HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Petitclerc, E. Articles by Brooks, P. C. Search for Related Content PubMed PubMed Citation Articles by Petitclerc, E. Articles by Brooks, P. C. Social Bookmarking                      What's this?

J Biol Chem, Vol. 275, Issue 11, 8051-8061, March 17, 2000

New Functions for Non-collagenous Domains of Human Collagen Type IV
NOVEL INTEGRIN LIGANDS INHIBITING ANGIOGENESIS AND TUMOR GROWTH IN VIVO^*

Eric Petitclerc^§, Ariel Boutaud^¶, Archie Prestayko^¶, Jingsong Xu, Yoshikazu Sado, Yoshifumi Ninomiya^**, Michael P. Sarras Jr., Billy G. Hudson, and Peter C. Brooks^§§

From the  Department of Biochemistry and Molecular Biology, University of Southern California School of Medicine, Los Angeles, California 90033, ^¶ BioStratum Inc., Durham, North Carolina 27713,  Division of Immunology, Shigei Medical Research Institute, 701 Okayama, Japan, ^** Department of Molecular Biology and Biochemistry, Okayama University Medical School, Okayama, Japan, and the  Department of Biochemistry and Molecular Biology, the University of Kansas Medical Center, Kansas City, Kansas 66160

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Collagen type IV is a major component of the basal lamina of blood vessels. Six genetically distinct collagen type IV chains have been identified and are distributed in a tissue-specific manner. Here we define a novel function for soluble non-collagenous (NC1) domains of the 2(IV), 3(IV), and 6(IV) chains of human collagen type IV in the regulation of angiogenesis and tumor growth. These NC1 domains were shown to regulate endothelial cell adhesion and migration by distinct _v and ₁ integrin-dependent mechanisms. Systemic administration of recombinant 2(IV), 3(IV), and 6(IV) NC1 domains potently inhibit angiogenesis and tumor growth, whereas 1(IV), 4(IV), and 5(IV) showed little if any effect. These findings suggest that specific NC1 domains of collagen type IV may represent an important new class of angiogenesis inhibitors.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The extracellular matrix (ECM)¹ is a complex network of fibrous proteins, proteoglycans, and other macromolecules that are interconnected in a mesh-like structure that provides mechanical support for cells and tissues (1, 2). Although the ECM provides structural support, recent studies provide intriguing evidence that a wealth of biochemical information is contained within its molecular structure (3, 4). A specialized form of ECM lining epithelial sheets and blood vessels is the basal lamina or basement membrane. The basement membrane is composed predominantly of collagen type IV, laminin, enactin/nidogen, and proteoglycans (5-7). This complex structure plays a pivotal role in the regulation of cellular proliferation, adhesion, migration, gene expression, and differentiation (8). Moreover, studies indicate that specific functional domains within ECM molecules such as fibronectin, laminin, and thrombospondin may regulate angiogenesis and tumor growth (9-11).

Interestingly, six genetically distinct collagen type IV chains have been identified (see Fig. 1) that have unique patterns of tissue distribution (12-15). The most widely distributed form of collagen type IV, composed of a network of two 1(IV) chains and one 2(IV) chain, is found in the basal lamina of many tissues, including blood vessels (12-15). Collagen type IV, like many matrix macromolecules, is organized into functional domains (see Fig. 1) (16). The 7 S domain at the amino terminus mediates association of collagen IV molecules into a tetramer, whereas the NC1 domain at the carboxyl terminus is thought to promote individual chain selection, lateral association, and dimerization during matrix assembly (17-19). In contrast, the central triple helical region of collagen type IV promotes cellular interactions by ligation of ₁ integrins (20, 21).

Previous studies utilizing a wide variety of species and experimental systems have established the importance of basement membranes, and type IV collagen in particular, in the process of morphogenesis. The fact that basement membrane formation is essential for morphogenesis and cell differentiation in such ancient and divergent invertebrate organisms as Hydra vulgaris, emphasizes the fundamental importance of cell/ECM interactions in biological systems (22). Experiments in which exogenously added NC1 domains of type IV collagen caused a perturbation of ECM formation and a subsequent blockage of morphogenesis in H. vulgaris also highlights the importance of this particular matrix component to growth, development, and cellular function (22). This work has been extended to vertebrates, where it has been shown that type IV collagen 3-deficient mice have auditory (23) and renal (24) deficits related to alterations in basement membrane structure and function. Similarly, deletion of the paired collagen type IV 5 and 6 genes leads to abnormalities in basement membrane structure and the development of such disease states as Alport syndrome and an inherited smooth muscle cancer termed diffuse leiomyomatosis (25).

From the standpoint of potential medical treatments, specialized forms of morphogenesis such as angiogenesis have received increased attention because of the importance of blood vessel formation in tumor growth (26-28). A wide variety of studies have determined that angiogenesis is functionally tied to cell/ECM interactions (29). More recent studies on cell/ECM interactions in angiogenesis have shown that this process can be modulated by the non-collagenous domains of type XV and XVIII collagen (30, 31). To more fully understand the molecular functions of NC protein domains, we generated recombinant NC1 domains from each of the six genetically distinct human collagen IV chains to evaluate whether these domains could regulate endothelial cell behavior.

