanoma cell line displays a low expression of MHC class I molecules favoring immune evasion and metastases in immunocompetent C57 BL/6 wild-type mice. Here, we generated metastases to the liver, an organ that is skewed towards immune tolerance, by intrasplenic injection of B16F10 cells in syngeneic C57 BL/6 compared to allogeneic Balb/c mice. Surprisingly, Balb/c mice, which usually display a pronounced M2 macrophage and Th2 T cell polarization, were ∼3 times more susceptible to metastasis than C57 BL/6 mice, despite a much higher M1 and Th1 T cell immune response. The anti-metastatic advantage of C57 BL/6 mice could be attributed to a more potent NK-cell mediated cytotoxicity against B16F10 cells. Our findings highlight the role of NK cells in innate anti-tumor immunity in the context of the liver - particularly against highly aggressive MHC I-deficient cancer cells. Moreover, the B16F10 model of melanoma liver metastasis is suited for developing novel therapies targeting innate NK cell related immunity in liver metastases and liver cancer.

Keywords: NK cells; cancer immunology; innate immunity; liver; macrophages; metastasis.

Figure 1.

Balb/c mice are highly susceptible to B16F10 liver metastasis compared to C57 BL/6 mice. (A-C) Balb/c mice (Bc) received intrasplenic injections of 100,000 B16F10luc cells (“100 k”), whereas C57 BL/6 mice (B6) received either 100,000 or 300,000 cells (“100 k” or “300 k”). Control animals were treated equally but received injections of PBS. Bioluminescence was recorded 7 and 14 days after intrasplenic injection of B16F10luc cells (A). Livers were harvested 14 days after injection of cells to assess metastatic spread (representative images are shown) (B). Biometric data (liver weight, liver to body weight ratio, number of metastatic nodules on the liver surface with 250 being the upper limit) were recorded and analyzed (C). (D-F) 100,000 B16F10luc cells (“100 k”) were injected in the tail vains of Balb/c and C57 BL/6 mice. Control animals were treated equally but received injections of PBS. Bioluminescence was recorded 7 and 14 days after intravenous injection of B16F10luc cells (D). Lungs were harvested 14 days after injection of cells and photographs were taken (representative images are shown) (E). Biometric data (lung weight, lung to body weight ratio, number of metastatic nodules on the lung surface with 250 being the highest number) were recorded and analyzed (F). Individual and mean values ± SD are shown. The number of samples per group is indicated above the corresponding column. One-way ANOVA; statistical significance was assessed using either the Sidak's multiple comparisons test or in case of metastatic nodules the Mann-Whitney test with Bonferroni correction.

Figure 2.

Murine livers with B16F10 metastases show strong myeloid and T cell immune cell infiltration. (A-C) Sections of metastasized livers from Balb/c and C57 BL/6 mice harvested 14 days after intrasplenic injection of B16F10 cells were stained with HE or antibodies against CD68 or CD3. Bars represent 200 µm (A) and 100 µm (B-C), respectively. Metastatic nodules are highlighted by white arrows. (D) Non-parenchymal cells were isolated from metastasized livers, dead cells excluded and the frequency of CD45⁺ cells determined by FACS. Individual and mean values ± SD are shown (n = 5 per group). (E) The number of cells per mm² in sections stained with antibodies against CD68 or CD3 was quantified. Individual and mean values ± SD are shown (n ≥ 4 per group).

Figure 3.

Balb/c but not C57 BL/6 mice launch a M1-polarized innate and T cell-dependent adaptive immune response against B16F10 liver metastases. (A) Non-parenchymal cells were isolated from metastasized livers from Balb/c and C57 BL/6 mice 14 days after intrasplenic injection of B16F10 cells, and the frequencies of macrophages (CD45⁺/ CD11b⁺/ F4/80⁺ cells), dendritic cells (CD45⁺/ CD11b⁻/ CD11 c⁺/ F4/80⁻), T killer cells (CD45⁺/CD8⁺), T helper cells (CD45⁺/CD4⁺) and B cells (CD45⁺/CD19⁺) as well as the strength of Ly6 c-expression on macrophages (MFI = mean fluorescence intensity) were determined by FACS. In addition, sections of metastasized livers were stained with the NK cell marker NKp46, and the number of positively stained cells per mm² were quantified. Individual and mean values ± SD are shown (n = 5 per group). (B) Expression of selected immune genes (grouped for function or dominant immune cell type) in metastatic livers of Balb/c vs C57 BL/6 mice as determined by RNA-Seq (n = 3 in each group). Gene expression is normalized to Reads Per Kilobase per Million mapped reads (RPKM). For better visualization the y-axis is log₁₀-scaled. False discovery rates (FDR) smaller than 0.01 for the differential gene expression between Balb/c and C57 BL/6 mice are indicated above the corresponding boxes.

Figure 4.

Analysis of the transcriptome maps the distinct differences in the immune responses against B16F10 metastases between Balb/c and C57 BL/6 mice. (A) MA-plot and numbers of differentially expressed genes between livers from Balb/c and C57 BL/6 mice with B16F10 metastasis (genes with log₂-fold change (log₂ FC ≥ 1 or ≤ −1) and false discovery rate (FDR) ≤ 0.01 are highlighted in red). (B-C) An IPA® was performed on genes differentially expressed between Balb/c and C57 BL/6 metastasized livers (log₂ FC > 1 or < −1). The top regulators identified for each strain along with the corresponding p-value are listed (B). Selected immune-related signaling pathways which were found to be significantly enriched in the data set by IPA® are shown with significantly different activation levels between Balb/c and C57 BL/6 mice (C; the bars represent a multiple testing corrected p value using the Benjamini-Hochberg method; the ratio value [depicted by the yellow line with boxes] tells the proportion of molecules in a given pathway that meet the cutoff criteria).

Figure 5.

Cytotoxic NK cells endow C57 BL/6 mice with immunity against B16F10 cells compared to Balb/c mice. (A) Sections of metastasized livers and spleens from Balb/c, C57 BL/6 and NK cell depleted (NKD) C57 BL/6 mice were stained with an antibody against NKp46. Spleens were harvested 10 minutes and livers 14 days after intrasplenic injection of B16F10 cells. Bars represent 50 µm. (B-D) Mice were treated as in Fig. 1 except that C57 BL/6 mice received intraperitoneal injections of anti-asialo GM1 antibody for NK cell-depletion (NKD) or injections of control IgG. Bioluminescence was recorded 14 days after intrasplenic injection of B16F10luc cells (A), followed by harvesting of livers for macroscopic metastasis assessment (B; representative images). Biometric data (liver weight, liver to body weight ratio, number of metastatic nodules on the liver surface) were recorded and analyzed (C). Individual and mean values ± SD are shown (the number of mice per group is indicated above the corresponding column). One-way ANOVA; statistical significance was assessed using either the Sidak's multiple comparisons test or in case of metastatic nodules the Mann-Whitney test with Bonferroni correction. (E) NK cells were isolated from the spleens of Balb/c and C57 BL/6 mice, respectively, by using indirect magnetic labeling with anti-NKp46 microbeads and subsequent magnetic separation, and incubated with B16F10 cells labeled with eFluor 670 in vitro. Cells were analyzed by FACS, and cytotoxicity was determined by measuring the extent of 7-AAD staining of B16F10 cells normalized to control cultures without effector cells. Mean values ± SD of the frequency of dead cells and of the MFI are shown (n = 4; MFI = mean fluorescent intensity). Two-way ANOVA; statistical significance was assessed using the Sidak's multiple comparisons test.