Disruption of Endothelial Cell Interactions with the Novel HU177 Cryptic Collagen Epitope Inhibits Angiogenesis

Angiogenesis or the formation of new blood vessels from preexisting vessels is a complex process involving a series of interconnected events (1–4). Although progress has been made in understanding key mechanisms that regulate angiogenesis, much still remains to be learned. In this regard, it is well established that endothelial cell interactions with extracellular matrix play important roles in the development of new blood vessels (5–8). In fact, studies on mutant and transgenic mice defective in expression of extracellular matrix components as well as mice defective in proteolytic enzymes have provided crucial insight into the importance of collagen metabolism in the angiogenic cascade (9–12). Although remodeling the extracellular matrix is a critical event in angiogenesis, the mechanisms by which extracellular matrix remodeling controls new blood vessel development are not completely understood. Studies have suggested that extracellular matrix remodeling may contribute to angiogenesis by releasing matrix-sequestered growth factors as well as generating soluble bioactive peptides that may stimulate or inhibit angiogenesis (13–15). Other studies suggest that proteolytic remodeling of extracellular matrix proteins may act to remove restrictive barriers and promote endothelial cell motility and invasion (16, 17). Although it is likely that a combination of mechanisms contribute to the regulation of vessel development, much less is known concerning the selective exposure and functional relevance of matrix-immobilized cryptic extracellular matrix epitopes to the angiogenic cascade.

In our previous studies, we used the immunologic technique of subtractive immunization to generate a series of monoclonal antibodies (mAb) specifically directed to cryptic collagen epitopes (18). One of the mAbs (HUIV26) recognized a cryptic epitope present exclusively in basement membrane collagen type IV (18, 19). The HUIV26 cryptic epitope could be exposed following thermal denaturation and proteolysis and was shown to be expressed within the extracellular matrix of angiogenic but not quiescent blood vessels (19–22). Exposure of this cryptic site played a role in angiogenesis because antibodies directed to this non-cellular element inhibited angiogenesis and tumor growth in animal models (19–22).

Given the complexity of the extracellular matrix, it is likely that additional functional epitopes exist that play unique roles in angiogenesis. We previously generated the mAb HU177 that selectively binds denatured and proteolyzed collagen (19). Studies have suggested that the HU177 cryptic epitope could be exposed within three-dimensional collagen gels following tumor- and vascular smooth muscle cell–mediated proteolysis (23, 24). In addition, other studies have shown the expression of the HU177 cryptic epitope within the medial wall of aortic explants from mice following ischemic injury and within the extracellular matrix of developing murine papillary membranes during angiogenesis (24, 25). We recently showed the selective temporal exposure of the HU177 epitope within ischemic muscle tissue, and this expression correlated with enhanced matrix metalloproteinase levels (26). However, to date, no evidence exists that the HU177 cryptic epitope plays a functional role in regulating invasive cellular behavior, or whether this epitope represents a clinically useful therapeutic target. In this regard, here, we provide experimental evidence that this second cryptic collagen epitope (HU177) is present in both interstitial collagen type I and basement membrane collagen type IV. Evidence is presented that this cryptic collagen epitope may be comprised of a peptide that includes the amino acid residues LPGFPG. Importantly, the HU177 cryptic epitope was selectively exposed not only in the interstitial matrix of tumors but also in the extracellular matrix of angiogenic but not quiescent vessels in vivo and played an important role in the ability of endothelial cells to attach and migrate on denatured collagen type IV. A mAb directed to the HU177 cryptic site selectively inhibited endothelial cell proliferation and induced the expression of the cyclin-dependent kinase (CDK) inhibitor p27^kip1. mAb HU177 inhibited endothelial cord formation in vitro and cytokine- and tumor-induced angiogenesis in vivo. Taken together, these novel findings provide evidence for a new physiologically relevant cryptic element and provide further support for the concept that targeting specific cryptic elements may represent an effective new strategy for the treatment of pathologic angiogenesis.

