Active immunotherapy induces antibody responses that target tumor angiogenesis

Abstract

The inhibition of VEGF signaling with antibodies or small molecules achieves clinical benefits in diverse solid malignancies. Nonetheless, therapeutic effects are usually not sustained, and most patients eventually succumb to progressive disease, indicating that antiangiogenic strategies require additional optimization. Vaccination with lethally irradiated, autologous tumor cells engineered to secrete granulocyte-macrophage colony stimulating factor (GM-CSF) and antibody blockade of cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) trigger a tumor vasculopathy in some long-term responding subjects. These reactions are characterized by disrupted tumor blood vessels in association with lymphocyte and granulocyte infiltrates and zonal areas of ischemic tumor necrosis. However, the mechanisms underlying this immune-mediated destruction of the tumor vasculature remain to be clarified. Here, we show that GM-CSF-secreting tumor cell vaccines and CTLA-4 blockade elicit a functionally important humoral reaction against multiple angiogenic cytokines. Antibodies to angiopoietin-1 and angiopoietin-2 block Tie-2 binding, downstream signaling, endothelial cell tube formation, and macrophage chemotaxis. Antibodies to macrophage inhibitory factor (MIF) attenuate macrophage Tie-2 expression and matrix metalloproteinase-9 (MMP-9) production. Together, these results delineate an immunotherapy-induced host response that broadly targets the angiogenic network in the tumor microenvironment.

Figure 1

Vaccinated melanoma patient K008 developed antibodies to multiple angiogenic cytokines. (A). Sera from K008, but not healthy donors (n=3) recognize VEGF-A, but not bFGF by immunoblotting (sera at 1:100). Control anti-cytokine antibodies are shown for comparison. (B). Longitudinal analysis of antibodies to VEGF family members in K008. Antigen specific IgG levels were determined with an ELISA (sera at 1:100). Arrows denote vaccinations. Day 0 is pre-vaccination. (C). Vaccination stimulated potent humoral reactions to angiopoietin-1 and -2 in K008. A subcutaneous nodule developed hemorrhagic necrosis after the fifth immunization. Day 0 is pre-vaccination.

Figure 2

Vaccine induced antibodies to angiopoietin-1 and -2 manifest blocking activity. (A). Post-vaccination K008 sera inhibit the binding of Tie-2 and angiopoietin-1 as measured with an ELISA (sera at 1:10). Day 0 is pre-vaccination. Full time-course studies were performed twice with similar results, whereas high and low blocking samples were analyzed two additional times, with equivalent findings. (B). Sera from K008, but not healthy donors inhibit angiopoietin-1 and -2 induced tube formation with HUVECs. K008 sera failed to block MFG-E8 stimulated tube formation. This experiment was performed two times with similar results. Pre-vaccination sera from K008 were not available for this assay. Representative high power fields are shown. Tube forming scores (averages of six replicates) were: donor (media) 0, (angiopoetin-1) 3.5, (angiopoietin-2) 3; K008 (MFG-E8) 2.67, (angiopoietin-1) 0.83, (angiopoietin-2) 0.33.

Figure 3

Vaccinated patients harbor neutralizing antibodies to angiopoietin-1 and -2. (A, B). Anti-angiopoietin-1 and -2 IgG antibodies were measured with an ELISA (sera 1:100) in normal donors (n=16) and vaccinated patients (n=16). (C). Antibodies in vaccinated patients block the binding of angiopoietin-1 and Tie-2.

Figure 4

Vaccination stimulated functional antibodies to angiopoietin-1 and -2. (A). Longitudinal analysis of metastatic melanoma patient M30 (sera 1:100). Day 0 is pre-vaccination. (B). Longitudinal analysis of metastatic non-small cell lung carcinoma patient L19. Day 0 is pre-vaccination. (C). Post-vaccination sera from M30 specifically blocks angiopoietin-1 and -2 induced tube formation. Serum was absorbed against recombinant cytokine as indicated. This experiment was performed twice with similar results. Tube forming scoring (average of two replicates): Pre-VAX (angiopoietin-1 and -2) 3.5, post-VAX (angiopoietin-1 and -2) 0, post-VAX absorbed (angiopoietin-1 and -2) 1.5. (D). Post-vaccination sera from M30 and L19 specifically block angiopoietin-2 stimulated migration of Tie-2 expressing monocytes. Cell migration in response to cytokine as compared to media is depicted. Serum was absorbed against recombinant cytokine as indicated. Similar results were observed in a second experiment.

Figure 5

Angiopoetin-1 and -2 trigger ERK phosphorylation. (A). HUVECs were treated with 200 ng/ml of angiopoietin-1 or 500 ng/ml angiopoietin-2 for varying times. Cell lysates were analyzed for phospho-ERK with immunoblotting. (B). HUVECs were treated for ten minutes with varying concentrations of angiopoietin-1 and -2, and then phospho-ERK was assayed with immunoblotting. (C). Phospho-ERK levels from (B) were analyzed with an ELISA.

Figure 6

Vaccine induced antibodies block angiopioetin-1/2 signaling. HUVECs were treated for ten minutes with 50 ng/ml of angiopoietin-1 or 200 ng/ml of angiopoietin-2 in the presence of PBS or 1 mg/ml of purified immunoglobulins from pre- or post-vaccination sera (with and without absorption with recombinant cytokine). Cell lysates were assayed for phospho-ERK with an ELISA. (A). Stage IV melanoma patient M30. (B). Stage IV lung carcinoma patient L19. Similar results were observed in two experiments.

Figure 7

Immunotherapy stimulated blocking antibodies to MIF. (A). Longitudinal analysis of anti-MIF antibodies in metastatic melanoma patient MEL15 (sera 1:500). Upward arrows denote vaccinations, downward arrows indicate anti-CTLA-4 mAb (MDX-010) infusions. (B). Longitudinal analysis of stage IV ovarian carcinoma patient OV65. (C). MEL15 sera obtained after immunotherapy inhibited MIF induced Tie-2 expression on monocytes. (D). Late MEL15 sera attenuated MIF stimulated MMP-9 production. Similar results were observed in three experiments; p=0.0314 for MEL15 early versus late sera.