Identification of candidate angiogenic inhibitors processed by matrix metalloproteinase 2 (MMP-2) in cell-based proteomic screens: disruption of vascular endothelial growth factor (VEGF)/heparin affin regulatory peptide (pleiotrophin) and VEGF/Connective tissue growth factor angiogenic inhibitory complexes by MMP-2 proteolysis

Abstract

Matrix metalloproteinases (MMPs) exert both pro- and antiangiogenic functions by the release of cytokines or proteolytically generated angiogenic inhibitors from extracellular matrix and basement membrane remodeling. In the Mmp2-/- mouse neovascularization is greatly reduced, but the mechanistic aspects of this remain unclear. Using isotope-coded affinity tag labeling of proteins analyzed by multidimensional liquid chromatography and tandem mass spectrometry we explored proteome differences between Mmp2-/- cells and those rescued by MMP-2 transfection. Proteome signatures that are hallmarks of proteolysis revealed cleavage of many known MMP-2 substrates in the cellular context. Proteomic evidence of MMP-2 processing of novel substrates was found. Insulin-like growth factor binding protein 6, follistatin-like 1, and cystatin C protein cleavage by MMP-2 was biochemically confirmed, and the cleavage sites in heparin affin regulatory peptide (HARP; pleiotrophin) and connective tissue growth factor (CTGF) were sequenced by matrix-assisted laser desorption ionization-time of flight mass spectrometry. MMP-2 processing of HARP and CTGF released vascular endothelial growth factor (VEGF) from angiogenic inhibitory complexes. The cleaved HARP N-terminal domain increased HARP-induced cell proliferation, whereas the HARP C-terminal domain was antagonistic and decreased cell proliferation and migration. Hence the unmasking of cytokines, such as VEGF, by metalloproteinase processing of their binding proteins is a new mechanism in the control of cytokine activation and angiogenesis.

FIG. 1.

Characterization of MMP-2 transfectants. (A) Medium conditioned for 48 h from immortalized Mmp2^−/− embryonic fibroblast transfectants expressing pGW1GH vector (vector) or human MMP-2 (MMP-2) with or without 20 μg/ml ConA was analyzed by gelatin zymography. Std, 5 ng recombinant human MMP-2. (B) Titration of ConA (0 to 20 μg/ml) was used to determine the lowest concentration necessary for robust activation of MMP-2 at 48 h. The migration of the zymogen form (proMMP-2) and active forms of MMP-2 is shown. The square indicates the concentration of ConA (14 μg/ml) used to activate MMP-2 in the ICAT experiments. (C) ICAT-labeled MMP-2 tryptic peptides identified at ≥99% confidence in 48-h-conditioned medium from ConA-treated Mmp2^−/− cells transfected with MMP-2 or vector alone are shown with their corresponding ICAT ratios. A ratio >1.0 reflects relative accumulation of the protein in conditioned medium.

FIG. 2.

Distribution of subcellular and extracellular proteins (A) identified by ICAT-labeled tryptic peptides detected in 48-h-conditioned medium from ConA-treated MMP-2 transfectants. For a full list of proteins from which these graphs were plotted, see Table S1 in the supplemental material. An ICAT ratio MMP-2/vector of ≥1.38 (B) indicates relative accumulation of potential substrate protein found as fragments in the conditioned medium, and an ICAT ratio of ≤0.76 (C) indicates relative depletion of proteins from conditioned medium by reduced expression or degradation, as was reflected by the smaller numbers of peptides identified per protein due to MMP-2 activity.

FIG. 3.

Confirmation of MMP-2 processing of human IGFBP-6 and CTGF. (A) Analysis of MMP-2 proteolysis of recombinant IGFBP-6 on silver-stained 15% SDS-PAGE gels. Arrows indicate cleaved protein fragments. Molecular weight markers (in thousands) are shown. ICAT ratios for IGFBP-6 identified in conditioned medium from Mmp2^−/− cells transfected with MMP-2 or vector alone and with ConA or treated with vehicle are shown, confirming that the increase in proteolytic fragments of this substrate occurs only in the MMP-2-expressing cells and not in the vector-transfected cells or due to ConA treatment alone. (B) MMP-2 cleavage of CTGF analyzed on a 10% SDS-PAGE gel. The schematic depicts the four domains of CTGF: IGFBP, von Willebrand factor type C (vWFC-1), thrombospondin type 1 repeat (TSP-1), and C-terminal cysteine knot (CYS), with the major MMP-2 cleavage site generating the 16,374-Da N-terminal fragment and the 18,745-Da C-terminal fragment indicated. Amino acid numbering commences from the mature protein. Arrows show full-length and cleaved proteins with corresponding MALDI-TOF MS-derived mass and Edman N-terminal sequence analysis. N/D, not detectable.

FIG. 4.