Antibodies and reagents. mAbs HU177 and HUIV26 were generated in our laboratory and have been described previously (18, 27). Normal mouse IgM was obtained from Pierce. Anti-p27^KIP1 and anti-β-tubulin antibodies were obtained from Santa Cruz Biotechnology. Anti-CD-31 antibody was obtained from BD PharMingen. Rhodamine and FITC-labeled Lycopersicon esculentum lectin were obtained form Vector Laboratories. Horseradish peroxidase–labeled goat anti-mouse and goat anti-rabbit antibodies were from Biosource. Purified collagen type I, collagen type IV, and fibronectin were obtained from Sigma. Growth factor–reduced Matrigel was obtained from Calbiochem. Bovine serum albumin (BSA), methanol, ethanol, acetone, and ornithine carbamyl transferase compound were obtained from Sigma. Synthetic peptides including p34 peptide (N-LPGFPGVAGPPGITGFPGFIGSRG-C), LPG peptide (N-LPGFPGC-C), and CVT peptide (N-CVTQGGLLMSRMI-C) were obtained from QED Bioscience.

Cells and cell culture. Human umbilical vein endothelial cell (HUVEC) were obtained from the American Type Culture Collection. CS-1 melanoma cells were a gift from Dr. David Cheresh (Scripps Research Institute, La Jolla, CA). CS-1 melanoma cells were maintained in RPMI (Life Technologies) in high glucose supplemented with 5% fetal bovine serum, glutamine, and penicillin/streptomycin at 37°C with 5% CO₂. HUVECs were maintained in endothelial cell growth medium in (Clontech) high glucose supplemented with 20% fetal bovine serum, endothelial cell growth supplements, and glutamine and penicillin/streptomycin at 37°C with 5% CO₂. All cells were maintained as subconfluent cultures and split 1:3, 24 h before use.

Solid-phase ELISA. Briefly, 96-well plates were coated (100 μL per well) with native or denatured extracellular matrix proteins including collagen type I (10.0 μg/mL), collagen type IV (10 μg/mL), fibronectin (10.0 μg/mL), or Matrigel (10.0 μg/mL) for 18 h at 4°C (18). The plates were blocked with 2.5% BSA in PBS. mAb HU177 or control antibody were resuspended (1.0 μg/mL) in 1.0% BSA in PBS and incubated for 1 h at 37°C. The plates were washed and incubated with horseradish peroxidase–conjugated goat anti-mouse secondary. The plates were washed and ELISA substrate was added (0.25 mg/mL o-phenylendiamine/H₂O₂) in citrate buffer (pH 5.0). In competition ELISA, plates were coated with denatured collagen type IV, and mAbs HU177 and HUIV26 (0.5 μg/mL) were resuspended in the presence or absence of LPG peptide of control nonspecific CVT peptide (500 μg/mL). Absorbance values were measured with an ELISA reader at 490 nm.

Cell adhesion assays. Briefly, 48-well plates were coated with native or denatured collagen type IV (1 μg/mL) at 4°C for 18 h. Subconfluent HUVECs (1 × 10⁵) were harvested and suspended in adhesion buffer (RPMI 1640 containing 1 mmol/L MgCl₂, 0.2 mmol/L MnCl₂, and 0.5% BSA) in the presence or absence of mAb HU177 or isotype-matched control and added to wells and allowed to attach for 10 to 15 min at 37°C. Non-attached cells were removed by washing, and attached cells were stained with crystal violet. Cell adhesion was quantified by measuring the absorbance of eluted crystal violet at 600 nm (7).

Cell migration assay. Briefly, membranes (8.0-μm pore size) from Transwell chambers were coated with native or denatured collagen IV (2.0 μg/mL) for 12 h at 4°C. The Transwells were washed and blocked with 1.0% BSA in PBS for 1 h at 37°C. Endothelial cells (HUVECs) were resuspended in buffer containing RPMI 1640, 1 mmol/L MgCl₂, 0.2 mmol/L MnCl₂, and 0.5% BSA in the presence or absence of mAb HU177 or an isotype-matched control. Endothelial cells were allowed to migrate for 2 to 4 h. Cell migration was quantified by measuring the absorbance of cell-associated dye at 600 nm (7).

