Sensitization of B16 tumor cells with a CXCR4 antagonist increases the efficacy of immunotherapy for established lung metastases

Abstract

Expression of the chemokine receptor CXCR4 by tumor cells promotes metastasis, possibly by activating prosurvival signals that render cancer cells resistant to immune attack. Inhibition of CXCR4 with a peptide antagonist, T22, blocks metastatic implantation of CXCR4-transduced B16 (CXCR4-luc-B16) melanoma cells in lung, but not the outgrowth of established metastases, raising the question of how T22 can best be used in a clinical setting. Herein, whereas the treatment of CXCR4-luc-B16 cells in vitro with the CXCR4 ligand CXCL12 did not reduce killing induced by cisplatin or cyclophosphamide, CXCL12 markedly reduced Fas-dependent killing by gp100-specific (pmel-1) CD8(+) T cells. T22 pretreatment restored sensitivity of CXCR4-luc-B16 cells to pmel-1 killing, even in the presence of CXCL12. Two immune-augmenting regimens were used in combination with T22 to treat experimental lung metastases. First, low-dose cyclophosphamide treatment (100 mg/kg) on day 5 in combination with T22 (days 4-7) yielded a approximately 70% reduction of B16 metastatic tumor burden in the lungs compared with cyclophosphamide treatment alone (P < 0.001). Furthermore, whereas anti-CTL antigen 4 (CTLA4) monoclonal antibody (mAb; or T22 treatment) alone had little effect on established B16 metastases, pretreatment with T22 (in combination with anti-CTLA4 mAb) resulted in a 50% reduction in lung tumor burden (P = 0.02). Thus, in vitro, CXCR4 antagonism with T22 renders B16 cells susceptible to killing by antigen-specific T cells. In vivo, T22 synergizes with cyclophosphamide or anti-CTLA4 mAb in the treatment of established lung metastases, suggesting a novel strategy for augmenting the efficacy of immunotherapy.

Figure 1

CXCR4 activation does not protect CXCR4-B16 from cell death induced by cyclophosphamide. CXCR4-luc-B16 cells were seeded in six-well plates (1.5 × 10⁵ per well) in DMEM with 0.5% FCS. On reaching 30% confluence, the cells were treated with CXCL12 (1 μg/mL) overnight if indicated. Cells were then treated with cyclophosphamide at the indicated concentrations for 24 h and then assessed for viability by the Annexin V assay. Representative of three experiments with similar results.

Figure 2

Activation of CXCR4 in B16 cells by CXCL12 prevents Fas-dependent pmel-1-induced cytotoxicity. A to C, after serum starvation and IFN-γ treatment, CXCR4-luc-B16 cells were labeled with calcein-AM and mixed with different ratios of pmel-1 cells (from WT as well as perforin- and FasL-deficient mice) for 3 h at 37°C in the presence of CXCL12, T22, zVAD-fmk (zVAD), and anti-FasL antibody. The supernatants were recovered and cell lysis fraction was calculated based on calcein release. The lysis fraction (%) was calculated by dividing the observed calcein release by the maximal (total) calcein release of the cells treated with 0.1% Triton X-100. A, pmel-1 cells from WT (◆), WT with 100 μg/mL anti-FasL antibody (●), FasL knockout (■), and perforin knockout mice (▲) were used. B, T22 was added to CXCR4-luc-B16 cells for 3 h at 37°C before treatment with CXCL12 (□). T22 was also added to CXCR4-luc-B16 cells without CXCL12 treatment (■). C, luc-B16 cells were used as target cells in the presence of pmel-1 cells (◆) with 1 μmol/L zVAD-fmk (●) or 1 μg/mL CXCL12 (▲). D, human MEL501A melanoma cells were serum starved overnight to increase CXCR4 expression, then treated with 1 mg/mL CXCL12 (▲), 1 μmol/L zVAD-fmk (●), 1 μg/mL T22 alone (■), T22 + CXCL12 (□), or left untreated (◆) overnight, labeled with calcein-AM, and then added to interleukin 2 (IL-2)–treated melanoma antigen–specific JR6C12 T cells at different E/T ratios for 3 h at 37°C. Percent cytotoxicity was calculated as for B16 cells. A to D, representative experiment from two or more experiments with consistent findings.

Figure 3

CXCR4 inhibition by T22 potentiates cyclophosphamide in reducing established lung metastases of CXCR4-luc-B16 melanoma in vivo. CXCR4-luc-B16 cells (4 × 10⁵ per mouse) were injected via tail vein of C57BL/6 mice (n = 5 per treatment group). Mice were euthanized on day 14 to measure luciferase activity in lungs. A, on day 5, the indicated doses of cyclophosphamide were administered i.p. to mice. B, in addition to cyclophosphamide treatment at 100 mg/kg on day 5, some of the mice were also treated with T22 40 μg (or the nonactive control peptide ALA, 40 μg) i.p. daily on day 4 to 7. Representative of three or more independent experiments. Luciferase activity is shown in relative light units.

Figure 4

An intact immune system is required for T22 to synergize with cyclophosphamide. A and B, CXCR4-luc-B16 cells (4 × 10⁵ per mouse) were injected into SCID/bge mice via tail vein. A, SCID/bge mice were treated with cyclophosphamide (100 mg/kg i.p.) on day 5. On day 14, tumor burden was assessed by measuring individual lung mass. Lungs from noninjected mice served as controls. B, SCID/bge mice were treated with cyclophosphamide (100 mg/kg i.p.) on day 5 with or without T22 (40 μg/mouse i.p.) daily on day 4 to 7. On day 14, lung tumor burden was measured by luciferase assay. The experiments depicted were done on different days with separate preparations of cells.

Figure 5

T22 pretreatment is required for anti-CTLA4-mediated reduction of established tumor burden in the lungs. A and B, CXCR4-B16 cells were inoculated via the tail vein in WT mice (n = 4 mice per treatment group, otherwise as described in Fig. 3). T22, when used, was administered i.p. (40 μg/mouse) daily on day 3 to 7 and day 9 to 10. Anti-CTLA4 mAb (clone 9H10), when used, was administered i.p. (250 μg/mouse) on day 3, 6, and 9. When combination treatment was administered, T22 injections were given i.p. 2 h before injection of anti-CTLA4 mAb. Mice were euthanized on day 13 for visual inspection of lung metastases (representative lungs from one of four mice used per group; A) and quantitative in vitro measurement of luciferase activity (B). Columns, mean; bars, SE.