Here we demonstrate that specific collagen IV NC1 domains provide novel _v and ₁ integrin-dependent binding sites for endothelial cells. Importantly, systemic administration of recombinant 2, 3, and 6(IV) NC1 domains potently inhibit angiogenesis and tumor growth in vivo. Taken together, these findings suggest that unique NC1 domains of collagen IV may regulate neovascularization by promoting integrin/endothelial cell interactions as well as directing matrix assembly of collagen IV within the subendothelial basement membrane. Thus, these NC1 domains may help define an important new class of anti-angiogenic molecules for the treatment of neovascular diseases.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents, Antibodies, and Chemicals-- Cortisone acetate, BSA, and PBS were obtained from Sigma. bFGF was from R&D Systems (Minneapolis, MN). Anti-fade mounting media was from Dako (Carpinteria, CA). Anti-Factor VIII-related antigen polyclonal antibody was from Biogenex (San Ramon, CA). OCT tissue-embedding medium was from Sakura Finetek (Torrance, CA). Anti-2(IV)NC1 domain mAb was kindly provided by Drs. Sado and Ninomiya (Okayama, Japan) (32). The function-blocking mAbs, LM609 (anti-_v3), P1F6 (anti-_v5), and P4C10 (anti-₁) were kindly provided by Dr. David A. Cheresh (The Scripps Research Institute, La Jolla, CA) and have been described previously (33, 34). The function-blocking mAbs 197Z3 (anti-₁), 1950Z (anti-₂), and 19522 (anti-₃) were obtained from Chemicon International (Temecula, CA). Goat anti-mouse fluorescein isothiocyanate and goat anti-rabbit rhodamine-conjugated IgGs were from BIOSOURCE International (Camarillo, CA).

Cells and Cell Culture-- Human umbilical vein endothelial cells (HUVECs), HT1080 human fibrosarcoma cells, and B16 murine melanoma cells were obtained from the ATCC (Manassas, VA). The CS1 hamster melanoma cell line was a kind gift from Dr. Caroline Damsky (University of California at San Francisco). HUVECs were cultured in M199, 20% FBS, 1 mM sodium pyruvate, 15 µg/ml endothelial cell growth supplement, heparin (50 U/ml), penicillin (100 U/ml), and streptomycin (100 µg/ml) and were used between passages 2 and 6. HT1080 cells were cultured in Dulbecco's modified Eagle's medium, 10% FBS, penicillin (100 U/ml), and streptomycin (100 µg/ml). CS1 and B16 melanoma cells were cultured in RPMI 1640 medium, 10% FBS, penicillin (100 U/ml), and streptomycin (100 µg/ml).

Expression of Recombinant NC1 Domains of Human Collagen Type IV-- Recombinant NC1 domains of human collagen type IV were expressed as fusion proteins containing a FLAG sequence from Kodak IBI (New Haven, CT) (35, 36). A modified eukaryotic expression vector derived from pRc/CMV was purchased from Invitrogen (San Diego, CA) and was used to express cDNAs in the mammalian expression system of human embryonic kidney 293 cells obtained from the ATCC (1573-CRL) (35). The expression vectors for 3(IV)NC1, 4(IV)NC1, and 5(IV)NC1 were constructed by polymerase chain reaction amplification of the genes from plasmids pDS3 MM19-14 and MD-6 (35) using the following primers: 5'GAT ATG CTA GCC GAC TAC AAG GAC GAC GAT GAC AAA CGT GGA GAC3' and 5'TAC ATA GGG CCC TCA GTG TCT TTT CTT CAT GCA C3' for 3(IV)NC1; 5'ATA ATG CTA GCC GAC TAC AAG GAC GAC GAT GAC AAG CCT GGA TAC CTC GGT GGC TTC C3' and 5'GCC GAG GGC CCC TAG CTA TAC TTC ACG CAG3' for 4(IV)NC1; 5'T GGT CCG CTA GCT GAC TAC AAG GAC GAC GAT GAC AAA GGT CCC CCT GG3' and 5'TAG AAT AGG GCC CTC TAG ATG CAT GCT CGA3' for 5(IV)NC1. The construction of the expression vectors for 1(IV)NC1, 2(IV)NC1, and 6(IV)NC1 domains was accomplished by polymerase chain reaction amplification of the genes from a cDNA library purchased from CLONTECH (Palo Alto, CA) using the following primers, respectively: 5'GCT AGC ATC TGT TGA TCA CGG CTT CC3' and 5'CCG CGG TAG CTG AGT CAG GCT TCA TTA TG3' for 1(IV)NC1; 5'GCT AGC CGT CAG CAT CGG CTA CCT CC3' and 5'GGG CCC TGG CAC GCG CCG GCT CAC AGG 3' for 2(IV)NC1; 5'GCT AGC GAG CAT GAG AGT GGG CTA CAC G3' and 5'GGG CCC GTG GCA GGT GCC ACC CTA CAG GC3' for 6(IV)NC1. The primers for 1, 2, and 6(IV) NC1 domains were designed based on the published sequences for human type IV collagen chains (35). The secreted proteins carried the FLAG sequence fused to the amino terminus of the respective full-length NC1 domains. The proteins were purified from serum-free conditioned media by affinity chromatography on anti-FLAG agarose. The recombinant protein pools were concentrated by Amicon YM10 ultrafiltrators, and the elution buffer was exchanged to PBS.