Cell proliferation assay. Non–tissue culture–treated 96-well plates were coated with denatured collagen IV (10 μg/mL) at 4°C for 18 h. Plates were blocked with 1% BSA in PBS. Subconfluent HUVECs were suspended in buffer containing 5% fetal bovine serum. Cells (5 × 10³ per well) were added to the wells in the presence or absence of mAb HU177 or an isotype-matched control (0-200 μg/mL) and incubated at 37°C for 24 h. Cell proliferation was quantified using the WST proliferation assay kit according to manufacturer's instructions.

Immunohistochemistry. Tumors were resected, washed, and embedded in OTC and snap frozen (27, 28). Frozen sections (4 μm) of melanoma were fixed by incubation for 30 s in 50% methanol 50% acetone. Tissues were blocked with 2.5% BSA followed by incubation with mAb HU177 (100 μg/mL) or polyclonal anti-CD31 in 2.5% BSA for 2 h at 37°C. Tissues were washed and incubated with FITC or rhodamine-conjugated secondary (1:300 dilution in 1.0% BSA in PBS). For colocalization studies, sections of CS-1 tumors were prepared as described and stained with rhodamine-labeled L. esculentum lectin (20 μg/mL) or anti-CD31 and mAb HU177 (100 μg/mL). For quantification of the percentage of HU177 staining tumor vessels, five distinct ×200 microscopic fields from three individual tumors were examined. The mean number of costained vessels compared with vessels lacking association of the HU177 epitope was counted within ×200 fields form each of three tumors. For apoptosis analysis, tissue sections were stained for apoptosis using Tunnel method (ApopTag) followed by staining with rhodamine-labeled L. esculentum lectin or anti-CD31. Photomicrographs were taken at a magnification of ×200 or ×630 under oil immersion. The relative levels of apoptosis were estimated by laser scanning image analysis of five ×200 fields from distinct tumors from each condition (7). Pixel densities from Kodak ID scans of stained tissues were quantified with Kodak ID version 4.0 image analysis software as described previously (7).

Matrigel endothelial cord formation assay. Briefly, 100 μL Matrigel (BD PharMingen) was added to wells of a 96-well culture plate and allowed to polymerize for 1 h at 37°C (28, 29). Subconfluent HUVECs (5 × 10³) were resuspended in 5% fetal bovine serum and added to the Matrigel in the presence or absence of mAb HU177 or an isotype-matched control (250 μg/mL). HUVECs were allowed to form endothelial cords for 24 h at 37°C. Endothelial cord branch points from four randomly selected microscopic fields (×200) were counted for each condition. Photographs were taken using a Sony digital camera.

Chick embryo angiogenesis assay. The chorioallantoic membranes of 10-day-old chicks (Charles River) were separated from the shell membrane (30). Filter discs containing buffer or basic fibroblast growth factor (40 ng) were place on the chorioallantoic membranes to induce angiogenesis. Twenty-four hours later, the embryos were treated by a single i.v. injection of mAb HU177 or an isotype-matched control (0-200 μg per embryo). The embryos were incubated for 3 days. At the end of the incubation period, the embryos were sacrificed, and the chorioallantoic membrane tissue was analyzed. Angiogenesis was quantified by counting the number of angiogenic branching blood vessels within the area of the filter disc. Seven to 10 embryos were used per condition.

Chick embryo tumor growth assay. CS-1 melanoma cells (5 × 10⁶) were seeded on the chorioallantoic membranes of 10-day-old chicks (27, 28). Twenty-four hours later, the embryos were injected with mAb HU177 or an isotype-matched control (200 μg per embryo). The embryos were incubated for 7 days. At the end of the incubation period, the embryos were sacrificed, tumors were removed, and wet weights were determined. Six to eight embryos were used per condition.

Chick embryo tumor angiogenesis assay. CS-1 melanoma cells were seeded on the chorioallantoic membranes of 10-day-old embryos (27, 28). Twenty-four hours later, the embryos were injected i.v. with mAb HU177 or an isotype-matched control (200 μg per embryo). At the end of the incubation period, the number of tumor associated branching blood vessels were counted with a stereomicroscope set at a defined focal length (×10). Vessel counts were made, and tumor angiogenesis was expressed as the mean number of branching blood vessels per tumor. Four to five tumors were counted for each condition and experiments.