Validation of MMP-2 processing of human follistatin-like 1 protein and HARP in vitro. (A) MMP-2 cleavage of follistatin-like 1 (FSTL1) on silver-stained 15% SDS-PAGE gel. Arrows indicate cleaved protein fragments. Molecular weight markers (in thousands) are shown. ICAT ratios for follistatin-like 1 identified in conditioned medium from Mmp2^−/− cells transfected with MMP-2 or vector alone and treated with ConA or vehicle are shown, confirming that the increase in proteolytic fragments occurs only in the MMP-2-expressing cells and not in the vector-transfected cells or due to ConA treatment alone. (B) SDS-PAGE analysis of MMP-2 cleavage of HARP analyzed on a 15% Tris-Tricine gel. The schematic depicts the four domains of HARP, with the two MMP-2 cleavage sites indicated. The corresponding recombinant analogues of MMP-2 cleaved HARP and are shown below. Amino acid numbering commences from the mature protein. Arrows show full-length and cleaved proteins, with corresponding MALDI-TOF MS-derived masses and N-terminal sequences determined by Edman sequencing. N/D, not detectable.

FIG. 5.

MMP-2 processing of CTGF is an inactivating cleavage of protein synthesis stimulation. Mmp2^−/− embryonic fibroblasts were cultured with CTGF or MMP-2-processed CTGF (0, 10, and 100 ng/ml) for 48 h. Protein concentration of the 48-h-conditioned medium was analyzed by bicinchoninic acid assay.

FIG. 6.

Proteolytic cleavage of HARP by MMP-2 inactivates HARP induction of cell proliferation and migration. Mitogenesis of serum-starved NIH 3T3 cells was determined by measuring [³H]thymidine incorporation over 18 h in the presence of (A) 4 nM recombinant HARP, HARP cleaved by MMP-2, MMP-2 alone, or buffer or (B) 40 nM N-HARP or C-HARP domains alone. Human umbilical vein endothelial cell migration was studied using Transwell chambers. Cells were incubated in the presence of (A) 4 nM recombinant HARP, HARP cleaved by MMP-2, MMP-2 alone, or buffer or (B) 40 nM N-HARP or C-HARP domains alone. After 4 h of incubation at 37°C, the migrated cells were stained using May-Grünwald-Giemsa solution and quantified by counting in three high-power microscopic fields (HPF)/well magnified 100×. Quantification of endothelial cell migration was determined as the mean of three fields/well for two independent experiments. Standard error bars are shown. **, Student's t test values (P < 0.01) comparing 4 nM HARP with all other conditions.

FIG. 7.

MMP-2-proteolyzed HARP antagonizes HARP-induced cell proliferation and migration. HARP (4 nM) stimulation of mitogenesis of serum-starved NIH 3T3 cells was determined by measuring [³H]thymidine incorporation over 18 h. (A) MMP-2-proteolyzed HARP (4 nM) was added with 4 nM HARP. Concentrations of 0 to 40 nM recombinant N-HARP (C) and C-HARP with or without 4 nM HARP (D) were also studied. Cell migration was quantified using Transwell chambers. (B) Human umbilical vein endothelial cells were incubated in the presence of MMP-2-proteolyzed HARP (4 nM) with 4 nM HARP. After 4 h of incubation at 37°C, the migrated cells were stained using May-Grünwald-Giemsa solution and quantified by counting in three high-power microscopic fields (HPF)/well magnified 100×. Quantification of endothelial cell migration was determined as the mean of three fields/well for two independent experiments. Standard error bars are shown. ** and *, Student's t test P values of <0.01 and <0.05, respectively.

FIG. 8.

MMP-2 proteolysis releases intact VEGF from HARP-VEGF and CTGF-VEGF complexes. (A) HARP or CTGF was complexed with VEGF for 24 h at 4°C and then incubated with MMP-2 for 24 h at 37°C. VEGF, HARP, and CTGF were also incubated with MMP-2 for 24 h at 37°C. The resulting cleavage fragments were analyzed by 15% Tris-Tricine SDS-PAGE and silver stained. HARP and CTGF were cleaved by MMP-2 whether incubated alone or when complexed with VEGF and generated identical cleavage fragments, indicating that VEGF did not mask these cleavage sites. VEGF was not processed by MMP-2. VEGF and CTGF comigrate at 38 kDa. Arrows indicate full-length and cleaved protein fragments. (B) VEGF was complexed with HARP or CTGF on a microtiter plate for 24 h prior to a 24-h incubation at 37°C with activated MMP-2. VEGF ELISA was used to quantify the residual VEGF bound to CTGF or HARP after MMP-2 cleavage.

FIG. 9.

Poteolytic unmasking of growth factor activities by MMP cleavage of cytokine binding proteins. With the four new substrates of MMP-2 discovered here (HARP, CTGF, IGFBP-6, and follistatin-like 1) we suggest that a common role for MMPs, and MMP-2 in particular, is to proteolytically liberate active cytokines from inhibitory complexes. Depicted in the schematic are the cleavage and inactivation of HARP/pleiotrophin and CTGF from inhibitory complexes with VEGF. Upon cleavage, VEGF and other cytokines (black squares) are mobilized and free to interact in an autocrine or paracrine manner with cells. These potent growth factor-regulatory activities complement any matrix-degradative roles in facilitating the neovascularization shown here or in other biological processes.