Cell Adhesion Assays-- Recombinant (IV)NC1 domains (25 µg/ml) were immobilized on 48-well non-tissue culture-treated plates as described previously (37). The coating efficiency of all six NC1 domains was essentially equal as measured by both total protein BCA and anti-FLAG enzyme-linked immunosorbent assays (data not shown). Wells were washed and incubated with 1% BSA in PBS for 1 h at 37 °C. Subconfluent HUVECs were harvested, washed, and resuspended in adhesion buffer containing RPMI 1640 medium, 1 mM MgCl₂, 0.2 mM MnCl₂, and 0.5% BSA. HUVECs (1 × 10⁵) resuspended in 200 µl of the adhesion buffer were added to each well and allowed to attach for 30 min at 37 °C. Function-blocking antibodies were added to the adhesion buffer at a final concentration of 25 µg/ml. The nonattached cells were removed by washing, and the attached cells were stained for 10 min with crystal violet as described (37). The wells were washed three times with PBS, and cell-associated crystal violet was eluted by addition of 100 µl of 10% acetic acid. Cell adhesion was quantified by measuring the optical density of eluted crystal violet at a wavelength of 600 nm with a microtiter plate reader.

Cell Migration Assays-- The bottom side of Transwell^TM migration chamber membranes (8.0-µm pore size) were coated with recombinant (IV)NC1 domains at a concentration of 25 µg/ml in PBS for 16 h at 4 °C as described previously (38, 39). Next, 600 µl of migration buffer (RPMI 1640, 1 mM MgCl₂, 0.2 mM MnCl₂, and 0.5% BSA) was added to the lower chamber. HUVECs (1 × 10⁵) were resuspended in migration buffer (100 µl), added to the upper chamber, and allowed to migrate for 8 h at 37 °C as described (38, 39). Cells remaining on the top of the membrane were removed, and cells that had migrated to the bottom side were fixed and stained with crystal violet. The Transwell^TM membranes were washed, and the cell-associated crystal violet was eluted with 10% acetic acid. Cell migration was quantified by measuring the optical density of eluted crystal violet at a wavelength of 600 nm with a microtiter plate reader as described previously (38, 39).

Cell Proliferation Assays-- Cells (HUVEC and CS1 melanoma) were plated in 6-well culture plates (25,000 cells/well) in the presence or absence of recombinant NC1 domains at concentration of 1, 10, and 30 µg/ml. HUVECs were incubated in M199, 20% FBS, 1 mM sodium pyruvate, 15 µg/ml endothelial cell growth supplement, heparin (50 U/ml), penicillin (100 U/ml), and streptomycin (100 µg/ml), and CS1 cells were incubated in RPMI 1640 medium, 10% FBS, penicillin (100 U/ml), and streptomycin (100 µg/ml). Cell counts were determined every 24 h up to 7 days with the use of a hemacytometer.

Chick Chorioallantoic Membrane Angiogenesis Assay-- Angiogenesis was induced on chick chorioallantoic membranes (CAMs) of 10-day-old chick embryos by placing a filter disc containing either bFGF (1 µg/ml) or VEGF (1 µg/ml) (approximately 20 µl/disc) on the CAMs in an area that was reduced in the number of pre-existing blood vessels as described previously (40). The embryos were treated with a single intravenous injection of recombinant (IV)NC1 domain (1.0-30.0 µg) in a total volume of 100 µl of sterile PBS. The embryos were allowed to incubate for a total of 3 days at which time the filter discs and surrounding CAM tissue were excised, washed, and analyzed. Angiogenesis was quantified by determining the number of branching blood vessels at a fixed focal length within the confined region of the filter disc by observers unaware of the experimental conditions. Angiogenesis was expressed as an angiogenic index, which was calculated as the mean number of blood vessel branches from experimental treated CAMs minus that of the control CAMs in the absence of cytokine (40). Angiogenesis experiments were completed three to five times with 5 to 15 embryos per condition.

Chick Embryo Tumor Growth Assays-- Single cell suspensions of CS1 melanoma (5 × 10⁶ per embryo), HT1080 fibrosarcoma (4 × 10⁵ per embryo), or B16 melanoma cells (5 × 10⁶ per embryo) were applied in a total volume of 40 µl of RPMI 1640 to the CAMs of 10-day-old chick embryos (41). Twenty-four hours later, the embryos received a single intravenous injection of recombinant NC1 domains (30.0 µg) in a total volume of 100 µl. Tumors were allowed to grow for a total of 7 days, the resulting tumors were removed and trimmed free of surrounding CAM tissue, and wet weights were determined. Tumor experiments were performed three to four times with 5 to 15 embryos per condition.