Real-time quantitative reverse transcription-PCR. Briefly, culture plates were coated with denatured collagen type IV (10.0 μg/mL). Equal numbers of endothelial cells were washed and added to the coated plates in the presence or absence of mAb HU177 or an isotype-matched control (100 μg/mL) and incubated for 12 h. Total RNA was isolated using RNeasy miniprep columns (Qiagen). Real-time fluorescence detection was carried out using an ABI Prism 7900 Sequence Detection System. All primers and probes were designed using Primer3 version 2 and Ensembl. Primer sets: p27^KIP1, CTTGTCGCTGTCTTGCACTC (forward) and AATCTGTCAGGCTGGTCTGC (reverse); β₂-microglobulin, AAAGATGAGTATGCCTGCCG (forward) and CCTCCATGATGATGCTGCTTACA (reverse). Samples were analyzed in triplicate. β₂-Microglobulin was used as an endogenous internal control (31).

Western blot analysis. Plates were coated with denatured collagen type IV (10.0 μg/mL). Equal numbers of endothelial cells were harvested, washed, and added to the coated plates in the presence or absence of mAb HU177 or an isotype-matched control (100 μg/mL) and incubated for 12 h. Cells were harvested, washed, and lysed in 1.0% Triton X-100 buffer containing 300 mmol/L NaCl, 50 mmol/L Tris (pH 7.0), and 1× protease inhibitor cocktail. Equal amounts (15 μg per lane) of cell lysates were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were probed with polyclonal antibodies directed to p27^KIP1 or β-tubulin. Western blots were visualized by chemiluminescence detection.

Statistical analysis. Statistical analysis was done using the InStat statistical program for Macintosh computers. Data were analyzed for statistical significance using Student's t test. P < 0.05 was considered significant.

Exposure of the HU177 cryptic collagen epitope in vitro and in vivo. To characterize and study the expression of the HU177 epitope, solid-phase ELISAs were carried out. Microtiter wells were coated with purified extracellular matrix proteins, including collagen type I, collagen type IV, fibronectin, and the basement membrane preparation Matrigel in their intact or denatured forms. As shown in Fig. 1A, mAb HU177 specifically bound denatured forms of collagen type I and type IV while exhibiting little reactivity with intact collagen. The HU177 cryptic epitope was also exposed following denaturation of Matrigel, whereas little reactivity was observed with intact Matrigel. Importantly, mAb HUI77 failed to react with fibronectin in either its native or denatured forms.

In previous studies, the HU177 cryptic epitope was selectively expressed within the extracellular matrix ischemic muscle and tumor tissues, whereas little if any was exposed within normal tissues (26, 27). To extend these studies and to examine whether the HU177 cryptic site could also be exposed within the extracellular matrix of tumor vessels, we examined malignant hamster melanoma (CS-1) tumors growing on the chorioallantoic membranes of chick embryos. As shown in Fig. 1B (top), the HU177 cryptic epitope was detected (brown) within the CS-1 melanoma tumors (right), whereas little nonspecific staining was detected with an isotype-matched control antibody (left). To confirm that the HU177 cryptic epitope could be associated with tumor blood vessels, co-stain analysis was carried out using mAb HU177 and FITC-labeled tomato lectin known to bind blood vessels (20). The HU177 cryptic epitope (red) colocalized (yellow) with CS-1 tumor blood vessels (green; bottom). Although the extracellular matrix surrounding many tumor blood vessels stained positive for the HU177 epitope, not all vessels were positive. In fact, quantification of the relative percentage of CS-1 tumor vessels that was associated with the HU177 epitope ranged from ∼50% to 75% within any given microscopic field (Table 1). Interestingly, whereas the majority of the HU177 staining within these CS-1 tumors seemed to be associated with vascular basement membranes, some scattered staining surrounding individual cells were observed. These findings confirm our previous studies and suggest that the HU177 cryptic epitope is not only exposed within the interstitial matrix of malignant tumors but can also be selectively found in association with tumor blood vessels.

To begin to identify potential amino acids that may comprise the HU177 cryptic epitope, solid-phase ELISAs were carried out with synthetic 25-mer peptides corresponding to sequences within the α2(IV) chain of human collagen type IV. mAb HU177 bound to immobilized peptide p34, a 25-mer peptide corresponding to collagen type IV amino acid residues (1047-1070) in solid-phase ELISAs (data not shown). To determine whether a smaller region within this 25-mer sequence contributes to the HU177 epitope, we examined a 6-mer sequence (LPGFPG). The amino acid sequence LPGFPG present in the original p34 peptide is a common sequence that is repeated within collagen type IV α1(IV) and α2(IV) chains. This sequence as well as several similar sequences (LPG{X}PG) are present several times in collagen type I and IV. Therefore, we synthesized the 6-mer peptide LPGFPG and engineered a cysteine (C) on the COOH terminus to assist in immobilization. As shown in Fig. 1C, the LPG peptide inhibited mAb HU177 binding to the denatured collagen, whereas a control nonspecific CVT peptide (CVTQGGLLMSRMS) had little effect. To further confirm the specificity of the LPG peptide, we examine the ability of the LPG peptide to block binding of mAb HUIV26, an isotype-matched antibody that is specifically directed to a second cryptic epitope within collagen type IV. As shown in Fig. 1D and E, the LPG peptide inhibited binding of mAb HU177 but failed to inhibit mAb HUIV26 binding to denatured collagen IV. Although it is clear that more structural studies, peptide mapping, and sequencing analysis will be required to precisely identify the exact sequence of the HUI77 cryptic epitope, these data are consistent with the possibility that the LPGFPG sequence represents at least part of the HU177 cryptic collagen type IV site.

mAb HU177 inhibits endothelial cell adhesion on denatured collagen type IV. Given that the HU177 epitope can be selectively exposed within tumor blood vessels, we chose to focus our functional characterization of the HU177 cryptic epitope within basement membrane collagen type IV. To examine whether the exposure of the HU177 cryptic site plays a role in endothelial cell interactions with denatured collagen IV, cell adhesion assays were carried out. Microtiter wells were coated with intact or denatured collagen type IV, and endothelial cells were allowed to bind in the presence or absence of mAb HU177 or an isotype-matched control. As shown in Fig. 2A, HUVECs bound to native collagen type IV and mAb HU177, and the control antibody had little effect over the dose range tested (0-200 μg/mL). In contrast, HUVEC adhesion to denatured collagen type IV was inhibited dose dependently with maximum inhibition (50%) observed at a concentration of 200 μg/mL (Fig. 2B). Identical concentrations of the isotype-matched control antibody had little effect.

mAb HU177 inhibits endothelial cell migration on denatured collagen type IV. To examine the effects of mAb HU177 on motility, HUVECs were allowed to migrate on coated membranes in the presence or absence of mAb HU177 or an isotype-matched control antibody (0-200 μg/mL). As shown in Fig. 3A, HUVECs migrate on native collagen type IV, and mAb HU177 and an isotype-matched control antibody had little effect. In contrast, mAb HU177 dose dependently inhibited HUVEC migration on denatured collagen type IV, with maximal inhibition of ∼50%, compared with either no treatment or incubation with an isotype-matched control (Fig. 3B). Collectively, these novel findings suggest that endothelial cells can attach and migrate on denatured collagen type IV, and that this migratory behavior is dependent in part on interactions with the HU177 cryptic site.

mAb HU177 selectively inhibits endothelial cell proliferation and up-regulates the CDK inhibitor p27^KIP1. Previous studies suggest that cellular interactions with extracellular matrix can regulate proliferation and expression of cell cycle control proteins such as CDK inhibitors (32–34). To examine the effects of mAb HU177 on HUVEC proliferation, we used the WST-1 proliferation assay. Microtiter wells were coated with native or denatured collagen type IV, and HUVECs were allowed to proliferate for 24 h in the presence or absence of mAb HU177 or an isotype-matched control (0-200 μg/mL). As shown in Fig. 4A, mAb HU177 and an isotype-matched control had little effect on HUVEC proliferation on native collagen type IV. In contrast, mAb HU177 inhibited HUVEC proliferation, with maximal inhibition of ∼50% compared with either no treatment or incubation with an isotype-matched control antibody (Fig. 4B). In similar studies, mAb HU177 showed little, if any, effect on CS-1 melanoma tumor cell proliferation in vitro, which was not surprising given that CS-1 tumor cells failed to interact with the HU177 cryptic epitope (data not shown).

