Exciting progress in the field of cancer immunotherapy has renewed the urgency of the need for basic studies of immunoregulation in both adaptive cell lineages and innate cell lineages. Here we found a central role for major histocompatibility complex (MHC) class I in controlling the phagocytic function of macrophages. Our results demonstrated that expression of the common MHC class I component β2-microglobulin (β2M) by cancer cells directly protected them from phagocytosis. We further showed that this protection was mediated by the inhibitory receptor LILRB1, whose expression was upregulated on the surface of macrophages, including tumor-associated macrophages. Disruption of either MHC class I or LILRB1 potentiated phagocytosis of tumor cells both in vitro and in vivo, which defines the MHC class I–LILRB1 signaling axis as an important regulator of the effector function of innate immune cells, a potential biomarker for therapeutic response to agents directed against the signal-regulatory protein CD47 and a potential target of anti-cancer immunotherapy.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank the members of the Weissman laboratory and the Stanford Stem Cell Center community for discussion and assistance; J.P. Volkmer, R. Majeti, K. Loh, N. Fernhoff, M. McCracken and J. Zee for input and advice; T. Storm, T. Naik, A. McCarty, F. Khameneh, J. Ho and P. Lovelace for logistical support; and G. Krampitz (Stanford University) for the APL1 cell line. Supported by the D.K. Ludwig Fund for Cancer Research (NIH R01CA086017 and NIHG R01GM100315 to I.L.W.), the Cancer Research Institute (Irvington Fellowship to R.L.M.), the Human Frontier Science Program Organization (postdoctoral fellowship to B.R.), the NIH (hematology training grant T32 HL120824-03 to B.R.), the University of Wisconsin Medical Scientists Training Program (GM008692 to L.J.B.), the National Research Service Award 1F30DK108561, the Paul and Daisy Soros Fellowship for New Americans (to J.M.T.) and the Stanford Medical Scientist Training Program (NIH GM07365 to A.A.B., K.W., A.M.R., J.M.T., B.M.G. and J.Y.C.).
Integrated Supplementary Information
(a) Representative flow cytometry plots of in vitro phagocytosis assay. Macrophages are defined as CD45+ events, and tumor cells as GFP+ events. CD45+, GFP+ events correspond to macrophages that have phagocytosed target cells. (b) Representative images from an automated microscopy-based approach for quantifying in vitro phagocytosis. Using a custom program on the GE Cytell, we defined human macrophages based on staining with anti-CD45-PE antibody (yellow), and tumor cells based on GFP expression (green). Phagocytic events are defined based on overlap of these two signals (right panel, white arrowheads). (c) Comparison of FACS-based quantification (left) or automated microscopy-based quantification (right) of DLD1 phagocytosis under conditions of PBS treatment, treatment with the anti-CD47 antibody Hu5F9-G4, or treatment with the anti-EGFR antibody Cetuximab.
(a) Log-transformed scatterplot of normalized phagocytic efficiency upon treatment with the anti-CD47 antibody Hu5F9-G4 (Y axis) plotted against surface expression of CD47, as measured by FACS analysis with the anti-human mouse monoclonal antibody B6H12 (X axis). There is no significant relationship between these two parameters, R2=0.094, p=0.217. (b) Histogram plots of HLA-A, B, C (top) and β2M expression (bottom) for two colon cancer cell lines: DLD1 (red) and HCT116 (magenta); two small cell lung cancer lines, NCI-H82 (blue), and NCI-H196 (purple); and a pancreatic neuroendocrine tumor APL1 (green), as measured by the BioLegend LegendScreen FACS-based antibody array system. Isotype control stain is indicated in blue, while specific stain is indicated in red. (c) Log-scale scatterplot of phagocytic efficiency in a panel of cell lines upon treatment with the anti-CD47 antibody Hu5F9-G4 (Y axis), plotted against surface expression of HLA-A, B, C as measured by FACS analysis with a pan-HLA-A, B, C binding antibody (X axis). The data is a subset of that shown in Figure 1b, in order to highlight the cell lines used for the experiment shown in Figure 2g.
(a) Flow cytometry-based measurement of the percentage of LILRB1 (blue) or LILRB2 (red) positive, non-immune cells from six samples of primary ascites fluid macrophages, and four samples of primary human colorectal carcinoma samples, as determined by fluorescence minus one controls. n.s. not significant, 2-way ANOVA with multiple comparisons correction. (b) Flow cytometry-based measurement of the percentage of LILRB1 (blue) or LILRB2 (red) positive tumor associated macrophages from four independent primary human colorectal carcinoma samples, as determined by fluorescence minus one controls. n.s. not significant, 2-way ANOVA with multiple comparisons correction. (c) Representative histogram plots from FACS analysis of LILRB1 and LILRB2 expression on primary human CD14+ peripheral blood monocytes (left), and day 7 ex vivo cultured macrophages derived from the same donor (right). Specific staining is indicated in blue; IgG control is indicated in red. (d) LILRB1 and LILRB2 expression levels, as measured by mean fluorescence intensity (MFI, left panel) or percent positive cells (right panel), in 4 pairs of primary monocytes (red) and ex vivo cultured macrophages from the same donors (blue). n.s., not significant. *** p < 0.001, 2-way ANOVA with multiple comparisons correction. (e) Flow cytometry-based measurement of phagocytosis by donor-derived macrophages of MHC I– DLD1 (black) and MHC I+ DLD1 (red). Values are normalized to the highest level of phagocytosis observed in a given experimental replicate. Error bars represent s.d. of assays performed with n = 8 biological replicates. * p < 0.01, ** p < 0.001, 2-way ANOVA with multiple comparisons correction.
Supplementary Figure 4. Mouse β2M does not efficiently form stable MHC class I complexes with human HLA α-chains
FACS histogram of parental DLD1 cells (black), DLD1 cells reconstituted with a human β2M transgene (red), or with a mouse β2M transgene (blue). While fully human β2M expression facilitates robust surface MHC class I expression, as detected by the pan-HLA antibody W6/32, expression of mouse β2M facilitates only low levels of surface MHC class I.
(a) Growth curve of MHC I– DLD1 (pink) and chimeric MHC I+ DLD1 (blue), as measured by cell counting at 1, 2, and 4 days. Error bars represent are s.d. of n = 3 replicates. (b) Flow cytometry analysis of MHC class I expression in a 50-50 in vitro co-culture of MHC I– APL1 and MHC I- APL1. Error bars represent s.d. of n = 3 replicates. (c) Flow cytometry-based quantification of the percentage of fully human MHC class I expression in a 50-50 in vitro co-culture of MHC I– APL1 and MHC I- APL1 cells pre-engraftment and at 28 days post engraftment into NSG mice. Error bars are s.d. of n = 5 mice. (d) Fluorescence microscopy of sections taken from mixed MHC I+ and MHC I– APL1 tumors injected subcutaneously into the flanks of NSG mice. Chimeric MHC class I+ cells are marked by GFP expression, while MHC I– cells are marked by RFP expression. Mouse macrophages are visualized by F4/80 staining (blue). A high degree of macrophage infiltration into the tumor is evident, and macrophages can be observed in the act of phagocytosis (white arrowhead). Scale bar represents 200 μm.
(a) Tumor volume of B16-F10 variants 15 days post-engraftment into immunocompetent C57Bl/6 wild type mice (black) or CCR2-/- mice (red). n.s. not significant. * p < 0.05, *** p < 0.001, Student’s t test. (b) FACS-based measurement of LILRB1 expression in primary human macrophages electroporated with either off-target sgRNA (black) or LILRB1-targeting sgRNA, as determined by MFI. MFI values are normalized to the highest value for each experimental replicate. Error bars represent the standard deviation of assays performed with n = 2 macrophage donors. (c) Frequency of allelic deletion of LILRB1 in primary human macrophages, as measured by TIDE analysis software, n = 2 macrophage donors. (d) Percentage of total macrophages with an M1-like phenotype in macrophages modified with off-target sgRNA (black) or LILRB1-targeting sgRNA (red) in n = 2 macrophage donors. (e) Gating strategy to define infiltrating immune cells within dissociated B16-F10 mouse tumors.