Immunofluorescence Analysis of Human Tumor Biopsies-- Biopsies of human malignant melanoma and bladder carcinoma were embedded in OCT and snap frozen as described previously (42). Tumor sections (4.0 µm) were incubated with 1.0% BSA in PBS for 2 h at 37 °C to block nonspecific binding sites. Following washes, the tissue sections were co-stained with a polyclonal antibody directed to Factor VIII-related antigen (1:100) and a mAb directed to the 2(IV)NC1 domain (1:500) in 1.0% PBS-BSA. The slides were allowed to incubate for 2 h at room temperature. The tissues were washed five times with PBS for 5 min each followed by incubation with rhodamine-conjugated goat anti-rabbit and fluorescein isothiocyanate-conjugated goat anti-mouse IgGs for 1 h at 37 °C. Tissue sections were washed as above and mounted with fluoromount anti-fade reagent and sealed. Photographs were taken with an Olympus microscope fitted with epifluorescence at 200× magnification as described previously (42).

Quantification of Tumor-associated Angiogenesis-- Tumor-associated angiogenesis was quantified by counting the number of branching blood vessels within a fixed focal length that were associated with the surface of the tumors as described previously (43). Vessel counts were made with the use of a stereo microscope and were completed for both the upper and lower surfaces of each tumor within a fixed focal length. Blood vessel counts were expressed as the mean number of surface vessels per tumor.

Statistical Analysis-- Statistical analysis was performed using Student's t test for unpaired samples or the Wilcoxon signed rank test. p values less than 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

NC1 Domains of Human Collagen Type IV Differentially Support Endothelial Cell Interactions-- Six genetically distinct collagen chains have been identified that can be organized into a number of specific triple helical collagen type IV molecules (Fig. 1). The formation of supramolecular networks of collagen type IV involves unique interactions between amino-terminal 7 S domains resulting in tetramers and between NC1 domains at the carboxyl terminus forming hexameric structures (Fig. 1). Interestingly, the NC1 domain is also thought to regulate individual chain selection as well as lateral association of individual collagen molecules during matrix assembly (17-19). Recent studies, however, suggest that collagen IV NC1 domains may also promote cellular interactions (44, 45). Therefore, we evaluated whether specific NC1 domains from collagen IV could interact with human endothelial cells. To facilitate these studies recombinant forms of all six genetically distinct human (IV)NC1 domains were generated; the amino acid sequences are provided in Table I.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Schematic representation of collagen type IV chains and supramolecular networks. A, type IV collagen comprises a family of six homologous chains. Each chain has a 7 S domain at the amino terminus, a long collagenous domain (approximately 1400 residues), and a non-collagenous domain (NC1) of approximately 230 residues at the carboxyl terminus. B, three  chains assemble into a triple helical protomer, as exemplified by the (1)2 2 molecule. Protomers interact head-to-head, end-to-end, and by lateral associations, forming networks of distinct chain compositions. The 1/2 network is common to all basement membranes.

View this table: [in this window] [in a new window]   Table I Sequence of recombinant human collagen type IV NC1 domains Amino acid sequences for the six NC1 domain recombinant fusion proteins of human collagen IV. The sequences underlined represent the collagen-IV NC1 domain. The remaining sequences represent the FLAG Tag sequence for each fusion protein. The arrow indicates the RGD site, a known amino acid binding sequence for AB integrin.

To assess the ability of these domains to interact with endothelial cells, they were immobilized on microtiter wells and human umbilical vein endothelial cells were allowed to attach for 30 min. As shown in Fig. 2A, 3 and 6 NC1 domains supported endothelial cell adhesion. Interestingly, 2 also promoted cell adhesion but was less potent. In contrast, 1, 4, and 5 showed little if any adhesion promoting ability. To further examine endothelial cell interactions with (IV)NC1 domains, the capacity of these collagen domains to support endothelial cell migration was evaluated by the well established method using Transwell^TM migration chambers (37-39). As shown in Fig. 2B, human endothelial cells migrated on NC1 2, 3, and 6, whereas little if any migration was seen on 1, 4, or 5. Taken together, these findings demonstrate that specific NC1 domains can directly interact with human endothelial cells yet differ significantly in their capacity to promote both adhesion and migration.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Collagen type IV NC1 domains differentially support endothelial cell interactions. Cell adhesion and migration were performed using 48-well plates or TranswellTM migration chambers coated with a concentration (25 µg/ml) of each of the six genetically distinct recombinant (IV)NC1 domains. Coating efficiencies of all six NC1 domains were essentially equal as determined by total protein and anti-FLAG enzyme-linked immunosorbent assays. A, HUVEC cell attachment to immobilized recombinant human (IV)NC1 domains. B, HUVEC cell migration on immobilized recombinant human (IV)NC1. The experiment was repeated three times with similar results. Data bars, mean optical density ± S.D. of triplicates wells.

NC1 Domains of Collagen Type IV Are Novel Ligands for Distinct _v and ₁ Integrin Receptors-- The integrin family of cell adhesion receptors has been shown to mediate cellular interactions with the ECM (46). However, little is known concerning the roles of integrins in the cellular interactions with NC1 domains of collagen. Thus, we examined the possibility that specific members of this family may directly bind the (IV)NC1 domains. To test this possibility, recombinant (IV)NC1 domains capable of supporting cellular adhesion and migration were immobilized on microtiter wells and human endothelial cells were allowed to interact in the presence or absence of specific function-blocking anti-integrin antibodies. As shown in Fig. 3A, human endothelial cells attached to immobilized 2(IV)NC1. Surprisingly, this cell attachment depended in large part on integrin _v3, because anti-_v3 mAb (LM609) inhibited cell adhesion by over 90%. Interestingly, 2 NC1 does not contain the classical RGD tripeptide sequence recognized by _v3 (Table I), suggesting the existence of a novel non-RGD _v3 binding site. Endothelial cell interactions with 2(IV)NC1 were also partially inhibited by the function-blocking mAb (P1F6) directed to the _v5 heterodimer (34) and a mAb (P4C10) that blocks the function of ₁-containing integrins (34), suggesting the involvement of _v5 and other ₁ integrin receptors.