Disruption of cellular interactions with extracellular matrix may alter expression of CDK inhibitors such as p27^KIP1 (32, 35). Therefore, we examined the effects of inhibiting endothelial cell interactions with the HU177 cryptic epitope on expression of p27^KIP1. Endothelial cells were resuspended in the presence or absence of mAb HU177 or an isotype-matched control and seeded on plates coated with denatured collagen type IV. At the end of the incubation period, cells were collected, and mRNA and whole-cell lysates were prepared. Expression of p27^KIP1 was quantified by real-time reverse transcription-PCR and Western blot analysis. As shown in Fig. 4C, inhibition of HUVEC interaction with the HU177 cryptic epitope resulted in an ∼3-fold increase in p27^KIP1 mRNA levels compared with controls. To confirm these findings at the protein level, Western blot analysis of total cell lysates from HUVECs were examined. As shown in Fig. 4D, the relative expression of p27^KIP1 but not tubulin was increased following disruption of HUVEC interactions with the HU177 cryptic epitope.

Role of HU177 cryptic epitope in endothelial cord formation and angiogenesis. Endothelial cell cord formation in Matrigel is thought to mimic certain steps in the angiogenic cascade. We examined the effects of the mAb HU177 on endothelial cord formation in Matrigel, a basement membrane preparation containing collagen type IV. HUVECs were resuspended in proliferation buffer in the presence or absence of HU177 or an isotype-matched control and seeded onto Matrigel. Endothelial cord formation was assessed 24 h later by light microscopy. As shown in Fig. 5A, treatment of endothelial cells with mAb HU177 (250 μg/mL) resulted in disruption of the number and extent of endothelial cords compared with no treatment or incubation with the control antibody, whereas lower concentrations of mAb HU177 had little effect (data not shown). To quantify cord formation, the endothelial branch points were counted within four microscopic fields for each experimental condition. As shown in Fig. 5B, mAb HU177 significantly (P < 0.05) inhibited endothelial cord formation by ∼50% compared with controls.

Endothelial cord formation does not completely recapitulate the integrated events of angiogenesis in vivo. Therefore, we examined the effects of systemically administered mAb HU177 on basic fibroblast growth factor–induced angiogenesis (30). Angiogenesis was induced by basic fibroblast growth factor in the chorioallantoic membranes of 10-day-old embryos. Twenty-four hours after basic fibroblast growth factor stimulation, embryos were treated with a single i.v. injection of mAb HU177 or an isotype-matched control. At the end of the 3-day incubation period, the embryos were sacrificed, and the chorioallantoic membranes were analyzed. As shown in Fig. 5C, mAb HU177 (200 μg) significantly (P < 0.05) inhibited angiogenesis by >90%, whereas lower concentrations had minimal effects (data not shown). In contrast, systemic administration of the isotype-matched control antibody had little if any effect.

Role of HU177 cryptic epitope in tumor growth and angiogenesis in vivo. Given our observation that selective targeting of the HU177 cryptic site can significantly inhibit endothelial cord formation and basic fibroblast growth factor–induced angiogenesis, we sought to confirm the potential role for the HU177 cryptic epitope in tumor angiogenesis and tumor growth. Briefly, CS-1 melanoma cells were seeded on the chorioallantoic membranes of 10-day old-chick embryos. Twenty-four hours later, the embryos were treated (i.v.) with mAb HU177 or an isotype-matched control, and tumor growth and angiogenesis were assessed. At the end of the 7-day incubation period, tumors were removed, and wet weights were determined. As shown in Fig. 6A, mAb HU177 (200 μg) significantly (P < 0.05) inhibited tumor growth by ∼50% compared with no treatment or treatment with an isotype-matched control. To examine the effects of mAb HU177 on tumor-associated angiogenesis, the CS-1 melanoma tumors were washed, and angiogenesis was quantified by counting the number of branching tumor blood vessels. As shown in Fig. 6B and C, mAb HU177 significantly inhibited the number of tumor-associated blood vessels by ∼60% compared with either no treatment or treatment with an isotype-matched control. To gain further insight into the potential mechanisms by which mAb HU177 may inhibit angiogenesis, we examined frozen sections of these CS-1 tumors for blood vessel apoptosis. Briefly, tissue sections of tumors from each experimental condition were co-stained by Tunnel (ApopTag) and rhodamine-labeled tomato lectin (L. esculentum lectin) known to bind blood vessels (20). As shown in Fig. 6D, little apoptosis was detected within tumor blood vessels from untreated or control antibody-treated animals. In contrast, apoptosis was detected within tumor-associated blood vessels from animals treated with mAb HU177. Moreover, some scattered apoptosis staining was detected outside of the blood vessels within the tumor mass. Although quantification of the relative levels of apoptosis staining within the tumor suggested a trend toward increased (50%) apoptosis staining in tumors treated with mAb HU177 compared with controls, this difference failed to exhibit statistical significance (P > 0.05) due to variation in staining (data not shown). Taken together, these findings confirm the antiangiogenic activity of mAb HU177 and are consistent with the possibility that endothelial cell interactions with the HU177 cryptic epitope may regulate endothelial cell proliferation and survival.

A number of recent studies have provided insight into the molecular mechanisms by which proteolytic enzymes contribute to invasive cell behavior (36, 37). It is becoming clear that remodeling of the extracellular matrix can regulate invasive processes such as angiogenesis in both a positive and negative manner by creating local microenvironments that may be either proangiogenic or antiangiogenic (38–41). Matrix metalloproteinase–mediated remodeling of extracellular matrix can create a less restrictive microenvironment that is conducive to endothelial cell invasion and migration, thus facilitating new blood vessel growth. Moreover, proteolytic enzymes can release extracellular matrix–sequestered mitogenic growth factors such as vascular endothelial growth factor, fibroblast growth factor, and insulin-like growth factor, thereby promoting endothelial cell proliferation and motility (13–15). Proteolytic enzymes can also release a number of endogenous angiogenesis inhibitors, including endostatin, tumstatin, and other fragments of extracellular matrix proteins, that may have antiangiogenic activity (42, 43). Thus, it is clear that a wealth of angiogenesis regulatory information is present within the complex integrated network of the extracellular matrix, and a more complete understanding of how this information is used by vascular cells is of critical importance.

Previous studies from our laboratory as well as others have suggested that proteolytic remodeling of extracellular matrix molecules may result in the exposure of unique cryptic extracellular matrix epitopes that play functional roles in angiogenesis, tumor growth, and metastasis (18, 19, 21, 44–47). In fact, endothelial and tumor cell interactions with the HUIV26 cryptic epitope within collagen type IV can regulate angiogenesis, tumor growth, and metastasis in vivo, thereby implicating these non-cellular epitopes as a potential new class of therapeutic targets (18–22). Evidence has suggested that the HU177 cryptic epitope can be exposed within three-dimensional collagen gels in vitro following tumor and vascular smooth muscle cell proteolysis and within murine papillary membranes during angiogenesis (23–25). Recently, we showed that the HU177 cryptic epitope could be exposed in vivo in a temporal manner within the extracellular matrix of ischemic muscle tissue and correlated with increased levels of active matrix metalloproteinases (26). Here, we confirm the selective in vivo exposure of this unique cryptic collagen site and expand our studies to show its selective colocalization within the extracellular matrix of tumor blood vessels. Moreover, we provide the first evidence that the HU177 cryptic epitope within collagen type IV may be comprised in part by the peptide sequence LPGFPG. However, several similar sequences (LPG{X}PG), were X can be a number of different residues, are also present within the individual chains of both collagen type I and collagen type IV. Thus, it cannot be completely ruled out that additional similar sequences may also contribute to the HU177 epitope. Therefore, it is clear that more in-depth studies will be necessary to identify which of the numerous highly similar sequences are specifically exposed in vivo and which may be responsible for regulating endothelial cell behavior.

