Natural Killer Cells Control Tumor Growth by Sensing a Growth Factor

Abstract

Many tumors produce platelet-derived growth factor (PDGF)-DD, which promotes cellular proliferation, epithelial-mesenchymal transition, stromal reaction, and angiogenesis through autocrine and paracrine PDGFRβ signaling. By screening a secretome library, we found that the human immunoreceptor NKp44, encoded by NCR2 and expressed on natural killer (NK) cells and innate lymphoid cells, recognizes PDGF-DD. PDGF-DD engagement of NKp44 triggered NK cell secretion of interferon gamma (IFN)-γ and tumor necrosis factor alpha (TNF-α) that induced tumor cell growth arrest. A distinctive transcriptional signature of PDGF-DD-induced cytokines and the downregulation of tumor cell-cycle genes correlated with NCR2 expression and greater survival in glioblastoma. NKp44 expression in mouse NK cells controlled the dissemination of tumors expressing PDGF-DD more effectively than control mice, an effect enhanced by blockade of the inhibitory receptor CD96 or CpG-oligonucleotide treatment. Thus, while cancer cell production of PDGF-DD supports tumor growth and stromal reaction, it concomitantly activates innate immune responses to tumor expansion.

Keywords: NK cell; NKp44; PDGF-D; cancer; cell cycle; cytokines; growth factor; immunosurveillance; innate lymphoid cells.

Figure 1.. PDGF-DD is a ligand for NKp44.

(A) A secretome library screen with NKp44-GFP reporter cells identified 30 proteins (red) as putative NKp44 ligands including PDGF-D; y-axis, screen 1; x-axis, screen 2 (normalized fluorescence signal, NFS). (B) GFP expression from NKp44-GFP reporter cells incubated with dilution series of the top 30 secretome library hits (PDGF-D, red; protein hits 1-29, black). Plate immobilized anti-NKp44 (blue) and anti-NKp30 (green) mAbs were used as positive and negative controls. (C) Schematic of PDGFD, PDGF-D, and PDGF-DD isoforms: signal peptide (SP, white); CUB domain (blue); growth factor domain (GFD, red); black triangle, protease cleavage site; black lines, disulfide bonds. (D, E) Recombinant PDGF-D and PDGF-DD activate NKp44-GFP reporter cells that is blocked by anti-PDGF-D. (F) NKp44-Fc binds to recombinant PDGF-DD but not PDGF-D in solid phase. BSA, recombinant Chikungunya virus (CHIK) E2 protein and IL-17A were used as negative controls. Recombinant CUB-TEV-PDGFD contains a TEV cleavage site between CUB and GFD. (G) Dose-dependent GFP expression from NKp44-GFP reporter cells stimulated with PDGF-DD. (H) PDGF-DD has 3.3 ± 0.4 μM affinity for NKp44 as determined by surface plasmon resonance (SPR). (I) PDGF-DD but not PDGF-D binds NKp44 by SPR. PDGF-D or PDGF-DD binding were investigated in the solid phase and NKp44 in the mobile phase (two-fold dilution series from 36 μM to 36 nM). A subset of the concentrations tested is shown. Data are represented as mean ± SEM (****, P ≤ 0.0001).

Figure 2.. PDGF-DD triggers cytokine secretion pathways in NK cells via NKp44 and DAP12.

(A) PDGF-DD induces calcium signaling in NK92 cells, which is blocked by anti-NKp44 mAb. (B) IL-2 cultured NK cells were stimulated with PDGF-DD for the indicated time points and phosphorylated AKT, ERK1/2 and FOXO3A and total ERK1/2 determined by immunoblotting. (C, D) PDGF-DD stimulates dose-dependent NK cell secretion of IFN-γ and TNF-α. NK cells were kept in IL-2 medium throughout the assay. (E, F) IFN-γ and TNF-α secretion by PDGF-DD-stimulated NK cells derived from either normal donor (Norm) or DAP12-deficient (D12) patient. Induction of IFN-γ and TNF-α is blocked by anti-NKp44. In the absence of PDGF-DD, IFN-γ and TNF-α can be induced by cross-linking NKp44 with GαM+α-NKp44 as surrogate NKp44 ligand. (G) Representative dotplots of intracellular IFN-γ and TNF-α staining of PDGF-DD-stimulated IL-2-cultured NK cells from normal (Norm) or DAP12-deficient (D12) donors (percentage expression indicated in each quadrant). PMA/i was used as positive control. (H) Representative histograms of NKp44 surface expression on IL-2-cultured NK cells from normal (Norm) or DAP12-deficient (D12) donors before and after PDGF-DD stimulation. (I) Volcano plot and heatmap analysis of RNA-seq transcriptional profile induced by PDGF-DD in 4 NK cell donors. Transcripts in red were upregulated at least 1.5-fold by PDGF-DD (NK;1-4_PD) versus unstimulated NK cells (NK;1-4), those in blue were similarly downregulated. (J) Top 10 highest scoring IPA Cellular Immune Response pathways generated from transcripts differentially expressed (±1.5-fold) in PDGF-DD-stimulated NK cells, as in (I). Ratios in columns indicate total PDGF-DD-regulated genes to the total number of genes in each pathway. Data represented as mean ± SEM (****, P ≤ 0.0001).