View larger version (56K): [in this window] [in a new window]   Fig. 3.   NC1 domains of collagen type IV promote endothelial cell adhesion by distinct integrin receptors. HUVEC cell adhesion was performed using 48-well plates coated with recombinant (IV)NC1 domains at 25 µg/ml. HUVECs were allowed to attached to the coated wells for 30 min in the presence or absence of function-blocking mAbs directed to distinct integrins receptors. A and B, HUVEC adhesion to 2(IV)NC1 domains. C, HUVEC adhesion to 3(IV)NC1 domain. D, HUVEC adhesion to 6(IV)NC1 domain. Data bars, mean optical density ± S.D. from triplicate wells. NT, no treatment; Anti-v3, mAb LM609 (25 µg/ml); Anti-v5, mAb P1F6 (25 µg/ml); Anti-1, P4C10 (25 µg/ml); Anti-1, mAb 1973Z (25 µg/ml); Anti-2, mAb 1950Z (25 µg/ml); Anti-3, mAb 19522 (25 µg/ml). The experiment was done three times with similar results.

To determine which ₁ integrins were involved in cellular interactions with 2(IV)NC1, we examined endothelial cell adhesion in the presence or absence of function-blocking antibodies directed to specific ₁ integrins. As shown in Fig. 3B, mAb directed to 3 inhibited cellular interactions with 2(IV)NC1 by approximately 50%, whereas mAbs directed to ₁ and ₂ integrins showed little effect. These findings suggest a role for ₃₁ integrin in this interaction. To determine whether integrin receptors were also involved in mediating endothelial cell interactions with 3(IV)NC1 and 6(IV)NC1, similar experiments were performed. As shown in Fig. 3 (C and D), endothelial cell adhesion to both 3 and 6(IV)NC1 domains were mediated by integrin _v3, because endothelial cell adhesion was blocked by over 90% with anti-_v3 mAb LM609, whereas little if any effect was seen with function-blocking mAbs directed to _v5 or ₁ integrins. Although recombinant 3(IV)NC1 domain contains the RGD tripeptide recognition motif, 2 and 6 domains do not (Table I). Taken together, these results suggest that 2 and 6(IV)NC1 domains contain novel non-RGD-dependent _v3 binding sites. These surprising results indicate that, although the central triple helical region of collagen IV can support cellular interactions, distinct (IV)NC1 domains can promote novel integrin-mediated cellular interactions by engagement of both _v and ₁ integrins. Moreover, because ligation of both _v3 as well as ₁ integrins plays important roles in the regulation of endothelial cell behavior (29, 43, 47), cellular interactions with these specific (IV)NC1 domains may contribute to the regulation of angiogenesis in vivo.

Specific NC1 Domains of Collagen Type IV Potently Inhibit Angiogenesis-- Collagen type IV is a major component of the subendothelial basement membrane. Moreover, cellular interactions with distinct forms of collagen may regulate endothelial cell behavior. In fact, the NC domain of collagen XVIII (Endostatin) blocks angiogenesis in vivo (31). Therefore, it is possible that specific NC1 domains within collagen type IV may also regulate angiogenesis. To test this possibility, we examined the effects of systemically administered NC1 domains from collagen type IV on cytokine-induced angiogenesis. To facilitate these studies angiogenesis was induced within the CAMs of 10-day-old chick embryos with bFGF. Twenty-four hours later, the embryos received a single intravascular injection of each of the six genetically distinct recombinant (IV)NC1 domains. As shown in Fig. 4A, 2, 3, and 6(IV)NC1 domains potently inhibited bFGF-induced angiogenesis, whereas similar injections of 1, 4, and 5(IV)NC1 had little if any effect. In fact, 2, 3, and 6 inhibited angiogenesis by 80-90% as compared with controls (Fig. 4B). The 2(IV) chain represents one of the most common collagen IV chains present within the subendothelial basement membrane of blood vessels (15, 16). Moreover, 2(IV)NC1 domain was shown to support endothelial cell interactions with multiple integrins. Therefore, we choose to further evaluate the effects of 2(IV)NC1 domain on angiogenesis. As shown in Fig. 4C, systemic administration of the 2(IV)NC1 domain resulted in a dose-dependent and saturable inhibition of angiogenesis within the chick CAM model, with a half-maximal inhibition of approximately 3.0 µg/embryo. In contrast, 5(IV)NC1 had little effect.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Recombinant human (IV)NC1 domains differentially inhibit bFGF-induced angiogenesis in vivo. Angiogenesis was examined in vivo using the chick embryo model. Embryos were treated with a single intravenous injection of recombinant (IV)NC1 domains at 30.0 µg/embryo with the exception that 3(IV)NC1 was used at 1.0 µg due to embryo death at higher concentrations. A, representative examples of the experimentally defined regions of the CAMs following treatment with (IV)NC1 domains. B, quantification of angiogenesis following treatment with recombinant human (IV)NC1 domains. C, dose-dependent inhibition of angiogenesis following systemic administration of 2(IV)NC1 and 5(IV)NC1 domains. Each experiment was done three to four times with similar results. Data bars, mean ± S.E. of the angiogenic index expressed as the percentage of control of untreated embryos.

