BACH2 restricts NK cell maturation and function, limiting immunity to cancer metastasis

Abstract

Natural killer (NK) cells are critical to immune surveillance against infections and cancer. Their role in immune surveillance requires that NK cells are present within tissues in a quiescent state. Mechanisms by which NK cells remain quiescent in tissues are incompletely elucidated. The transcriptional repressor BACH2 plays a critical role within the adaptive immune system, but its function within innate lymphocytes has been unclear. Here, we show that BACH2 acts as an intrinsic negative regulator of NK cell maturation and function. BACH2 is expressed within developing and mature NK cells and promotes the maintenance of immature NK cells by restricting their maturation in the presence of weak stimulatory signals. Loss of BACH2 within NK cells results in accumulation of activated NK cells with unrestrained cytotoxic function within tissues, which mediate augmented immune surveillance to pulmonary cancer metastasis. These findings establish a critical function of BACH2 as a global negative regulator of innate cytotoxic function and tumor immune surveillance by NK cells.

Figure 1.

Expression of Bach2 in developing and mature NK cells. (A) Flow cytometry of Bach2^tdRFP expression by NK cell progenitors within BM (left) and replicate measurements (right). (B and C) Gating strategy for progressive NK cell maturation states from the BM (B) and spleen (C). (D) Bach2^tdRFP expression (left) and replicate measurements (right) in NK cell subsets defined in B from mice of indicated genotypes. (E) Bach2^tdRFP expression (left) and replicate measurements (right) in the NK cell subsets defined in C from mice of indicated genotypes. (F) Immunoblot analysis of BACH2 expression in NK cell subsets sorted from Bach2^Flag animals using an anti-Flag antibody. Numbers show relative expression when compared to an anti–β-actin control. (G) Flow cytometry of Bach2^Flag expression in subsets of NK cells from the spleen (left) and replicate measurements (right). Bach2^tdRFP expression is plotted as ΔMFI (mean fluorescence intensity), calculated as the difference between RFP signal in samples from Bach2^tdRFP/+ animals compared to the mean RFP signal from WT controls (D–E). BACH2^Flag expression is plotted as ΔMFI, calculated as the difference between anti-Flag signal in Bach2^Flag animals and the mean WT control signal. Data representative of two independent experiments with four to six mice per group (A–E and G). ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. Ordinary one-way ANOVA (D, E, and G). Bars and error are mean and SEM.

Figure S1.

Bach2 deficiency in innate lymphocytes is not associated with a lethal inflammatory phenotype. (A) Genomic tracks depict histograms of normalized reads (y axis) plotted by genome position (x axis) near and within the Bach2 locus for BM-derived CD27⁺ CD11b⁻ and CD27⁻ CD11b⁺ NK cells (GSE109517; Zook et al., 2018). Yellow boxes highlight differentially accessible peak regions. (B) Percentages of NK cells of indicated phenotypes within NK cells from blood samples of WT and Bach2^−/− mice. (C) Representative histograms (left) and replicate measurements (right) of KLRG1 expression on NK cells in blood samples from mice of indicated genotypes. Numbers in gates indicate percentages. (D) Kaplan-Meier plot showing survival of WT, Bach2^−/−, Rag2^−/−, and Rag2^−/− Bach2^−/− animals. Logrank test; P = 0.0069. (E) Representative histological images of lungs taken from aged animals of indicated genotypes. Scale bar = 100 µm. (F) Replicate measurements of percentage alveolar involvement of lungs, as shown in E. (G) Histopathology scoring of gut samples taken from aged animals of indicated genotypes. ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. Data representative of two independent experiments with six to eight mice per group (B–F). Unpaired two-tailed Student’s t test (B and C). Ordinary one-way ANOVA (F). Kruskal–Wallis test (G). Bars and error are mean and SEM.

Figure S2.

Regulation of NK cell maturation by BACH2. (A) Representative plots of CD27 and CD11b expression (top) and percentages of NK cells of indicated phenotypes (bottom) within NK cells from spleens of Rag^−/− and Rag^−/− Bach2^−/− mice. (B) Representative histograms (left) and replicate measurements (right) of KLRG1 expression on splenic NK cells from mice of indicated genotypes. (C) Representative histograms (left) and replicate measurements (right) of KLRG1 expression on lung NK cells from mice of indicated genotypes. (D) Representative Granzyme B expression (MFI) in splenic NK cells (left) and replicate measurements in spleen and lung NK cells (right) from mice of indicated genotypes. (E) Percentage of NK cells of indicated phenotypes (left) and KLRG1 expression on NK cells (right) within the CD45.1⁺ CD45.2⁺ NK compartments from the spleens of WT:WT (black dots) or Bach2^−/−:WT (red dots) chimeras. Numbers in gates indicate percentages. Data representative of two independent experiments with five to six mice per group. ns, not significant (P > 0.05); **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. Unpaired two-tailed Student’s t test. Bars and error are mean and SEM.