Figure 3.. PDGF-DD/NKp44 interaction induces cytokine secretion by ILC1 and ILC3.

(A) Representative histograms of intracellular TNF-α content of freshly isolated tonsil ILC3s stimulated with PDGF-DD and/or IL-23 (percentage cytokine-positive cells is indicated in each histogram). (B) Percentage of TNF-α secreting cells in PDGF-DD-stimulated tonsil ILC3s from 6 different donors (***, P ≤ 0.001). (C, D) ILC1 derived in vitro by IL-2-induced conversion of tonsil ILC3 secrete IFN-γ and TNF-α following stimulation with PDGF-DD. Anti-NKp44 but not control anti-NKp46 mAb blocks cytokine secretion. Data represented as mean (n = 3) ± SEM (****, P ≤ 0.0001). (E) Volcano plot and heatmap analysis of transcripts differentially expressed between PDGF-DD-stimulated (ILC1_1/2PD) versus unstimulated tonsilar ILC1 (ILC1_1/2) isolated from two different donors. Transcripts represented in red were upregulated at least 1.5-fold in PDGF-DD-stimulated versus unstimulated ILC1, and those in blue were similarly downregulated. (F) Top 5 highest scoring IPA Cellular Immune Response pathways generated from transcripts differentially expressed (±1.5-fold) in PDGF-DD-stimulated tonsil ILC1. Ratios in columns indicate total PDGF-DD-regulated genes to the total number of genes in each pathway.

Figure 4.. PDGF-DD binding to NKp44 induces NK cell secretion of cytokines that mediate tumor growth arrest.

(A) Meljuso melanoma cells were cultured with supernatants from NK cells stimulated with either PDGF-DD plus IgG1 (NK^DD+IgG sup) or PDGF-DD plus anti-NKp44 (NK^DD+α-NKp44 sup) or recombinant IFN-γ and TNF-α. Representative dotplots (left panel) show BrdU incorporation (y-axis) and 7-AAD staining for total DNA content (x-axis). The percentages of cells in different stages of the cell cycle (G1, S and G2/M) are indicated in each gate. Cells in the different stages of the cell cycle are quantified (right panel). (B, C) Colo38 or Meljuso melanoma cells were cultured in either NK^DD+IgG sup, NK^DD+α-NKp44 sup or complete medium, washed, then re-plated in complete medium and the percentage (%) growth relative to control medium recorded for 3 passages. (D) Colo38 cells were cultured in either complete medium or NK^DD+IgG sup with (+) or without anti-IFN-γ and anti-TNF-α (Abs) before washing and passaging in complete medium as described above. (E-G) Colo38 or Meljuso cells were pre-cultured with NK^DD+IgG sup, NK^DD+α-NKp44 sup or complete medium. IL-8 or IL-6 secretion in culture medium was measured at the end of the first passage. (H) Representative dotplots of CADM1 (assessed by CRTAM-Fc), CD112, CD155 and CD95 (Fas) expression from Colo38 cells exposed to NK^DD+IgG sup or NK^DD+α-NKp44 sups (24h). (I) RNA-seq profiles of human melanoma (Colo38 and Meljuso), breast cancer (MCF7) and ovarian cancer (OVCA) cell lines exposed to either NK^DD+IgG sup or NK^DD+α-NKp44 sup (48h). Transcripts in red were upregulated at least 1.5-fold and those in blue were similarly downregulated. Data are represented as mean ± SEM (***, P ≤ 0.001; ****, P ≤ 0.0001; ns, not significant).

Figure 5.. Core signatures of PDGF-DD-activated NK cells and growth arrested tumor cells correlate with NCR2 expression in TCGA GBM cohort.

(A, B) PDGF-D expression in tissue sections of GBM, as determined by immunohistochemistry (representative of 5 cases). Arrowheads indicate PDGF-D⁺ hyperplastic blood vessels. (C) Expression of signature core cytokine genes by NK cells from different donors kept in medium alone (Unstim), or stimulated with PDGF-DD or PMA/i. (D) Expression of signature core cell cycle genes in different tumor cells exposed to medium alone, NK^DD+α-NKp44 supernatant, or NK^DD+IgG supernatant. (E) Correlation data for individual cytokine genes (y-axis) in the core signatures of PDGF-DD-activated NK cells versus NCR2 expression (x-axis) extracted for TCGA GBM. Each dot represents a tumor case (n = 539 patients). The statistical significance of the correlation was determined using the Pearson’s correlation coefficient. A red linear regression line is shown in each plot. (F) Infiltration of NKp44⁺ (red) and CD3ε⁺ (green) lymphocytes in tissue sections of a GBM case (representative of 8) including a case of gliosarcoma, as assessed by immunofluorescence. (G) Correlation data for NCR2 (y-axis) versus NCR3 (x-axis) expression in the TCGA GBM cohort. (H) Correlation data for individual cell cycle genes (y-axis) in core signatures of growth arrested tumor cells versus NCR2 expression (x-axis) extracted for TCGA GBM. (I, J) Kaplan-Meier survival curves for the first (Q1) and fourth quartile (Q4) of the canonical cytokine and canonical cell cycle variate of NCR2 expression, respectively.

Figure 6.. NKp44 restricts tumor cells expressing PDGFD in vivo.

(A) Representative dotplots of NKp44 expression in CD3⁻NK1.1⁺ NK cells (NK), CD3⁺NK1.1⁺ NKT cells (NKT), and CD3⁺NK1.1⁻ T lymphocytes (T) isolated from spleen and mesenteric lymph nodes (mLN) of NCR2-tg (founder #1) and non-tg mice. The percentages of cells are indicated in each gate. (B) NKp44 expression on NK cells from spleens of NCR2-tg and non-tg mice activated in vitro with IL-2 and IL-15 ± IL-1β and TNF-α. (C) GFP expression of B16F10 melanoma cells stably transduced with pMX-IRES-eGFP retroviral vector (B16-pMX) or pMX-IRES-eGFP encoding PDGFD (B16-PDGFD). (D, E) Tissue culture supernatants from B16-PDGFD cells elicit IFN-γ and TNF-α secretion from human NK cells, which is blocked by anti-PDGF-D or anti-NKp44. (F) Quantification (left panel) and representative photos (right panel) of surface lung metastases formed in NCR2-tg or non-tg littermate mice injected with B16-PDGFD cells. (G) Mean tumor area (μM²) ± SEM and representative histochemical images of lung metastases from NCR2-tg (n = 6) or non-tg (n = 7) mice injected with B16-PDGFD cells. (H, I) Day 17 quantitative RT-PCR of Ifng and Tnf transcripts in lungs of NCR2-tg and non-tg mice injected with B16-PDGFD cells. (J) Day 17 surface lung metastases in NCR2-tg injected with B16-PDGFD cells and treated with either control or anti-NKp44 antibodies. (K) Day 17 surface lung metastases in NCR2-tg and non-tg mice injected with B16-pMX control cells (ns, not significant). Data are represented as mean ± SEM (*, P ≤ 0.05; **, P ≤ 0.01; ****, P ≤ 0.0001; ns, not significant).

Figure 7.. PDGF-DD binding to NKp44 augments immunotherapies.

(A) NK cell expression of NKp44 and CD96 in the lungs of NCR2-tg and non-tg mice. (B) Representative plots of NK cells from day 17 lungs and mLN of NCR2-tg mice injected with B16-PDGFD cells, showing expression of NKp44, CD96, PD-1, or TIGIT. (C) Percentages of NKp44⁺ NK cells expressing CD96, PD-1, or TIGIT in day 17 lungs and mLN of NCR2-tg mice injected with B16-PDGFD cells. (D) The CD96 ligand CD155, and CD95 are upregulated on B16F10 cells stimulated with IFN-γ and TNF-α. (E) Quantification of day 17 surface lung metastases from NCR2-tg and non-tg mice injected with B16-PDGFD cells and treated with anti-CD96 mAb or control antibodies. (F) Mean tumor volumes ± SEM from NCR2-tg or non-tg mice injected subcutaneously with B16-PDGFD cells and treated with CpG-ODN or left untreated (arrows indicate intratumoral CpG-ODN injections). (G, H) Representative dotplots and percentage of NKp44⁺ cells in NK cells isolated from B16-PDGFD tumors of NCR2-tg or non-tg mice treated with CpG-ODN or untreated.