2(IV)NC1 Inhibits VEGF-induced Angiogenesis-- The 2(IV)NC1 domain was shown to interact with multiple integrins, including _v3, _v5, and ₃₁. Moreover, previous studies have suggested that specific cytokines may induce angiogenesis, which is dependent in part on distinct integrin receptors (48). In fact, previous studies indicated that angiogenesis induced by bFGF could be blocked by antagonists of the _v3 integrin, whereas _v3 antagonists had only minimal effects on angiogenesis induced by VEGF (48). In contrast, an antagonist of _v5 potently inhibited angiogenesis induced by VEGF, while having no effect on angiogenesis induced by bFGF (48). Because 2(IV)NC1 interacted with both _v3 and _v5, we assessed its effects on VEGF-induced angiogenesis in vivo. As shown in Fig. 5, systemic administration of 2(IV)NC1 inhibited VEGF-induced angiogenesis by 80%, whereas 5(IV)NC1 shows little if any effect. These findings suggest that the anti-angiogenic activity of 2(IV)NC1 is not restricted to angiogenesis induced by a single cytokine.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Recombinant human (IV)NC1 domains differentially inhibit VEGF-induced angiogenesis in vivo. Angiogenesis was examined in vivo using the chick embryo model. Embryos were treated with a single intravenous injection of recombinant (IV)NC1 domains at 30.0 µg/embryo. Quantification of angiogenesis followed treatment with either 2(IV)NC1 or 5(IV)NC1. Data bars, mean ± S.E. of the angiogenic index expressed as the percentage of control of untreated embryos.

Distinct NC1 Domains of Collagen Type IV Inhibit Tumor Growth-- The growth and metastatic dissemination of tumors depends in part on angiogenesis (26, 27). Therefore, we evaluated the effects of each of the six genetically distinct (IV)NC1 domains for their effects on solid tumor growth. A single-cell suspension of CS1 melanoma tumor cells were applied to the CAMs of 10-day-old chick embryos. Twenty-four hours later the embryos were injected systemically with either recombinant (IV)NC1 domains or controls. As shown in Fig. 6, CS1 melanoma tumor growth was inhibited by approximately 60% by a single injection of recombinant 2, 3, and 6(IV)NC1 domains. In contrast, similar injections of recombinant 1, 4, and 5 had little if any effect. These data suggest that 2, 3, and 6(IV)NC1 domains are potent inhibitors of tumor growth, because these NC1 domains were injected only once at 30.0 µg per embryo. Thus, systemic administration of distinct NC1 domains of collagen type IV not only disrupt angiogenesis but also inhibit tumor growth.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Recombinant human (IV)NC1 domains differentially inhibit tumor growth in vivo. CS1 hamster melanoma cells (5 × 106 cells/egg) were applied to the CAMs of 10-day-old chick embryos. Twenty-four hours later, embryos were treated with a single injection of recombinant human (IV)NC1 domains at a concentration of 30.0 µg/embryo. Data bars, mean ± S.E. of tumor weights, expressed as the percentage of control of untreated tumors. Data were derived from three to five independent experiments with 5 to 15 embryos per condition.

The 2(IV)NC1 Domain of Collagen Type IV Is Associated with the Subendothelial Basement Membrane of Human Tumor Blood Vessels in Vivo-- The six chains that comprise the distinct forms of collagen IV have unique patterns of tissue distributions (12-16). In fact, the most commonly distributed form of collagen IV is composed of two 1 chains and one 2 chain, whereas 3(IV) and 6(IV) chains are limited in distribution (12). Moreover, our findings indicate that 2 (IV)NC1 domain inhibited angiogenesis in a dose-dependent and saturable manner with a half-maximal inhibition at approximately 3.0 µg per embryo (Fig. 4C). Given its potency and tissue distribution within the subendothelial basement membrane, combined with the fact that 2(IV)NC1 interacts with multiple integrins, we chose to focus our attention on the 2(IV)NC1 domain for further analysis. During human angiogenesis and tumor growth, proteolytic remodeling of the subendothelial basement membrane likely occurs, and thus, particular regions of collagen IV may be disrupted or destroyed (49). Therefore, we examined whether the 2(IV)NC1 domain of collagen IV remains associated with the subendothelial basement membrane during human tumor angiogenesis in vivo. To facilitate these studies, we examined biopsies from both human melanoma and bladder carcinoma. These human tumor biopsies were co-stained with a monoclonal antibody directed to 2(IV)NC1 and a polyclonal antibody directed to Factor VIII-related antigen, a known marker of blood vessels. As shown in Fig. 7, 2(IV)NC1 domain was specifically detected within the subendothelial basement membrane of tumor blood vessels from both human melanoma and bladder carcinoma. These findings indicate that the 2(IV)NC1 domain remains associated with the subendothelial basement membrane and suggest that it is available for cellular interactions with endothelial cells during human tumor angiogenesis.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Immunolocalization of 2(IV)NC1 domain within human tumor biopsies. Analysis of human tumor biopsies from patients with either malignant melanoma or bladder carcinoma for the expression of 2(IV)NC1 domain. Frozen sections of human tumor biopsies were co-stained with a polyclonal antibody directed to Factor VIII-related antigen to mark the blood vessels and a mAb directed to 2(IV)NC1 domain. Left panels, human melanoma biopsy. Right panels, human bladder carcinoma biopsy. Red indicates human tumor-associated blood vessels. Green represents expression of 2(IV)NC1 domain. Yellow staining represents co-localization (overlap of green and red). Photomicrographs were taken at a magnification of 200×.

