Deficiency of factor-inhibiting HIF creates a tumor-promoting immune microenvironment

Abstract

Hypoxia signaling influences tumor development through both cell-intrinsic and -extrinsic pathways. Inhibiting hypoxia-inducible factor (HIF) function has recently been approved as a cancer treatment strategy. Hence, it is important to understand how regulators of HIF may affect tumor growth under physiological conditions. Here we report that in aging mice factor-inhibiting HIF (FIH), one of the most studied negative regulators of HIF, is a haploinsufficient suppressor of spontaneous B cell lymphomas, particular pulmonary B cell lymphomas. FIH deficiency alters immune composition in aged mice and creates a tumor-supportive immune environment demonstrated in syngeneic mouse tumor models. Mechanistically, FIH-defective myeloid cells acquire tumor-supportive properties in response to signals secreted by cancer cells or produced in the tumor microenvironment with enhanced arginase expression and cytokine-directed migration. Together, these data demonstrate that under physiological conditions, FIH plays a key role in maintaining immune homeostasis and can suppress tumorigenesis through a cell-extrinsic pathway.

Keywords: B cell lymphoma; factor-inhibiting HIF; hypoxia-inducible factor; tumor microenvironment; tumor suppression.

Fig. 1.

FIH deficiency increases spontaneous tumorigenesis. (A) Overall survival curve of FIH^+/+ (N = 14) FIH^+/Δ1-2 (N = 21), and FIH^Δ1-2/Δ1-2 (N = 22) mice over 125 wk. ns, non-significant by the Mantel–Cox test. N denotes the number of mice. (B) Tumor-free survival curve of FIH^+/+ (N = 14) FIH^+/Δ1-2 (N = 21), and FIH^Δ1-2/Δ1-2 (N = 22) mice over 125 wk. * indicates P < 0.05 by the Mantel–Cox test. Throughout the results, ns label and asterisks indicate the comparison between FIH^+/+ and FIH^+/Δ1-2 (blue) and between FIH^+/+ and FIH^Δ1-2/Δ1-2 (red), except where otherwise indicated. (C) The spectrum and frequency of tumors observed in FIH^+/+ (N =14), FIH^+/Δ1-2 (N = 21), and FIH^Δ1-2/Δ1-2 (N = 22) animals. Eight cancer types are colored and tumor frequencies (number of mice with detectable tumors divided by number of mice with the indicated genotype) are indicated. FIH^Δ1-2/Δ1-2 mice showed increased tumor incidence compared to FIH^+/+ mice (*P = 0.0203, χ² test). (D) B cell lymphoma frequency in indicated organs and tissues (number of mice with B cell lymphoma at the indicated site divided by number of mice with the indicated genotype) in FIH^+/+(N = 14), FIH^+/Δ1-2 (N = 21), and FIH^Δ1-2/Δ1-2 (N = 22) animals (* indicates P < 0.05 and ** indicates P < 0.01 by the χ² test). (E) Pie charts showing the proportions of B cell lymphomas of indicated grade (number of B cell lymphomas of indicated grade divided by the total number of B cell lymphomas found within the group) in FIH^+/+ (n of B lymphomas=7), FIH^+/Δ1-2 (n of B lymphomas=30) and FIH^Δ1-2/Δ1-2 (n of B lymphomas = 41). ** indicates P < 0.01 and **** indicates P < 0.0001 compared with FIH^+/+ by the χ² test.

Fig. 2.

FIH deficiency affects the immune cell composition in aged mice but not young adult mice. (A) Flow cytometry profiling of immune cells (pre-pro B cells (B220⁺CD43⁺CD24^loBP-1⁻IgM⁻IgD⁻), pro B cells (B220⁺CD43⁺CD24^midBP-1⁻IgM⁻IgD⁻), proliferating pre B cells (B220⁺CD43⁺CD24^hiBP-1⁺IgM⁻IgD⁻), pre B cells (B220⁺CD43⁻CD24^hiBP-1⁺IgM⁻IgD⁻), immature B cells (B220⁺CD43⁻CD24^midIgM⁺IgD⁻), mature B cells (B220⁺CD43⁻CD24^loIgM⁺IgD⁺), CD4⁺ T cells, CD8⁺ T cells, CD11b⁺ myeloid cells, inflammatory monocytes (CD11b⁺Ly6G⁻Ly6C^hi), and neutrophils (CD11b⁺Ly6G⁺Ly6C⁺) in the indicated tissues of 18-wk-old (young) FIH^+/+ and FIH^Δ1-2/Δ1-2 mice (Bone marrow, spleen, and lung tissues were derived from 4 FIH^+/+ mice and 4 FIH^Δ1-2/Δ1-2 mice, whereas thymus tissues were obtained from 3 FIH^+/+ and 3 FIH^Δ1-2/Δ1-2 mice). (B) Flow cytometry analysis of B cell subsets in the spleens of 18-wk-old (young, Left) and 104-wk-old (aged, Right) mice FIH^+/+ mice (N = 4) and FIH^Δ1-2/Δ1-2 mice (N = 3). Follicular (Fo) B cells (CD19⁺B220⁺CD21^midCD23⁺), marginal zone (Mz) B cells (CD19⁺B220⁺CD21⁺CD23^mid), transitional (Trans) B cells (CD19⁺B220⁺CD21⁻CD23⁻). Inflammatory Mo, inflammatory monocytes. (C) Flow cytometry profiling of immune cell subtypes in the indicated tissues of 104-wk-old (aged) FIH^+/+ (N = 4) and FIH^Δ1-2/Δ1-2 (N = 3) mice. (D) mRNA levels of Tgfb, Il10, Il6, and Tnfa in the lung, liver, and small intestine (SI) of FIH^+/+ and FIH^Δ1-2/Δ1-2 mice of 104 wk (aged) determined by RT-qPCR. Horizontal lines indicate mean with SD. * indicates P < 0.05, ** indicates P < 0.01 by two-tailed t tests. Outliers identified by the Grubbs test were excluded from the dataset.

Fig. 3.

FIH-defective mice show enhanced growth of LLC tumors. FIH^+/+, FIH^+/Δ1-2, and FIH^Δ1-2/Δ1-2 mice were subjected to subcutaneous injection of 4 × 10⁵ LLC cells or 1 × 10⁵ B16 cells on both flanks. n denotes number of tumors. (A) Average LLC tumor volumes from day 1 to day 14. Tumor volumes and number of tumors measured at each time point are shown in SI Appendix, Table S2. Statistically significant differences between FIH^+/+ vs. FIH^+/Δ1-2 and FIH^Δ1-2/Δ1-2 are indicated by blue and red asterisks, respectively. (B) Average weights of LLC tumors collected on day 14. Tumor weights are shown in SI Appendix, Table S2. (C) Mean B16 tumor volumes from day 1 to 14 are shown. Tumor volumes, number of tumors measured and number of mice with measurable tumors at each time point are shown in SI Appendix, Table S3. (D) Flow cytometry analysis of the percentages of CD4 and CD8 T cells in all LLC tumors collected on day 14 post inoculation (left). Representative FACS plots of CD8 T cells (right). (E) Flow cytometry analysis of the percentages of tumor-associated macrophages (TAMs, CD11b⁺F4/80⁺), monocytic-myeloid derived suppressor cells (M-MDSCs, CD11b⁺CD11c⁻Ly-6G⁻Ly6C^hi), and granulocytic-myeloid derived suppressor cells (G-MDSCs, CD11b⁺CD11c⁻Ly-6G⁺Ly6C⁺) in LLC tumors collected on day 14. Bars indicate mean with SD. * indicates P < 0.05, by two-tailed t tests.

Fig. 4.

FIH myeloid knockout (MKO) mice show enhanced subcutaneous growth of LLC tumors but not spontaneous tumor development. (A and B) A cohort of WT (N = 20) and FIH MKO (N = 28) mice were monitored for 125 wk. The overall survival (A) and tumor-free survival curves (B) are shown. (C–G) WT and FIH MKO mice were subjected to subcutaneous injection of 4 × 10⁵ LLC cells at the left flank. (C) Average tumor volumes from day 1 to 13. Mean tumor sizes at each time point are shown in SI Appendix, Table S4. (D) Average weights of LLC tumors from WT and FIH MKO animals at the end of the study (pool of four studies). Individual points represent single tumors. (E) Quantification of TAMs, M-MDSCs, and G-MDSCs in LLC tumors by flow cytometry (pool of four studies). (F) Average percentage of TAMs, M-MDSCs, and G-MDSCs in LLC tumors positive for arginase 1 (ARG1) determined by flow cytometry (pool of four studies). (G) Average percentage of CD4 and CD8 T cells from LLC tumors measured by flow cytometry (pool of three studies). Upper case N denotes number of animals while lower case n denotes number of tumors hereafter. Error bars indicate the mean ± SD in panels C–G. * indicates P < 0.05, ** indicates P < 0.01, and *** indicates P < 0.001, two-tailed t test.

Fig. 5.

Myeloid deletion of HIF2α suppresses LLC tumor growth and abolishes the effect of myeloid FIH expression. (A and B) 4 × 10⁵ LLC cells were administered to the left flank of WT (N = 27) and HIF1α MKO (N = 14) animals (9 to 13 wk old). Pooled results from two independent experiments are presented. (A) Average tumor volumes from day 1 to 14. Mean tumor volumes at each time point are detailed in SI Appendix, Table S6. (B) Average weights of LLC tumors from WT and HIF1α MKO animals at the end of the study. (C and D) 4 × 10⁵ LLC cells were administered to the left flank of WT (N = 17), HIF2α MKO (N = 23), and HIF2α FIH MDKO (myeloid double knockout) (N = 17) animals (9 to 13 wk old). Pooled results from three independent experiments are presented. (C) Average tumor volumes from day 1 to 14. Mean tumor volumes at each time point are shown in SI Appendix, Table S7. * indicates P < 0.05 and ** indicates P < 0.01 by the two-tailed t test. Asterisks denote comparison between WT and HIF2α MKO mice (green) or between WT and HIF2α FIH MDKO mice (orange). (D) Average weights of LLC tumors from WT, HIF2α MKO and HIF2α FIH MDKO animals at the end of the study. Mean ± SD is indicated in all panels. ** indicates P < 0.005 by the two-tailed t test. (E) Murine macrophage cell line J774 cells were knocked down with either FIH, HIF1α, or HIF2α, or a combination of FIH/HIF1α or FIH/HIF2α, followed by stimulation with LPS or IL-4 for 48 h. Arg1 expression was detected by qRT-PCR.

Fig. 6.

FIH deletion enhances Arg1 expression in vitro. (A) BMDMs harvested from FIH^+/+ (N = 4) and FIH^Δ1-2/Δ1-2 (N = 4) mice were stimulated with LPS followed by bulk RNA sequencing. A clustered heat map of the differentially expressed genes from the FIH^+/+ and FIH^Δ1-2/Δ1-2 mice is shown. A gene was considered to be differentially expressed if its absolute log2 fold change is >1 and false discovery rate (FDR)< 0.05. Color key indicates the z-scores of normalized expression values. (B) BMDMs isolated from FIH^+/+ and FIH^Δ1-2/Δ1-2 mice were treated with M1–polarizing (5 ng/mL LPS and 1 ng/mL IFNγ) or M2–polarizing (10 ng/mL IL-4) agent for 24 h. mRNA level of Arg1 was determined by RT-qPCR. UT, untreated. Bars indicate mean ± SD. **** indicates P < 0.0001 by two-tailed t tests. (C) Western blots showing ARG1 and NOS2 protein levels in FIH^+/+ and FIH^Δ1-2/Δ1-2 BMDMs in response to LPS+IFN or IL-4 treatments at the indicated time points. Following the detection of ARG1 expression, the same blot was incubated with anti-β-actin antibody as a loading control. (D) Western blots showing the expression levels of ARG1 and NOS2 in FIH^+/+ and FIH^Δ1-2/Δ1-2 BMDMs infected with Mycobacterium bovis BCG at one multiplicity of infection (MOI) and collected at indicated hours post infection. β-actin was used as a loading control in the re-probed blot. (E) RT-qPCR analysis of Hif1an, Vegf, Arg1, Nos2, and Il6 expression in TAMs derived from LLC tumors grown in FIH^+/+ (N = 3) and FIH^Δ1-2/Δ1-2 (N = 4) mice. (F) FIH^+/+ and FIH^Δ1-2/Δ1-2 BMDMs were treated with medium conditioned by LLC or B16 cells (tumor-conditioned medium, TCM). Arg1 expression was detected by RT-qPCR (Left) and western blotting (Right, β-tubulin as loading control). Horizontal lines and error bars represent mean with SD. * indicates P < 0.05 by two-tailed t tests.

Fig. 7.

Loss of FIH promotes macrophage migration toward CCL5 and C5a and promotes ARG1+, F4/80+, and F4/80+/ARG1+ M2-like macrophage infiltration in LLC tumors. (A) Migration of BMDMs derived from FIH^+/+ and FIH^Δ1-2/Δ1-2 mice toward CCL5 (10 ng/mL; FIH^+/+ N = 4, FIH^Δ1-2/Δ1-2 N = 3) and C5a (10 ng/mL; FIH^+/+ N = 3, FIH^Δ1-2/Δ1-2 N = 3) was recorded by xCELLigence RTCA-DP instrument. Migration traces and max-min analyses of the traces are shown. (B) Expression levels of CCL5 receptor (CCR5) and C5a receptor (C5aR) in FIH^+/+ (N = 4) and FIH^Δ1-2/Δ1-2 (N = 3) BMDMs as determined by flow cytometry. gMFI, geometric mean fluorescence intensity. Horizontal lines and error bars represent mean ± SD. * indicates P < 0.05, ** indicates P < 0.01, and *** indicates P < 0.001, by two-tailed t tests. (C) Whole section scanning by confocal microscopy showing dual immunofluorescence staining of ARG1 (red), F4/80 (green), and DAPI (blue) in LLC tumors grown in FIH^+/+ and FIH^+/Δ1-2 mice (three replicates shown for each genotype, from three different mice). Zoomed-in regions (Scale bar, 20 µm) correspond to the adjacent white box outline on the main image. (Scale bar, 200, 500 or 1,000 µm, as indicated.)