Figure 2.

Cell-intrinsic regulation of NK cell maturation by BACH2. (A) Schematic diagram of the generation of competitive mixed BM chimeras. (B) Expression of CD27 and CD11b (left) and percentage of NK cells of indicated phenotypes (right) within the CD45.2⁺ NK compartments from the spleens of WT:WT (black dots) or Bach2^−/−:WT (red dots) chimeras. (C) Representative histograms (left) and replicate measurements (right) of KLRG1 expression on NK cells from the splenic CD45.2⁺ populations of WT:WT or Bach2^−/−:WT chimeras. (D) Replicate measurements of Granzyme B expression (MFI) in spleen and lung NK cells gated from the CD45.2⁺ population of mixed BM chimeras. (E) PCR genotyping of sorted NK, T, and B cells from spleens of mice with the indicated genotypes. NK cells have cell-specific excision of Bach2 only in mice that also possess the Ncr1^iCre gene. (F) Flow cytometry analysis of the expression of Ly49 receptors in Bach2^fl/fl Ncr1^iCre+ animals and controls. (G) Representative plots of CD27 and CD11b expression (left) and percentage of NK cells of indicated phenotypes (right) within the spleens of conditional KO animals and relevant controls. (H) Representative histograms (left) and replicate measurements (right) of KLRG1 expression on NK cells from the spleens of mice of indicated genotypes. (I) Granzyme B expression (MFI) of splenic NK cells from mice of indicated genotypes. Numbers in gates indicate percentages. Data representative of two (A–E) and four (G–I) independent experiments with five to seven mice per group. ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. Unpaired two-tailed Student’s t test. Bars and error are mean and SEM.

Figure 3.

Regulation of the global NK cell maturation program by BACH2. (A) Heatmap showing differentially expressed genes between NK cells sorted from Bach2^fl/fl Ncr1^iCre+ (cKO) mice and Ncr1^iCre+ (cWT) controls (P_adj ≤ 0.2). Also shown is the expression of these genes in NK cell maturation subsets sorted from the same animals. cWT/cKO bulk NK samples and cWT/cKO NK subset gene expression (fragments per million; FPM) independently normalized to row maxima. Genes are hierarchically clustered on the y axis. Data are representative of three biological replicates per group. (B) Average gene expression within the five identified clusters from A in cWT and cKO NK cells and NK cell subsets. (C) Venn diagram showing the overlap between significantly downregulated genes in NK cell maturation subsets sorted from cKO compared to cWT animals. (D) Venn diagram showing the overlap between significantly upregulated genes in NK cell maturation subsets sorted from cKO compared to cWT animals. (E) Normalized expression of selected genes in NK cell subsets sorted from cWT and cKO animals. (F) Representative histograms of Granzyme B expression (left) and replicate measurements of Granzyme B MFI (right) in indicated cell subsets from Rag2^−/− and Rag2^−/− Bach2^−/− animals (six mice per group). ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. RNA-Seq from three independent biological replicates per genotype is shown. Unpaired two-tailed Student’s t test. Bars and error are mean and SEM.

Figure S3.

