A novel antiangiogenic and vascular normalization therapy targeted against human CD160 receptor

Abstract

Angiogenesis plays an essential role in several diseases of the eye and in the growth of solid tumors, but existing antiangiogenic therapies have limited benefits in several cases. We report the antiangiogenic effects of a monoclonal antibody, CL1-R2, in several animal models of neovascularization. CL1-R2 recognizes human CD160, a membrane receptor which is conserved in various mammal species. We show that CD160 is expressed on the endothelial cells of newly formed blood vessels in human colon carcinoma and mouse B16 melanoma but not in vessels of healthy tissues. CL1-R2 reduced fibroblast growth factor 2-induced neovascularization in the rabbit cornea, in a mouse model of oxygen-induced retinopathy, and in a mouse Matrigel plug assay. Treatment of B16 melanoma-bearing mice with CL1-R2 combined with cyclophosphamide chemotherapy caused regression of the tumor vasculature and normalization of the remaining vessels as shown by Doppler ultrasonography, intravital microscopy, and histology. These studies validate CD160 as a potential new target in cases of human pathological ocular and tumor neoangiogenesis that do not respond or become resistant to existing antiangiogenic drugs.

Figure 1.

CL1-R2 inhibits FGF2-induced rabbit corneal neoangiogenesis. (A) Representative photographs of corneas treated with control IgG1 (Ctrl) or CL1-R2 mAb. Neovascularization was assessed 8 d after corneal grafting of FGF2-containing implants. Arrows, implants; arrowhead, neovessels. Bar, 2 mm. (B) Quantitative analysis of neovascularization in Ctrl- or CL1-R2–treated corneas. Values are means ± SEM obtained from four independent experiments, n = 5 rabbits/group/experiment. ***, P < 0.0001 (Mann-Whitney U test).

Figure 2.

Intravitreal injection of CL1-R2 reduces retinal neovascularization in a murine model of oxygen-induced retinopathy. Retinal neovascularization was induced by exposing 7-d-old mice to 75 ± 2% oxygen for 5 d, and then to normoxia. Animals were injected intravitreally with control IgG1 or CL1-R2 or received no injection. (A) Flat-mounted retinas harvested from 17-d-old pups were perfused with FITC-dextran and observed by fluorescence microscopy to assess retinal vasculature. Representative flat-mounted retinas are from normal mice (Normal retina), mice exposed to hyperoxic conditions without injection (Mock), or mice injected intravitreally with 5 µg IgG1 (Ctrl) or 5 µg CL1-R2. Central avascular zone, arrowheads; neovascular tufts, asterisks. Data are representative of three independent experiments (n = 6–11 mice/group). Bars: (top) 770 µm; (bottom) 335 µm. (B) Histological analyses (periodic acid–Schiff staining) of whole eye tissue sections. Representative photomicrographs of eye sections from normal mice (Normal retina; n = 3) and mice exposed to hyperoxic conditions without injection (Mock; n = 6) or injected intravitreally with IgG1 (Ctrl; n = 9) or CL1-R2 (n = 11). Longitudinal and transverse aberrant microvessels, arrows; isolated nuclei of endothelial cells not involved in tube formation, asterisks. GCL, ganglion cell layer; INL, inner nuclear layer. Bar, 50 µm. (C) Von Willebrand Factor staining of whole eye tissue sections. Representative photomicrographs of eye sections from normal mice (Normal retina) and mice exposed to hyperoxic conditions without injection (Mock) or injected intravitreally with IgG1 (Ctrl) or CL1-R2 mAb. After paraffin removal, sections were rehydrated, treated with 0.3% Triton X-100, and incubated with polyclonal antibody against Von Willebrand factor, followed by Alexa Fluor 488 secondary antibody (green) and counterstaining with DAPI (blue). GCL, ganglion cell layer; INL, inner nuclear layer. Bar, 25 µm. (D and E). Quantitative assessment of the retinal vascularization illustrated in B for 6 (Mock), 9 (Ctrl), or 11 (CL1-R2) mice/group. Endothelial cell nuclei and vessel lumens were counted on four to eight sections per eye stained with periodic acid–Schiff reagent. The mean number of endothelial cell (EC) nuclei (D) and vessel lumen (E) presented in the histograms were determined using a Poisson regression model for clustered data. 95% confidence intervals of the mean number estimates are figured as error bars. P-values were corrected for post-hoc group comparisons by the Bonferroni method. **, P < 0.001.