Recombinant 2(IV)NC1 Domain Inhibits the Growth of Tumors from Distinct Histological Origins-- Previous studies suggest that the growth and expansion of solid tumors depend on angiogenesis (26, 27). Therefore, we examined the effects of recombinant 2(IV)NC1 domain on the growth of two additional tumor types, including B16 murine melanoma and HT1080 human fibrosarcoma. As shown in Fig. 8 (A and B), a single injection of 30.0 µg of 2(IV)NC1 domain inhibited the growth of both B16 melanoma and HT1080 fibrosarcoma by approximately 40% and 50%, respectively. In contrast, similar injections of 5(IV)NC1 had little if any effect. These findings confirm the antitumor activity of the 2(IV)NC1 domain in vivo. To determine whether the antitumor activity of 2(IV)NC1 domain was associated with its anti-angiogenic activity, tumor angiogenesis was examined. Blood vessels associated with the surface of CS1 melanoma tumors from control or 2(IV)NC1-treated were evaluated. As shown in Fig. 8C, untreated and 5(IV)NC1-treated CS1 melanoma tumors were associated with numerous small branching blood vessels. In contrast, CS1 tumors treated with 2(IV)NC1 domain had significantly fewer small branching tumor blood vessels. In fact, quantification of these tumor blood vessels showed an approximately 50% reduction as compared with either control or 5(IV)NC1-treated tumors (p < 0.01). Interestingly, 2(IV)NC1 domain had little if any effects on tumor cell proliferation in vitro (data not shown). These findings suggest that the inhibition of tumor growth was due, at least in part, to inhibition of tumor angiogenesis.

View larger version (50K): [in this window] [in a new window]   Fig. 8.   Recombinant human 2(IV)NC1 Domain inhibits the growth of tumors from distinct histological origins. B16 murine melanoma and HT1080 human fibrosarcoma were applied to the CAMs of 10-day-old chick embryos. Embryos were then treated with a single intravenous injection of recombinant 2(IV)NC1 or 5(IV)NC1 domains at a concentration of 30.0 µg per embryo. A, quantification of B16 melanoma tumor growth following systemic administration of recombinant human (IV)NC1 domains. B, quantification of HT1080 fibrosarcoma tumor growth following systemic administration of recombinant human (IV)NC1 domains. Data bars, mean tumor weights ± S.E. from 5 to 10 embryos per condition. C, representative CS1 tumors treated with either recombinant 2(IV)NC1 or 5(IV)NC1 domains. Note significant reduction in the extent of tumor-associated angiogenesis (small branching tumor-associated blood vessels). D, quantification of CS1 tumor angiogenesis following systemic administration of recombinant human (IV)NC1 domains. Data bars, mean number of surface-associated vessels per tumor ± S.D. Data were derived from two experiments with 5 to 10 tumors per condition. The difference seen between the 2(IV)NC1-treated group and the controls was significant (p < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

While the ECM is thought to play a role in angiogenesis, little information is available concerning the mechanisms by which specific ECM proteins such as collagen contribute to this process. Elegant studies by Ingber and Folkman have implicated collagen metabolism in regulating angiogenesis (4, 50). Therefore, the possibility exists that specific domains within genetically distinct forms of collagen may actively regulate neovascularization. This notion is consistent with studies showing the anti-angiogenic activity of the non-collagenous domain of collagen XVIII (Endostatin) as well as the homologous NC region of collagen XV(Restin) (30, 31). However, little is known concerning the cell-surface receptors for these collagen fragments or the mechanisms by which they function. To this end, angiogenesis likely depends in part on integrin-dependent endothelial cell interactions with distinct forms of collagen (51-53). Moreover, function-blocking antibodies directed to integrins that bind collagen type IV inhibit angiogenesis in vitro and in vivo (51-53). Therefore, integrin-mediated endothelial cell interaction with collagen IV is likely an important control point in neovascularization.

Previous studies have suggested that cellular interactions with collagen IV is mediated by integrin ligation of specific regions within the central triple helical domain (21, 22, 54). In this report, we provide evidence that the genetically distinct non-collagenous domains of the 2, 3, and 6 chains of collagen IV can directly interact with endothelial cells by engaging distinct _v and ₁ integrin receptors. Surprisingly, the classical vitronectin receptor _v3 can promote endothelial cell interactions with 2, 3, and 6 (IV)NC1 domains. Although recombinant 3(IV)NC1 contains the typical RGD tripeptide recognized by _v3, both 2 and 6(IV)NC1 lack this sequence. Thus, both 2 and 6(IV)NC1 contain novel non-RGD-dependent _v3 binding sites. Importantly, endothelial cell interactions with these unique _v3 binding sequences may contribute to the regulation of angiogenesis in vivo, because ligation of _v3 has been shown to regulate neovascularization in numerous animal models (42, 43, 47). Moreover, blocking _v3 ligation has been shown to disrupt signal transduction pathways leading to altered expression of Bcl-2, Bax, and the DNA binding activity of p53 (55).