Our current studies provide the first evidence that the HU177 cryptic epitope may play a functional role in the ability of endothelial cells to interact with denatured and remodeled collagen. A mAb directed to this cryptic epitope selectively inhibited endothelial cell adhesion and migration on denatured collagen type IV but not on native collagen or other extracellular matrix proteins. Although endothelial cell interactions with the extracellular matrix may regulate proliferation, our previous studies failed to indicate a role for the HUIV26 cryptic epitope in proliferation (19, 20). In contrast, endothelial cell interactions with the distinct HU177 cryptic epitope, as opposed to the HUIV26 epitope, seems to regulated proliferation because a mAb directed to this site inhibited endothelial cell proliferation by ∼50%. Importantly, CS-1 melanoma cells failed to interact with the HU177 epitope; thus, mAb HU177 showed little, if any, direct effect of CS-1 tumor cell proliferation. The antiproliferative activity observed in endothelial cells was associated with enhanced expression of the CDK inhibitor p27^KIP1 at both the mRNA and protein levels. These finding are consistent with the known role of p27^KIP1 in regulating cell cycle control (32–35). Given the complexity of mechanisms regulating expression and functional activity of p27^KIP1, additional studies will be required to determine whether up-regulation of p27^KIP1 plays a functional role in the antiproliferative activity observed in our studies. Taken together, these data provide additional evidence that the HUIV26 and HU177 cryptic collagen epitopes are distinct and likely regulate different cellular events.

Given the roles of cell adhesion, migration, and proliferation in new vessel development, we examined the effects of HUI77 cryptic epitope in angiogenesis. mAb HU177 inhibited endothelial cell cord formation in vitro and both cytokine and tumor angiogenesis in vivo. The reduction in tumor angiogenesis was associated with a corresponding reduction in the growth of CS-1 tumors. These findings are in agreement with our previous studies with the HUIV26 cryptic site and provide further evidence for the utility of the unique approach of selectively targeting non-cellular cryptic epitopes for the treatment of pathologic angiogenesis. Interestingly, co-staining analysis of the tumors revealed apoptosis within blood vessel from animals treated with mAb HU177, whereas little, if any, apoptosis could be detected within vessels from either untreated or control treated tumors. In contrast, scattered apoptosis was detected throughout the tumor mass outside of vessels, with a trend toward elevated levels of apoptosis staining within the tumors treated with mAb HU177. However, the increased levels of apoptotic staining outside the vessels between conditions did not reach statistical significance. Although it is clear that additional studies will be required to determine whether the antiangiogenic activity of mAb HU177 is associated with the selective induction of blood vessel apoptosis, our findings are consistent with previous studies suggesting that disruption of endothelial cell-extracellular matrix interactions and the up-regulation of p27^KIP1 may induce apoptosis (32–35, 47). It is of interest to point out that previous studies have suggested that the regulation of p27^KIP1 can be altered by modulation of the phosphatidylinositol 3-kinase/AKT signaling pathway (48). Disruption of integrin-mediated cellular interactions with interstitial collagen can regulate cell cycle control, proliferation, and expression of CDK inhibitors such as p21^CIP1 and p27^KIP1 (32–35). Given these studies, it would be interesting to speculate that a unique integrin-mediated interaction with the HU177 cryptic collagen epitope may act to selectively suppress expression p27^KIP1, facilitate endothelial cell adhesion and motility, and promote endothelial cell proliferation and survival during angiogenesis. Moreover, given that CS-1 melanoma tumor cells lacked the ability to interact with the HU177 cryptic epitope, it was not surprising that mAb HU177 seemed to have little direct effect on these CS-1 tumor cells.

Taken together, our studies provide evidence of a second functional matrix-immobilized cryptic extracellular matrix epitope within collagen that plays a positive role in angiogenesis. Importantly, our studies also suggest a highly selective and novel approach for specifically targeting non-cellular epitopes for the treatment of pathologic angiogenesis and malignant tumor growth.