Figure 3.

CL1-R2 inhibits FGF2-induced neoangiogenesis in a Matrigel plug model. (A, Top) Photographs of representative FGF2-containing Matrigel plugs taken from control IgG1- and CL1-R2–treated mice 7 d after plug implantation. (A, Bottom) Representative sections of FGF2-containing Matrigel plugs taken from control IgG1- and CL1-R2–treated mice on day 7. Endothelial cells were labeled with isolectin B4 (brown) and the sections were counterstained with hematoxylin (blue). Arrowheads indicate isolectin B4-positive vessels. Bar, 50 µm. Representative images of five independent experiments. (B) Quantification of hemoglobin in FGF2-containing Matrigel plugs removed from control IgG1- and CL1-R2–treated mice. Horizontal bars represent the mean hemoglobin content. n = 11 IgG1-treated mice; n = 22 CL1-R2–treated mice. *, P = 0.0153 (Mann-Whitney U test). Pooled data are from five independent experiments.

Figure 4.

Diminished vascular density of B16 melanoma tumors in CL1-R2–treated mice. (A) Immunohistochemical assessment of vascular density in B16 tumor sections taken on day 16 from mice treated with control IgG1 (Ctrl) or CL1-R2, both in combination with cyclophosphamide (CycloP). Endothelial cells in representative sections were stained with isolectin B4 (brown) and counterstained with hematoxylin (blue). Images are representative of five mice in each group. Bar, 50 µm. (B) Quantification of the data shown in A. Vessel density was counted in four independent areas on each B16 tumor section. ***, P < 0.0001 (one-way ANOVA test). Error bars represent SEM. (C) 3 d after B16 cell injection, mice received repeated i.p. injections of CL1-R2 or IgG1 (Ctrl). Cyclophosphamide was then added to the drinking water. Mice were examined by DUS to identify intratumoral blood vessels 12, 15, and 18 d after the first CL1-R2 injection. Power Doppler mode color DUS ultrasound images with intratumoral blood vessel signals are shown in red/yellow. Representative images obtained from the same tumor cross section in a single mouse from day 12 to 18 are shown for each treatment. Bar, 5 mm. Side bar: Doppler power, maximum 4.3. (D) Quantification of the number of intratumoral vessels in IgG1-treated (black bars) and CL1-R2–treated (gray bars) mice determined from multiple cross sections of the tumor. Tumor vessel number on the ordinate represents the mean number of total intratumoral blood vessels per tumor because the whole tumor was scanned at each time point for the detection of blood vessels. Data are expressed as means ± SEM (n = 5 mice in each group). *, P = 0.0203 for day 12; **, P = 0.0032 for day 15; **, P = 0.0077 for day 18 (Student’s t test). (E) B16 tumor growth in untreated (PBS), CL1-R2–treated (500 µg, three times a week), or control mice in combination with cyclophosphamide treatment. Tumor volumes were measured at the indicated times after tumor inoculation. Data from four experiments (n = 5 mice/group) were pooled. Data represent mean tumor size values ± SEM. ***, P < 0.0001 (two-way ANOVA test).

Figure 5.

Normalization of B16-GFP melanoma tumor vessels in CL1-R2–treated mice observed by in vivo intravital fluorescence microscopy. (A, Left) Host vessels surrounding the tumor in control Ig or CL1-R2–treated mice in combination with cyclophosphamide at day 9 after tumor implantation. Contrast was enhanced by i.v injection of 150 kD TRITC-labeled dextran. Arrowheads and arrows point to the same tumor vessels at day 0 (d0) and day 9 (d9), respectively. Bars, 40 µm. (A, Right) Quantification of the data in the left panel. Data represent mean values ± SEM (n = 5 mice/group). ***, P < 0.0001 (two-way ANOVA test). (B, Left) Tumor vascular area 12 or 14 d after CL1-R2 treatment or control IgG1. Bars, 40 µm. (B, Right) Quantification of the tumor vascular areas in the left panels. Values are means ± SEM (n = 5 mice/group). *, P = 0.0114 (day 12); *, P = 0.0292 (day 14; Student’s t test). (C, Left) Intratumoral microvasculature in day-14 control IgG1- or CL1-R2–treated mice. Bars, 40 µm. (C, Right) Quantitative analysis of vascular network complexity based on microvessel fractal dimension measurements in IgG1-treated (Ctrl) and CL1-R2–treated mice. Values are means ± SEM (n = 5 mice/group). *, P = 0.0211 (Student’s t test). (D, Left) Microangiography of IgG1-treated (Ctrl) and CL1-R2–treated tumors on day 14 after implantation. Bars, 40 µm. (D, Right) Quantitative data from a representative experiment (plotted in the box/whisker format) comparing IgG1-treated (Ctrl) and CL1-R2–treated tumors on day 14 after implantation (n = 5 mice/group). Box plots show median values (horizontal line inside the whiskers box), box boundaries represent quartiles, and error bars represent the lowest and highest values. Images and quantitative data are representative of three independent experiments.

Figure 6.

The B16 tumor vasculature in CL1-R2–treated mice is mature and functional. All data presented were obtained from mice 16 d after tumor implantation (A, Left) Representative B16 tumor vessels stained with isolectin B4 (brown staining; counterstaining, hematoxylin) from mice treated with CL1-R2 or control Ig in combination with cyclophosphamide. Bar, 50 µm. (A, Right) Histogram showing the number of vessels with an open lumen. Number of fields: four. n = 5 mice per group. *, P = 0.0341 (Student’s t test). (B, Left) CD34-positive endothelial cells (green) are covered by SMA (red) pericyte cells in B16 tumors from CL1-R2–treated mice (bottom) compared with IgG1-treated Ctrl mice (top). Nuclei are labeled blue. Images were taken with confocal fluorescence microscopy. Bar, 20 µm. (B, Right) Quantification of the number of endothelial cells covered by pericytes. The ratio of total area red staining (SMA) to green staining (CD34) was calculated (10 fields per tumor, 5 tumors). Values are means ± SEM. ***, P < 0.0001. (C, Left) Cryosections of B16 tumors from control and CL1-R2–treated mice stained with isolectin B4 (red). Hoechst staining (blue) indicates perfused vessels. Insets show images at higher magnification. Bars, 50 µm. (C, Right) Quantification. n = 5. Number of fields: 4. **, P = 0.0047 (unpaired Student’s t test with Welch’s correction). Control = 24.59 ± 7.086 (n = 5); CL1-R2 = 65.52 ± 1.158 (n = 5). (D, Left) Histopathology of B16 tumors stained with hematoxylin and eosin. Necrosis is indicated by arrowheads. Bar, 200 µm. (D, Right) Quantification of necrotic areas in B16 tumors from CL1-R2–treated and control mice. n = 5 mice/group. Number of fields: 4. **, P = 0.0092 (Student’s t test with Welch’s correction). Images and quantitative data are representative of three independent experiments. The data in A, C, and D are means ± SEM of four different microscope fields for each tumor and a total of five different tumors in each treatment group.

Figure 7.

Expression of CD160 in tumor blood vessels but not in the blood vessels of healthy tissues. (A) Representative sections of B16 melanoma at day 16 after their subcutaneous injection (top) and healthy mouse heart (bottom), both stained with either CL1-R2 (CD160) or isolectin B4, an endothelial cell marker. Bar, 50 µm. Images are representative of four independent experiments (five mice/group). (B) Sections of human colon tumor (top) and healthy colon (bottom) stained with CL1-R2 (CD160) or a mAb against CD31, a human endothelial cell marker. Images are representative of two patient biopsies. Control IgG1 did not stain and is not shown. Bar, 50 µm.