In recent years, a great deal of effort has been focused on developing novel anti-angiogenesis antagonists. The majority of these antagonists has been directed to angiogenic cytokines, cell adhesion receptors, and proteolytic enzymes (56). In contrast, relatively little attention has been focused on the possibility that specific domains of ECM molecules may represent an important new class of therapeutic targets. Here, we provide evidence for the first time that systemic administration of genetically distinct NC1 domains from the 2, 3, and 6 chains of human collagen IV, potently inhibit cytokine-induced angiogenesis in vivo. Moreover, these distinct (IV)NC1 domains also inhibited the growth of multiple tumors of distinct histological origin. The inhibition of angiogenesis and tumor growth observed in these studies was specific, because similar NC1 domains derived from the 1, 4, and 5 chains of collagen IV showed little if any effect. Moreover, 2(IV)NC1 inhibited angiogenesis in vivo in a dose-dependent and saturable manner with a half-maximal inhibition observed at approximately 3.0 µg per embryo. Importantly, the anti-angiogenic activity of 2(IV)NC1 was not restricted to angiogenesis induced by bFGF because VEGF-induced neovascularization was also potently inhibited.

Although the antitumor activity of these (IV)NC1 domains is likely due in part to their anti-angiogenic activity, it is possible that they may also have a direct effect on the tumor cells by disrupting tumor cell-NC1 domain interactions. However, the recombinant NC1 domains utilized in our studies failed to inhibit either endothelial or tumor cell proliferation in vitro at the concentration tested, suggesting alternative mechanisms of action. Interestingly, soluble NC1 domain has been shown to block collagen IV matrix assembly by binding along the length of the central helical rod-like region of collagen IV thereby disrupting lateral associations of individual collagen IV molecules (57). In addition, these soluble NC1 domains may also disrupt the carboxyl-terminal association of triple helical collagen IV molecules that occur during assembly of collagen networks. Thus, it is possible that the anti-angiogenic activity of these NC1 domains is due in part to disruption of collagen IV matrix assembly. Importantly, a ₁ integrin binding site is located approximately 100 nm from the amino terminus near a region in collagen IV where soluble NC1 domain has been suggested to bind (58). Interestingly, a function-blocking mAb directed to ₁ integrins was found previously to disrupt collagen IV deposition and assembly in an in vitro culture system (59, 60). Thus, it is possible that the anti-angiogenic activity of recombinant (IV)NC1 may be due in part to disruption of endothelial cell integrin interactions with this site.

Although the exact mechanisms by which these (IV)NC1 domains inhibit angiogenesis and tumor growth in vivo are not completely understood, it is possible that the inhibitory activity is associated with the disruption of _v and/or ₁ integrin-dependent mechanisms. Recent studies by O'Reilly et al. (31) and Ramchandran et al. (30) demonstrate anti-angiogenic activity of NC domains of collagen XVIII (Endostatin) as well as a highly homologous region within the NC domain of collagen XV (Restin). However, little is known concerning cellular receptors for these collagen fragments or the mechanisms by which they function. Moreover, these NC1 domains are not highly homologous to the NC1 domains of collagen IV. However, it would be tempting to speculate that these NC domains of collagen types XV and XVIII may inhibit angiogenesis by disruption of similar integrin-dependent mechanisms given the capacity of integrins to interact with genetically distinct forms of collagen. Our findings, in conjunction with previous reports, strongly suggest that specific regulatory domains are present within the NC regions of genetically distinct collagen molecules and that these domains may represent an important and powerful new class of anti-angiogenic compounds for the treatment of human neovascular diseases.

	ACKNOWLEDGEMENTS

We thank Dorothy Rodriguez, Jenny Kim, and Parvin Todd for their excellent technical assistance.

	FOOTNOTES

^* This work was supported in part by a grant from BioStratum Inc. and by Grants R29 CA74132-01 (to P.C.B.) and DK18381 (to B.G.H.) from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Supported by fellowships from the Medical Research Council of Canada and the Fonds de la Recherche en Santé du Québec.

^§§ To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, University of Southern California School of Medicine, Norris Cancer Center, Topping Tower Rm. 6409, 1441 Eastlake Ave., Los Angeles, CA 90033. Tel.: 323-865-0510; Fax: 323-865-0514; E-mail: pbrooks@hsc.usc.edu.

	ABBREVIATIONS

The abbreviations used are: ECM, extracellular matrix; NC, non-collagenous domain; CAM, chorioallantoic membrane; bFGF, basic fibroblast growth factor; VEGF, vascular endothelial growth factor; HUVEC, human umbilical vein endothelial cells; BSA, bovine serum albumin; PBS, phosphate-buffered saline; mAb, monoclonal antibody; FBS, fetal bovine serum.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES