Physiological levels of tumstatin, a fragment of collagen IV alpha3 chain, are generated by MMP-9 proteolysis and suppress angiogenesis via alphaV beta3 integrin

Abstract

We demonstrate a physiological role for tumstatin, a cleavage fragment of the alpha3 chain of type IV collagen (Col IValpha3), which is present in the circulation. Mice with a genetic deletion of Col IValpha3 show accelerated tumor growth associated with enhanced pathological angiogenesis, while angiogenesis associated with development and tissue repair are unaffected. Supplementing Col IValpha3-deficient mice with recombinant tumstatin to a normal physiological concentration abolishes the increased rate of tumor growth. The suppressive effects of tumstatin require alphaVbeta3 integrin expressed on pathological, but not on physiological, angiogenic blood vessels. Mice deficient in matrix metalloproteinase-9, which cleaves tumstatin efficiently from Col IValpha3, have decreased circulating tumstatin and accelerated growth of tumor. These results indicate that MMP-generated fragments of basement membrane collagen can have endogenous function as integrin-mediated suppressors of pathologic angiogenesis and tumor growth.

Figure 1. Tumstatin and Col IVα3/tumstatin^−/− mice

A: Schematic illustration of the type IV collagen α3 chain and tumstatin. Tumstatin (solid shading) is indicated in relation to type IV collagen α3 chain. B: Circulating levels of tumstatin in mouse blood. The plasma from ten wild-type (T^+/+) and ten Col IVα3/tumstatin^−/− (T^−/−) mice was examined using direct ELISA with anti-tumstatin antibody. Standard curve was calculated using recombinant tumstatin protein and anti-tumstatin antibody. ** indicates p < 0.01; compared to T^−/−. The results are shown as the mean ± SEM. C and D: Litter size and development of Col IVα3/tumstatin^−/− mice. To evaluate litter size at birth, an average number of eight separate litters delivered by T^+/+ and T^−/− parents was counted. In order to evaluate early development of pups, a 6 week survival rate for eight separate litters from both groups was calculated. The results are shown as the mean ± SD.

Figure 2. Tumor burden studies in Col IVα3/tumstatin-deficient mice

A: LLC tumor growth in Col IVα3/tumstatin^−/− mice. Twelve age- and sex-matched Col IVα3/tumstatin^−/− mice and six wild-type mice were used in this experiment. Recombinant tumstatin protein was intravenously injected into seven Col IVα3/tumstatin^−/− mice daily (300 ng) for 8 days starting at day 18 in sterile PBS, while only PBS was injected into the other five Col IVα3/tumstatin^−/− mice. Data are representative of four such independent experiments. The results are shown as the mean ± SEM. ** indicates p < 0.01; * indicates p < 0.05; compared to Col IVα3/tumstatin^−/− mice without injection. Scale bar: 1 cm. B: Matrigel plug assay in Col IVα3/tumstatin^−/− mice. Sections of each Matrigel plug were stained with HE and the numbers of blood vessels (arrow) in ten fields at 200× magnification were counted. Each column represents the mean ± SEM of six mice in each group. * indicates p < 0.05; compared to Col IVα3/tumstatin^−/− mice. Scale bar: 50 μm. C: Blood vessel quantification in LLC tumors. Frozen sections (4 μm) from tumor tissue were stained with anti-CD31 antibody, and the number of CD31-positive blood vessels (arrow) was counted. The results are shown as the mean ± SEM. ** indicates p < 0.01; compared to Col IVα3/tumstatin^−/− mice. Scale bar: 50 μm. D: VEGF receptor 2 (VEGFR2)-positive circulating endothelial cells in mice. The blood cells from peripheral blood attached to the slide were stained with anti-VEGFR2 antibody, and the number of positive cells (arrow) was counted. The results are shown as the mean ± SEM. ** indicates p < 0.01; compared to Col IVα3/tumstatin^−/− mice. Scale bar: 50 μm.

Figure 3. Repair-associated angiogenesis in Col IVα3/tumstatin-deficient mice

A: Skin wound healing in Col IVα3/tumstatin^−/− mice. One centimeter diameter skin wounds were induced in five wild-type (T ^+/+), five Col IVα3/tumstatin^+/− (T^+/−), and five Col IVα3/tumstatin^−/− (T^−/−) mice. The areas were measured periodically as indicated and the ratio in relation to the original wound size at day 0 was calculated at each time point. There was insignificant difference in the wound healing among the groups at all time points. The results are shown as the mean ± SEM. The PAS-stained wound areas at day 10 are shown here. Scale bar: 50 μm. B: Quantification of blood vessels in the skin. Frozen sections (4 μm) were stained with the anti-CD31 antibody and the numbers of CD31-positive blood vessels (arrow) were counted. The results are shown as the mean ± SEM. Scale bar: 50 μm. C and D: Liver regeneration in Col IVα3/tumstatin^−/− mice. Seventy percent of the liver was removed in T^+/+, T^+/−, and T^−/− mice with five mice per group. The proliferating hepatic cells were double-stained with BrdU and DAPI (arrow). Frozen sections (4 μm) from liver tissue were stained with anti-von Willebrand Factor (vWF) antibody to evaluate blood vessels (arrow). The results are shown as the mean ± SEM. Scale bar: 50 μm.

Figure 4. Tumstatin activity in β3 integrin-deficient mice

A: Proliferation assay in β3 integrin^−/− endothelial cells. MLEC were isolated from wild-type (β3^+/+) and β3 integrin^−/− (β3^−/−) mice as previously described (Maeshima et al., 2002). Proliferation was assessed by [³H]-thymidine incorporation. All groups represent triplicate analyses. The results are shown as the mean ± SEM. ** indicates p < 0.01; compared to control cells. B: Expression of β1 and β3 integrins in blood vessels of Matrigel plug. β1 and β3 integrin-positive blood vessels are indicated with arrows. Scale bar: 50 μm. C: Matrigel plug assay in β3 integrin^−/− mice. Sections of each Matrigel plug were stained with HE and the number of blood vessels (arrow) was counted. Each column represents the mean ± SEM of 2–4 mice in each group. Data are representative of two independent Matrigel plug experiments. ** indicates p < 0.01; compared to each control groups. † indicates p < 0.05; compared to the control group in β3 integrin^+/+ mice. Scale bar: 50 μm.

Figure 5. Expression of β3 and β1 integrins in tumors, healing skin wound, and regenerating livers

A: Immunohistochemical staining with anti-β3 integrin (β3) antibody and CD31 or vWF was performed to assess localization of β3 integrin in blood vessels. Anti-CD31 antibody was used for the tumor and skin tissues and anti-vWF antibody was used for the liver tissue. β3 integrin-positive blood vessels (merge) are indicated with arrows. Scale bar: 50 μm. B: β1 integrin expression in the tumor, skin wound, and regenerating liver tissues. β1 integrin-positive blood vessels (merge) are shown here. β1 integrin-positive vessels are indicated with arrows. Scale bar: 50 μm.

Figure 6. Expression of β3 and β1 integrins in tumors of various sizes

A: Immunohistochemical staining with anti-β3 integrin (β3) antibody and CD31 was performed to assess localization of β3 integrin in the tumor blood vessels with volumes of 0.1, 0.5 and 3.0 cm³ in wild-type mice. β3 integrin-positive blood vessels (merge) are indicated with arrows. Scale bar: 50 μm. B: β1 integrin expression in tumors with volumes of 0.1, 0.5, and 3.0 cm³. β1 integrin-positive vessels (merge) are indicated with arrows. Scale bar: 50 μm. C: Quantification of β1/β3 integrin-positive blood vessels in the tumor. The ratio of β3 integrin-positive blood vessels to CD31-positive vessels (open bar) was calculated in the tumors with various volumes. As a control, the ratio of β1 integrin-positive blood vessels to vWF-positive vessels (closed bar) was also calculated. ** indicates p < 0.01; compared to β3 integrin-positive blood vessels in 0.1 cm³ of tumor. †† indicates p<0.01; compared to β3 integrin-positive blood vessels in 0.2 cm³ of tumor.

Figure 7. Degradation of basement membrane preparation by active MMP-9

A: Western blot of BM degraded by MMP-9. MMP-9 enzyme was added to GBM at a 1:100; enzyme:substrate ratio, and the generation of tumstatin was detected by SDS-PAGE and immunoblotting using anti-tumstatin antibody. To establish MMP specificity in generation of tumstatin, 20 mM EDTA was added to the MMP/GBM mixture to inactivate MMP. The expected dimer (black arrow) and monomer (white arrow) of tumstatin are indicated. Monomers and dimers of type IV collagen NC1 domain have been previously described (Kalluri et al., 1997b). B: Quantification of tumstatin production by MMP-9 by ELISA. MMP-9 enzyme was added to 1 mg of GBM, and the production of soluble tumstatin in the supernatant was analyzed by ELISA using anti-tumstatin antibody. The results are shown as the mean ± SD. ** indicates p < 0.01; compared to the control group without enzyme.

Figure 8. Enhanced tumor growth in MMP-9-deficient mice

A: LLC tumor growth in MMP-9^−/− mice. Eight age- and sex-matched MMP-9^−/− mice and ten wild-type mice were used for this experiment. While at the very early stages (between 50 mm³ and 100 mm³), MMP-9^−/− tumors grew slightly slower; after that (especially >500 mm³), they grew at a much faster rate compared to wild-type tumors. Circulating levels of tumstatin were 350 ng/ml in wild-type mice compared to 141 ng/ml in the MMP-9^−/− mice. When the tumor volume reached 2.0 cm³, tumstatin was intravenously injected into four MMP-9^−/− mice (200 ng/day) for 14 days. Their tumor growth was compared to wild-type tumors starting at a size of 2.0 cm³. For these experiments, the tumors in the wild-type and MMP-9^−/− mice were grown until the both sets of tumors reached 2.0 cm³ (the wild-type tumors take 28 days to reach this size, while tumors in MMP-9^−/− mice take 21 days to reach this size). The accelerated tumor growth in MMP-9^−/− mice was suppressed by exogenous tumstatin, when compared with wild-type mice. The results are shown as the mean ± SEM. ** indicates p < 0.01; * indicates p < 0.05; compared to wild-type mice. Scale bar: 1 cm. B: Blood vessel quantification in LLC tumors. Frozen sections (4 μm) from tumor tissue were stained with anti-CD31 antibody and the number of CD31-positive blood vessels (arrow) counted. The results are shown as the mean ± SEM. ** indicates p < 0.01; compared to MMP-9^−/− mice. Scale bar: 50 μm. C: VEGFR2-positive circulating endothelial cells in mice. Cells from peripheral blood were stained with VEGFR2 antibody and the numbers of positive cells (arrow) were counted. The results are shown as the mean ± SEM. ** indicates p < 0.01; compared to MMP-9^−/− mice. Scale bar: 50 μm.