Transcriptional control of NK cell maturation by BACH2. (A) Schematic diagram of the generation of mixed BM chimeras as a source of WT and Bach2^−/− NK cells sorted for RNA-Seq. (B) Schematic diagram showing sorted NK populations from WT mice for RNA-Seq. (C) Heatmap showing differentially expressed genes between WT and Bach2^−/− NK cells (P_adj ≤ 0.2). Also shown is the expression of these genes in NK cell maturation subsets sorted from WT animals. WT/KO and NK subset gene expression (FPM) independently normalized to row maxima. Genes are hierarchically clustered on the y axis. Data are representative of four biological replicates per group. (D) Representative KLRG1 expression (left) and replicate measurements (right) of NK cells sorted from the spleens of mice with indicated genotypes after 4 d of culture in indicated conditions. (E) Replicate measurements showing the percentage of Annexin V⁺ NK cells of the indicated genotypes after 4 d in culture with the indicated cytokines. (F) Replicate measurements showing the percentage of Annexin V⁺ NK cells from indicated subsets and genotypes after sorting by FACS and following 4 d in culture with IL-15. (G) Representative plots showing CTV staining profiles of NK cells sorted into indicated subsets from mice of indicated genotypes after 4 d in culture with IL-15. (H) Replicate measurements showing the percentage of proliferating NK cells (gated as all NK cells not in the highest CTV peak) sorted from mice of indicated genotypes into indicated subsets after 4 d in culture with IL-15. (I) Tumor volume at indicated time point after subcutaneous administration of B16-F10 melanoma cells into animals of indicated genotypes. (J) Tumor volume at indicated time points after subcutaneous administration of RMA-S lymphoma cells into animals of indicated genotypes. Tumor growth after administration of 1 × 10⁶ (left) and 2 × 10⁵ (right) RMA-S cells is shown. RNA-Seq from four independent biological replicates per genotype are shown (C). Data representative of two independent experiments (D–F, I, and J). Data from three or four technical replicates (D–H) or six to eight mice per group (I and J). ns, not significant (P > 0.05); ***, P ≤ 0.001; ****, P ≤ 0.0001. Unpaired two-tailed Student’s t test (D–F and H–J). Bars and error are mean and SEM.

Figure 4.

BACH2 regulates shared transcriptional programs in NK and CD8⁺ T cells. (A) ssGSEA of cWT and cKO NK cells and subsets with immunologic gene sets (ImmuneSigDB; P_val < 0.05, |FC| > 1.5). (B) GSEA of differentially upregulated (top) and differentially downregulated (bottom) genes in Bach2^−/− CD8⁺ T cells within the global transcriptional differences between cKO and cWT NK cells. NES, normalized enrichment score; FDR, false discovery rate. (C) Heatmap showing expression of genes which are significantly upregulated in cKO NK cells (P_adj ≤ 0.2) and expressed by both NK and CD8⁺ T cells. Genes with peak-called BACH2-binding sites between 15 kb upstream of the transcription start site and 3 kb downstream of the transcription end site of that gene within CD8⁺ T cells are indicated. cWT/cKO NK and WT/KO CD8⁺ T cell gene expression data (reads per million; RPM) independently normalized to row maxima. Genes are hierarchically clustered along the y axis. (D) Representative alignments of ChIP-Seq, ATAC-Seq, and RNA-Seq data at selected loci in indicated cell types, showing increased mRNA expression of genes. RNA-Seq from three independent biological replicates per genotype is shown.

Figure 5.

BACH2 maintains NK cell quiescence and restricts immune surveillance to pulmonary metastasis. (A) Replicate measurements of the percentage of IFN-γ⁺ NK cells sorted from the spleens of mice with indicated genotypes after 4 d of culture in indicated conditions. (B) Representative Granzyme B expression (MFI; left) and replicate measurements (right) of NK cells sorted from the spleens of mice with indicated genotypes after 4 d of culture in indicated conditions. (C) Replicate measurements of the percentage of IFN-γ⁺ NK cells (left) and NK cell Granzyme B expression (right) after sorting from the spleens of mice with indicated genotypes and 16 h of culture in indicated conditions. (D) Representative plot (left) showing CTV staining and replicate measurements (right) of the percentage of NK cells at each stage of division after isolation from the spleens of Rag2^−/− and Rag2^−/− Bach2^−/− mice by FACS and following 4 d in culture with IL-15. (E) Frequency of B16 metastases on the lobular surfaces (left) and representative photographs (right) of lungs from mice of indicated genotypes 16 d after intravenous injection of B16 cells. Data pooled from five representative experiments with six to eight mice per group. (F) Granzyme B expression (MFI) of lung NK cells from conditional KO animals and relevant controls 16 d following intravenous injection of B16 cells. (G) Counts of B16 cells after overnight co-culture in indicated ratios with NK cells sorted from mice of indicated genotypes. (H) Frequency of B16 metastases on the pleural surfaces of lungs from mice of indicated genotypes and treated intraperitoneally with the indicated antibodies 15 d following intravenous injection of B16 cells. (I) Representative images (left) and quantification of relative tumor area to total lung area (right) of metastases within lungs of mice of indicated genotypes 13 d after intravenous administration of syngeneic LL/2 cells. Relative percentage of total cross-sectional lung area occupied by cancer metastases is shown. Scale bar = 3,000 µm. Data representative of two independent experiments (A, B, D, G, and I). Data from three to five technical replicates (A–D and G) or five to nine mice per group (F–I). ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001. Unpaired two-tailed Student’s t test. Bars and error are mean and SEM.