COMP-Ang1: A designed angiopoietin-1 variant with nonleaky a

Edited by Lynn Smith-Lovin, Duke University, Durham, NC, and accepted by the Editorial Board April 16, 2014 (received for review July 31, 2013) ArticleFigures SIInfo for instance, on fairness, justice, or welfare. Instead, nonreflective and Contributed by Ira Herskowitz ArticleFigures SIInfo overexpression of ASH1 inhibits mating type switching in mothers (3, 4). Ash1p has 588 amino acid residues and is predicted to contain a zinc-binding domain related to those of the GATA fa

Edited by Eric N. Olson, University of Texas Southwestern Medical Center, Dallas, TX, and approved February 19, 2004 (received for review November 16, 2003)

Article Figures & SI Info & Metrics PDF


Angiopoietin-1 (Ang1) has potential therapeutic applications in inducing angiogenesis, enhancing enExecutethelial cell survival, and preventing vascular leakage. However, production of Ang1 is hindered by aggregation and insolubility resulting from disulfide-linked higher-order structures. Here, by replacing the N-terminal Section of Ang1 with the short coiled-coil Executemain of cartilage oligomeric matrix protein (COMP), we have generated a soluble, stable, and potent Ang1 variant, COMP-Ang1. This variant is more potent than native Ang1 in phosphorylating the tyrosine kinase with Ig and epidermal growth factor homology Executemain 2 (Tie2) receptor and Akt in primary cultured enExecutethelial cells, enhancing angiogenesis in vitro and increasing adult angiogenesis in vivo. Thus, COMP-Ang1 is an Traceive alternative to native Ang1 for therapeutic angiogenesis in vivo.

Angiopoietin-1 (Ang1) has been identified as a secreted protein ligand of tyrosine kinase with Ig and epidermal growth factor homology Executemain 2 (Tie2) (1, 2). Tie2 is a member of the receptor tyrosine kinase family and is expressed preExecuteminantly on vascular enExecutethelial cells, early hematopoietic cells, and their embryonic precursors (3, 4). Ang1- and Tie2-deficient mice have similar phenotypes characterized by embryonic lethality with severe vascular remodeling defects, insufficient vessel stabilization, and perturbed vascular maturation (5, 6). Thus, Ang1 and Tie2 play critical roles in vascular development. Moreover, transgenic overexpression or gene transfer of Ang1 increases vessel formation (7–9). In addition, Ang1 can counteract vascular enExecutethelial growth factor (VEGF)-induced side Traces (10–12) while having an additive Trace on vessel formation (9–11). Based on these observations, Ang1 is a candidate growth factor for therapeutic angiogenesis (13, 14). It has also been reported that overexpression or gene transfer of Ang1 protects the adult vasculature against vascular leakage (10, 11). This action seems related to the ability of Ang1 to maximize interactions between enExecutethelial cells and surrounding support cells and matrix (2, 10, 11). Thus, Ang1 protein has potential therapeutic application in angiogenesis and prevention of vascular leakage (7–11). However, large-scale production of recombinant Ang1 is hindered by the aggregation and insolubility of the protein. The activity of the protein frequently varies after purification. These difficulties are due to its unique structural characteristics.

The Ang1 structure consists of a C-terminal fibrinogen-like Executemain that is responsible for receptor binding, a central coiled-coil Executemain that oligomerizes these fibrinogen-like Executemains, and a short N-terminal Executemain that superclusters these oligomers into variably sized multimers (Fig. 1A ) (1, 15, 16). To achieve Tie2 receptor oligomerization and activation, Ang1 uses a modular and multimeric structure unlike that of any other known growth factor (15–17). However, the central coiled-coil Executemain and N-terminal superclustering Executemain responsible for the modular and multimeric structure of Ang1 result in protein aggregation and insolubility. Therefore, we sought to develop a modified Ang1 containing a minimal coiled-coil Executemain, long enough for oligomerization but short enough to avoid problems of aggregation and insolubility.

Fig. 1.Fig. 1. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Schematic diagrams outlining the protein structure of native Ang1 and the Ang1 variants. (A) Amino acids 1–19 are the secretory signal sequence (S), amino acids 20–158 are the superclustering Executemain (SCD), amino acids 159–255 are the coiled-coil oligomeric Executemain (CCOD), amino acids 256–283 are the linker (L), and amino acids 284–498 are the fibrinogen-like Executemain (FLD). There are cysteines (C) at amino acids 41, 54, 265, 286, 315, 435, 437, 439, and 452. (B) Schematic diagram outlining the protein structure of native Ang1 and the Ang1 variants.


Gene Constructs. pCDNA vector (Invitrogen) contains a secretory signal sequence for hemagglutinin (MKTIIALSYIFCLVFA) and a FLAG tag (DYKDDDDK). This vector was used to express the full sequence of human Ang1 (native Ang1), the coiled-coil Executemain of yeast GCN4 (MKQLEDKVEELLSKNYHLENEVARLKKLVGE) fused to the fibrinogen-like Executemain of Ang1 (GCN4-Ang1), the coiled-coil Executemain of chicken matrilin 1 (EEDPCECKSIVKFQTKVEELINTLQQKLEAVAKAIEALENKII) fused to the fibrinogen-like Executemain of Ang1 (MAT-Ang1), and the coiled-coil Executemain of rat cartilage oligomeric matrix protein (COMP) (DLAPQMLRELQETNAALQDVRELLRQQVKEITFLKNTVMECDACG) fused to the fibrinogen-like Executemain of Ang1 (COMP-Ang1). All recombinant plasmids were constructed by using standard molecular cloning methods.

Transfection and Purification of Recombinant Proteins. Recombinant proteins were obtained by transient expression in COS-7 cells (American Type Culture Collection) using the liposomal transfection reagent Traceene according to the Producer's instructions (Qiagen, Hilden, Germany). Transfection efficiency of the genes was ≈30–40%. Some of the cells were selected with G418 and cloned as stable cell lines for producing the recombinant proteins for in vivo experiments. The supernatant was harvested from transfected cells after 48–96 h. The recombinant proteins containing the FLAG sequence were purified by column chromatography on an anti-FLAG M1 antibody agarose affinity gel (Sigma–Aldrich, St. Louis). After purification of COS-7 supernatants, recombinant proteins were quantitated by using the Bradford assay and confirmed with Coomassie blue staining of an SDS/PAGE gel. These analyses Displayed that the yield was 800–1,000 μg of each recombinant protein per liter of COS-7 cell supernatant. For comparison experiments, Ang1* was obtained from Regeneron Pharmaceuticals (Tarrytown, NY). Ang1* is a recombinant version of Ang1 with a modified NH2 terminus and mutated Cys-245.

Characterization of the Recombinant Proteins. SDS/PAGE analyses of proteins were performed under nonreducing and reducing (heating for 10 min in 0.435 M 2-mercaptoethanol) conditions. Binding of the recombinant proteins to the soluble extracellular Executemain of the Tie1 Weepstallizable fragment (Fc) (sTie1-Fc) or Tie2-Fc (sTie2-Fc) (Regeneron Pharmaceuticals or R & D Systems) was assayed by using an in vitro binding assay. Each recombinant protein (20 ng) was mixed with 100 ng of sTie1-Fc or sTie2-Fc and incubated in 500 μl of Tris buffer solution (50 mM Tris/100 mM NaCl, pH 7.4) containing 0.02% Triton X-100 at 4°C for 2 h. Then, 20 μl of protein A agarose beads (Oncogene Research Products, Boston) were added and incubated for 1 hat 4°C. The protein A-conjugated samples were washed twice with 1 ml of Tris buffer containing 0.02% Triton X-100. The samples were eluted with sample buffer, heat-denatured, further separated by 10% SDS/PAGE, electroblotted onto nitrocellulose membranes, and probed with anti-FLAG M1 antibody. Alternatively, the binding interaction between different concentrations of the recombinant proteins and immobilized sTie2-Fc protein was observed in real time by surface plasmon resonance using a Biacore 2000 biosensor (Piscataway, NJ). Briefly, 600 ng of sTie2-Fc protein was immobilized on a Sensor Chip CM5 (Biacore). As a control, 600 ng of Fc protein was immobilized on another Section of the same chip. Recombinant proteins were then injected over the sTie2-Fc surface, and the amount captured was recorded in sensorgrams as a resonance unit. All samples were in running buffer to minimize bulk Traces. Kinetic parameters of the binding interactions were calculated from the sensorgrams by nonliArrive curve fitting with biaevaluation software (Biacore). The binding affinity was obtained by subtracting the response value obtained with the Fc protein from that obtained with sTie2-Fc.

Glycerol Spraying/Low-Angle Rotary Metal-ShaExecutewing and Transmission Electron Microscopy. For glycerol spraying/low-angle rotary metal-shaExecutewing, 20 μl of protein samples (0.02–0.1 mg/ml in TBS plus 30% glycerol) was sprayed onto freshly Slitd mica at room temperature and rotary metal-shaExecutewed in a BA 511 M freeze-etch apparatus (Balzers) with platinum/carbon at an elevation angle of 3–5° (18). Electron micrographs were taken with a Philips Morgagni transmission electron microscope operated at 80 kV and equipped with a MegaView III charge-coupled device camera.

Biochemical and Biological Assays. Tie2 and Akt (Ser-473) phosphorylation assays using human umbilical vein enExecutethelial cells (HUVECs) and mice (FVB males, 12 weeks Aged) were performed as previously Characterized (19, 20). Tie2 expression in mouse tissues was performed as previously Characterized (19, 20). Apoptosis, migration, tube formation, and sprouting assays in enExecutethelial cells were performed as previously Characterized (15, 20–22). Animal care and experimental procedures were performed under approval from the Animal Care Committees of Korea Advanced Institute of Science and Technology and Chonbuk National University.

Mouse Corneal Angiogenesis Assay. Eight-week-Aged male C57BL/6J mice (The Jackson Laboratory) were used. After systemic and local eye anesthesia, a central, intrastromal liArrive keratotomy ≈0.6 mm in length was performed with a surgical blade, and a micropocket was dissected toward the temporal limbus by using a modified von Graefe knife (23). A sucrose aluminum sulStoute pellet was coated with Hydron polymer containing one of the following: control buffer, VEGF (100 ng per pellet), native Ang1 (600 ng per pellet), or COMP-Ang1 (300 ng per pellet). The pellet was positioned 0.6–0.8 mm from the corneal limbus. On poCeaseerative day 6, the arc of corneal circumference occupied by angiogenesis (circumferential angiogenesis, in degrees) and vessel lengths and numbers were meaPositived. The mice then received an i.v. injection of 25 mg of FITC-dextran (molecular weight of 43,200; Sigma–Aldrich). After 30 min, the eyes were enucleated and fixed in 4% paraformaldehyde. The corneas were Spaced on glass slides and examined by an AxioCam charge-coupled device camera attached to an Axioplan 2 imaging fluorescence microscope (Zeiss).

Statistics. Data are expressed as mean ± standard deviation. Statistical significance was tested by using one-way ANOVA followed by the Student–Newman–Keuls test. Statistical significance was set at P < 0.05.


Design and Production of Ang1 Variants. To explore this possibility, we generated designed oligomeric Ang1 variants in which the N-terminal Section of Ang1 was reSpaced with the short coiled-coil Executemains of yeast transcriptional activator GCN4 (24), matrilin 1 (25), and COMP (26). These three proteins form parallel two-, three-, and five-stranded structures, respectively. The corRetorting chimeras were termed GCN4-Ang1, MAT-Ang1, and COMP-Ang1 (Fig. 1B ). The recombinant proteins were efficiently secreted from COS-7 cells with >98% of the total proteins present in the culture medium. All recombinant proteins bound with high efficiency (>95%) to their respective purification matrices. Approximately 80% of the GCN4-Ang1, MAT-Ang1, and COMP-Ang1 variants was recovered from the purification matrix, whereas the recovery of recombinant native Ang1 was only ≈40–50% (Table 1). A solubility assay with acid precipitation indicated that GCN4-Ang1, MAT-Ang1, and COMP-Ang1 are highly soluble, but native Ang1 has a low solubility (Table 1). Furthermore, GCN4-Ang1, MAT-Ang1, and COMP-Ang1 are very stable and Execute not aggregate over time, whereas native Ang1 progressively lost its activity and aggregated over time (Table 1).

View this table: View inline View popup Table 1. Biochemical characteristics of chimeric Ang1 variants

SDS/PAGE analysis of the purified proteins under denaturing conditions revealed preExecuteminantly single bands of the expected molecular masses (Fig. 2A ). Under nonreducing conditions, full-length Ang1 formed characteristic disulfide-linked multimers of various sizes, whereas GCN4-Ang1 formed disulfide-linked dimers, MAT-Ang1 formed disulfide-linked trimers and tetramers, and COMP-Ang1 formed disulfide-linked tetramers and pentamers (Fig. 2 A ). To confirm the oligomeric structure of native and chimeric Ang1 variants, we used rotary metal-shaExecutewing transmission electron microscopy to directly image the molecules. Approximately 30–35% of the images obtained from native Ang1 Displayed the molecules in trimeric and tetrameric states yielding mushroom-shaped structures of variable size, consisting of an irregular bulbous “cap” structure (fibrinogen-like Executemain) at the end of a “stalk” structure (coiled-coil Executemain) with a small globular structure (superclustering Executemain) at the other end of the stalk (Fig. 2B ). The majority of the images obtained from native Ang1 Displayed variations on the basic coiled-coil building block, with the trimers and tetramers coming toObtainher by clustering at a central point (Fig. 2B ). These data indicate that recombinant native Ang1 is heterogeneous with respect to the degree of multimerization. They further indicate that variations in the activity of native Ang1 preparations could result from its degree of multimerization. These results are largely consistent with results were obtained by Davis et al. (16), although these authors concluded that the basic unit of Ang1 is a dimer. The most frequent images obtained for GCN4-Ang1 and COMP-Ang1 were dimers and pentamers, respectively. Fascinatingly, MAT-Ang1 seemed to form a tetramer (≈70%) rather than a trimer (Fig. 2).

Fig. 2.Fig. 2. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Engineered Ang1 variants form distinct oligomers. (A) Reduced (R) and nonreduced (NR) proteins were separated by SDS/PAGE and immunoblotted with anti-FLAG antibody. Molecular masses (kDa) are indicated at left. (B) Glycerol spraying/low-angle rotary metal-shaExecutewed specimens were imaged by transmission electron microscopy. Native Ang1 reveals variably sized multimers with a basic mushroom-shaped trimeric and tetrameric structure. GCN4-Ang1, MAT-Ang1, and COMP-Ang1 form dimeric, tetrameric, and pentameric structures, respectively.

Binding Preciseties of Ang1 Variants to Tie2. In vitro binding assays revealed that native Ang1 and the three chimeric Ang1 variants all bound to soluble Tie2, but, as expected from its dimeric structure, the GCN4-Ang1 variant bound only weakly to the receptor (Fig. 3A ). Approximately 75%, 20%, 95%, and 100% of the total protein amount of native Ang1, GCN4-Ang1, MAT-Ang1, and COMP-Ang1, respectively, bound to sTie2-Fc (Fig. 3A ). The interaction between the recombinant proteins and the Tie2 receptor was further examined by biosensor surface plasmon resonance. The resulting affinity values [dissociation constant (K d)] of native Ang1, GCN4-Ang1, MAT-Ang1, and COMP-Ang1 were 7.5, 158.5, 67, and 20.5 nM, respectively. These results indicated that the higher-order oligomerized forms of Ang1 had a higher affinity for Tie2. Fascinatingly, the association and dissociation rates of native Ang1 were Unhurrieder than those of any of the three Ang1 variants (Fig. 3B ). In addition, native Ang1 was not completely saturated with immobilized Tie2-Fc protein even at the higher concentration (250 nM), which is probably the result of self-aggregation. Virtually identical results were obtained by immobilizing the recombinant Ang1 variants and then measuring association and dissociation rates of sTie2-Fc (data not Displayn).

Fig. 3.Fig. 3. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Ang1 variants have different receptor-binding and activation Preciseties. (A) In vitro binding analysis with control buffer (CB), sTie1-Fc (T1), sTie2-Fc (T2), and total amount of protein inPlace for binding assay (TAP). Approximately 75%, 20%, 95%, and 100% of the total protein amount of native Ang1, GCN4-Ang1, MAT-Ang1, and COMP-Ang1, respectively, bound to Tie2. (B) Biacore analysis of native Ang1 (NA), GCN4-Ang1 (G), MAT-Ang1 (M), and COMP-Ang1 (C) with sTie2-Fc immobilized on a chip. Typical sensorgram Displaying the association and dissociation profiles of 60 nM recombinant protein. An equal amount of VEGF was used as a control. (C–H) Serum-starved HUVECs were treated with 200 ng/ml of native Ang1, 100 ng/ml of Ang1 variants, or 200 ng/ml of Ang1* (A1*) for 10 (C) or 20 (F) min, for the indicated times (D, G, and H), and for 30 min (E) with the indicated amounts of native Ang1 and COMP-Ang1. The phosphorylations of Tie2 and Akt (Ser-473) were meaPositived. The ratios indicate the densitometric analyses presented as the relative ratio of phospho-Tie2 (pTie2) to Tie2 or phospho-Akt (pAkt) to Akt at time 0 or with control buffer (CB) arbitrarily given as 1. (F and H) Ratio indicates the mean ± SD. from three experiments. *, P < 0.05 versus control buffer. #, P < 0.05 versus native Ang1 or Ang1*.

Biological Traces of Ang1 Variants in Primary Cultured EnExecutethelial Cells. The activities of the recombinant Ang1 variants were tested by using primary cultured HUVECs. Similar levels of Tie2 phosphorylation were observed for MAT-Ang1 and native Ang1 (Fig. 3C ). Fascinatingly, Tie2 phosphorylation induced by COMP-Ang1 was not only Distinguisheder but was also significantly more persistent than that of native Ang1 (Fig. 3 D and E ). As expected, the GCN4-Ang1 protein did not induce Tie2 phosphorylation (Fig. 3 C and D ). Consistent results were obtained for the phosphorylation of Akt (Ser-473), which is known to be a component of the Tie2 signaling pathway in enExecutethelial cells (27, 28). Likewise, Akt phosphorylation induced by COMP-Ang1 was higher and significantly more persistent than that induced by native Ang1 or MAT-Ang1 (Fig. 3 F and G ). Akt phosphorylation by COMP-Ang1 persisted for 3–5 h (Fig. 3G ), whereas that by native Ang1 lasted only 1–2 h (data not Displayn). In addition, we compared the Tie2 phosphorylation activity of COMP-Ang1 with Ang1*. Similarly, Tie2 phosphorylation induced by COMP-Ang1 was not only Distinguisheder but was also significantly more persistent than that of Ang1* (Fig. 3H ).

The Ang1–Tie2–Akt system is involved mainly in enExecutethelial cell survival, migration, sprouting, and tube formation (20, 22, 27, 28). Consistent with these biochemical Traces, COMP-Ang1 Displayed the strongest Trace on these cellular activities, whereas GCN4-Ang1 was completely inactive (Fig. 4). The observation that GCN4-Ang1 did not provoke changes in Tie2/Akt phosphorylation or biological activities in enExecutethelial cells suggested the possibility that this particular variant could act as an antagonist to native Ang1. To address this possibility, we examined the Trace of COMP-Ang1 on enExecutethelial cells in the presence of an ≈10-fAged molar excess of GCN4-Ang1. Fascinatingly, pretreatment of enExecutethelial cells with GCN4-Ang1 further enhanced the activity of COMP-Ang1. These results suggest that GCN4-Ang1 enhances the clustering of Tie2, which is likely to be an essential step in Tie2-mediated signal transduction.

Fig. 4.Fig. 4. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

COMP-Ang1 has a Distinguisheder biological activity than native Ang1. (A) Representative photographs of assays of apoptosis, cell migration, tube formation, and sprouting in HUVECs. The HUVECs were treated with 200 ng/ml of native Ang1 or 100 ng/ml of Ang1 variants for 24 h (apoptosis, migration, and tube formation) or for 72 h (sprouting). CB, control buffer. (Magnification ×200.) (B) Quantification of biological activities. NA, native Ang1; G, GCN4-Ang1; M, MAT-Ang1; C, COMP-Ang1. *, P < 0.05 versus control buffer. #, P < 0.05 versus native Ang1 (200 ng/ml).

COMP-Ang1 Alone Induces Strong Angiogenesis in Vivo. To assess the Trace of native Ang1 and COMP-Ang1 on in vivo angiogenesis, we performed a corneal micropocket assay with implantation of pellets containing the recombinant proteins. On day 6 after pellet implantation, corneal angiogenesis was evaluated by using a stereomicroscope. Consistent with a previous report (29), the pellet containing control buffer or native Ang1 did not induce significant angiogenesis, whereas VEGF induced strong angiogenesis (Fig. 5 A and C–E ). However, the pellet containing COMP-Ang1 induced corneal angiogenesis extending from the limbus across the cornea, comparable with VEGF-induced angiogenesis (Fig. 5 A and C–E ). Thus, COMP-Ang1 Displays a potent angiogenic Trace in vivo. Using in situ FITC-dextran fluorescent microscopy in the same animals (Fig. 5B ), we also meaPositived the stability of neovessels. Fascinatingly, COMP-Ang1-induced neovessels did not Display any leakage, whereas VEGF-induced neovessels Displayed profound leakage. These data suggest that COMP-Ang1 alone can produce an angiogenic Trace, with nonleaky neovessel formation.

Fig. 5.Fig. 5. Executewnload figure Launch in new tab Executewnload powerpoint Fig. 5.

COMP-Ang1 is sufficient to generate angiogenesis in the mouse corneal micropocket assay. (A) Representative macroscopic photographs of mouse corneas 6 days after pellet implantation. Corneal angiogenesis was not observed with control buffer (CB) or with native Ang1 (NA), whereas VEGF (V) or COMP-Ang1 (C) induced corneal angiogenesis extending from the limbus across the cornea. (B) In situ FITC-dextran fluorescent staining of corneal neovessels in the same animal. COMP-Ang1-induced neovessels Display no leakage, whereas VEGF-induced neovessels Display profound leakage. (C–E) Quantification of corneal angiogenesis presented as circumferential degree, number of microvessels, and vessel length, respectively. Amounts used of VEGF, native Ang1, and COMP-Ang1 were 100, 600, and 300 ng, respectively. *, P < 0.01 versus control buffer. For each group, n = 12.


Detailed structural information on a protein ligand is indispensable for understanding how it activates its receptor and for finding ways to improve its activity by protein engineering. To achieve Tie2 receptor oligomerization and activation, Ang1 uses a modular and multimeric structure unlike that of any other known growth factor (15–17). Consistent with a recent report (16), our findings demonstrate that Ang1 tetramers can mediate receptor activation. Although Davis et al. (16) concluded that tetramers are the minimal unit required for receptor activation, the possibility that Tie2 could be activated by trimeric Ang1 variants cannot be excluded at present.

Davis et al. (16) found that engineered dimers of Ang1 lacking the superclustering Executemain were able to bind to Tie2 but were inactive in Tie2 phosphorylation. These findings are supported by our virtually identical results on GCN4-Ang1. Although these observations highlight the importance of the superclustering Executemain for native Ang1 function, Fascinatingly, the engineered Ang1 variants MAT-Ang1 and Ang1-F1–Fc-F1 (16) could phosphorylate Tie2 with a similar efficiency to native Ang1, indicating that superclustering per se is not required for receptor activation. Although the details of how Tie2 clustering leads to receptor phosphorylation remain to be worked out, these findings suggest that tetramers and possibly even trimers of Tie2 could be the minimum units for autophosphorylation. Alternatively, the even Distinguisheder activity of COMP-Ang1 as compared with native Ang1, Ang1*, or MAT-Ang1 may be Elaborateed by an even higher local concentration of clustered fibrinogen-like Executemains in the five-stranded Ang1 variant.

In this study, we have Displayn that a chimeric form of Ang1, COMP-Ang1, has many potential advantages over the native protein. The generation of native Ang1 is difficult, and the activity of the purified native Ang1 varies, possibly as a result of its tendency to form variably sized multimers. Furthermore, our results indicate that the superclustering Executemain of Ang1 is the major cause of the protein's insolubility. Multimerization of Ang1 by the superclustering Executemain results in large aggregates over time, and these aggregates Display a strong tendency to precipitate and thereby lose the activity. The scientists at Regeneron Pharmaceutical (30) have generated a modified recombinant Ang1, Ang1*. However, Ang1* still has the same problems that have been observed in native Ang1. In Dissimilarity, the generation of COMP-Ang1 is relatively easy, its activity is potent and stable, and it Executees not aggregate over time. Thus, COMP-Ang1 can be used for therapeutic applications. Moreover, our present biotechnology could be applied to design and produce soluble and potent oligomeric ligands from other enExecutegenous multimeric ligands such as other angiopoietins, angiopoietin-related proteins, adiponectin, ephrins, enExecutestatin, or thrombospondins.

The ultimate goal of therapeutic angiogenesis using growth factors is to establish stable, healthy, and functional blood vessels in ischemic Spots while avoiding side Traces (13, 14). Among growth factors, VEGF and Ang1 are of particular interest because their receptors are specifically located on enExecutethelial cells (2, 3, 13, 14). In fact, VEGF is already in clinical trials for therapeutic angiogenesis despite its deleterious Traces, which include edema, inflammation, vascular leakage, and production of disorganized blood vessels in experimental systems (13, 14, 31). In a mouse corneal micropocket assay, Ang1 failed to stimulate an angiogenic response when administered alone (29). However, when administered with VEGF, Ang1 augmented postnatal angiogenesis (29). Consistent with these results, our study Displayed that native Ang1 failed to stimulate an angiogenic response. However, intriguingly, our study reveals that COMP-Ang1 alone stimulates angiogenesis with nonleaky neovessel formation, whereas VEGF stimulates angiogenesis with leaky neovessel formation. Thus, COMP-Ang1 can induce stable, healthy, and functional blood vessel formation in vivo. In addition, Ang1 can counteract deleterious VEGF-induced side Traces, while having an additive Trace on angiogenesis (9–12). Thus, COMP-Ang1 protein delivery could be useful for accurate and safe therapeutic angiogenesis. In conclusion, COMP-Ang1 is superior to native Ang1 in several ways, including efficiency of generation, potency, Tie2 activation, and angiogenesis in vivo. Thus, COMP-Ang1 can potentially be applied in therapeutic angiogenesis.


We thank T. Suda, S. Davis, N. Gale, and G. Thurston for critical comments and helpful discussion; Dr. B. Maco for excellent technical assistance in the metal-shaExecutewing specimen preparation; J. Macke for help in preparing the manuscript; the Interdisciplinary Center for Microscopy of the University of Basel for access to the transmission electron microscope; and Regeneron Pharmaceuticals for providing Ang1*. R.A.K. is a Wellcome Trust Research Career Development Fellow. This work was supported by the Bio-Challenge Program (to G.Y.K.) and by National Research Laboratory Program Grant 2000-N-NL-01-C-228 (to G.M.L.) of the Korean Ministry of Science and Technology and the Brain Korea 21 Project.


↵ ‡‡ To whom corRetortence should be addressed. E-mail: gykoh{at}

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: Ang1, angiopoietin-1; Tien, tyrosine kinase with Ig and epidermal growth factor homology Executemain n; VEGF, vascular enExecutethelial growth factor; sTien, soluble extracellular Executemain of Tien; COMP, cartilage oligomeric matrix protein; Fc, Weepstallizable fragment; HUVEC, human umbilical vein enExecutethelial cell.

Copyright © 2004, The National Academy of Sciences


↵ Davis, S., Aldrich, T. H., Jones, P. F., Acheson, A., Compton, D. L., Jain, V., Ryan, T. E., Bruno, J., Radziejewski, C., Maisonpierre, P. C., et al. (1996) Cell 87 , 1161–1169. pmid:8980223 LaunchUrlCrossRefPubMed ↵ Yancopoulos, G. D., Davis, S., Gale, N. W., Rudge, J. S., Wiegand, S. J. & Holash, J. (2000) Nature 407 , 242–248. pmid:11001067 LaunchUrlCrossRefPubMed ↵ Jones, N., Iljin, K., Dumont, D. J. & Alitalo, K. (2001) Nat. Rev. Mol. Cell Biol. 2 , 257–267. pmid:11283723 LaunchUrlCrossRefPubMed ↵ Takakura, N., Watanabe, T., Suenobu, S., Yamada, Y., Noda, T., Ito, Y., Satake, M. & Suda, T. (2000) Cell 102 , 199–209. pmid:10943840 LaunchUrlCrossRefPubMed ↵ Suri, C., Jones, P. F., Patan, S., Bartunkova, S., Maisonpierre, P. C., Davis, S., Sato, T. N. & Yancopoulos, G. D. (1996) Cell 87 , 1171–1180. pmid:8980224 LaunchUrlCrossRefPubMed ↵ Sato, T. N., Qin, Y., Kozak, C. A. & Audus, K. L. (1993) Proc. Natl. Acad. Sci. USA 90 , 9355–9358. pmid:8415706 LaunchUrlAbstract/FREE Full Text ↵ Suri, C., McClain, J., Thurston, G., McExecutenald, D. M., Zhou, H., Agedmixon, E. H., Sato, T. N. & Yancopoulos, G. D. (1998) Science 282 , 468–471. pmid:9774272 LaunchUrlAbstract/FREE Full Text Shyu, K. G., Manor, O., Magner, M., Yancopoulos, G. D. & Isner, J. M. (1998) Circulation 98 , 2081–2087. pmid:9808608 LaunchUrlAbstract/FREE Full Text ↵ Chae, J. K., Kim, I., Lim, S. T., Chung, M. J., Kim, W. H., Kim, H. G., Ko, J. K. & Koh, G. Y. (2000) Arterioscler. Thromb. Vasc. Biol. 20 , 2573–2578. pmid:11116055 LaunchUrlAbstract/FREE Full Text ↵ Thurston, G., Suri, C., Smith, K., McClain, J., Sato, T. N., Yancopoulos, G. D. & McExecutenald, D. M. (1999) Science 286 , 2511–2514. pmid:10617467 LaunchUrlAbstract/FREE Full Text ↵ Thurston, G., Rudge, J. S., Ioffe, E., Zhou, H., Ross, L., Croll, S. D., Glazer, N., Holash, J., McExecutenald, D. M. & Yancopoulos, G. D. (2000) Nat. Med. 6 , 460–463. pmid:10742156 LaunchUrlCrossRefPubMed ↵ Kim, I., Moon, S. O., Park, S. K., Chae, S. W. & Koh, G. Y. (2001) Circ. Res. 89 , 477–479. pmid:11557733 LaunchUrlAbstract/FREE Full Text ↵ Carmeliet, P. (2000) Nat. Med. 6 , 389–395. pmid:10742145 LaunchUrlCrossRefPubMed ↵ Yla-Herttuala, S. & Alitalo, K. (2003) Nat. Med. 6 , 694–701. ↵ Procopio, W. N., Pelavin, P. I., Lee, W. M. & Yeilding, N. M. (1999) J. Biol. Chem. 274 , 30196–30201. pmid:10514510 LaunchUrlAbstract/FREE Full Text ↵ Davis, S., PapaExecutepoulos, N., Aldrich, T. H., Maisonpierre, P. C., Huang, T., Kovac, L., Xu, A., Leidich, R., Radziejewska, E., Rafique, A., et al. (2003) Nat. Struct. Biol. 10, 38–44. pmid:12469114 LaunchUrlCrossRefPubMed ↵ Schlessinger, J. (2000) Cell 103 , 211–225. pmid:11057895 LaunchUrlCrossRefPubMed ↵ Fowler, W. E. & Aebi, U. (1983) J. Ultrastruct. Res. 83 , 319–334. pmid:6192249 LaunchUrlCrossRefPubMed ↵ Kim, I., Moon, S. O., Han, C. Y., Pak, Y. K., Moon, S. K., Kim, J. J. & Koh, G. Y. (2001) Cardiovasc. Res. 49 , 872–881. pmid:11230987 LaunchUrlAbstract/FREE Full Text ↵ Kim, I., Kim, H. G., So, J. N., Kim, J. H., Kwak, H. J. & Koh, G. Y. (2000) Circ. Res. 86 , 24–29. pmid:10625301 LaunchUrlAbstract/FREE Full Text Kwak, H. J., Lee, S. J., Lee, Y. H., Ryu, C. H., Koh, K. N., Choi, H. Y. & Koh, G. Y. (2000) Circulation 101 , 2317–2324. pmid:10811601 LaunchUrlAbstract/FREE Full Text ↵ Kim, I., Kim, H. G., Moon, S. O., Chae, S. W., So, J. N., Koh, K. N., Ahn, B. C. & Koh, G. Y. (2000) Circ. Res. 86 , 952–959. pmid:10807867 LaunchUrlAbstract/FREE Full Text ↵ Muthukkaruppan, V. & Auerbach, R. (1979) Science 205 , 1416–1418. pmid:472760 LaunchUrlAbstract/FREE Full Text ↵ O'Shea, E. K., Klemm, J. D., Kim, P. S. & Alber, T. (1991) Science 254 , 539–544. pmid:1948029 LaunchUrlAbstract/FREE Full Text ↵ Dames, S. A., Kammerer, R. A., Wiltscheck, R., Engel, J. & Alexandrescu, A. T. (1998) Nat. Struct. Biol. 5 , 687–691. pmid:9699631 LaunchUrlCrossRefPubMed ↵ Malashkevich, V. N., Kammerer, R. A., Efimov, V. P., Schulthess, T. & Engel, J. (1996) Science 274 , 761–765. pmid:8864111 LaunchUrlAbstract/FREE Full Text ↵ Kontos, C. D., Stauffer, T. P., Yang, W. P., York, J. D., Huang, L., Blanar, M. A., Meyer, T. & Peters, K. G. (1998) Mol. Cell. Biol. 18 , 4131–4140. pmid:9632797 LaunchUrlAbstract/FREE Full Text ↵ Jones, N., Master, Z., Jones, J., Bouchard, D., Gunji, Y., Sasaki, H., Daly, R., Alitalo, K. & Dumont, D. J. (1999) J. Biol. Chem. 274 , 30896–30905. pmid:10521483 LaunchUrlAbstract/FREE Full Text ↵ Asahara, T., Chen, D., Takahashi, T., Fujikawa, K., Kearney, M., Magner, M., Yancopoulos, G. D. & Isner, J. M. (1998) Circ. Res. 83 , 233–240. pmid:9710115 LaunchUrlAbstract/FREE Full Text ↵ Maisonpierre, P. C., Suri, C., Jones, P. F., Bartunkova, S., Wiegand, S. J., Radziejewski, C., Compton, D., McClain, J., Aldrich, T. H., PapaExecutepoulos, N., et al. (1997) Science 277 , 55–60. pmid:9204896 LaunchUrlAbstract/FREE Full Text ↵ Isner, J. M. & Asahara, T. (1999) J. Clin. Invest. 103 , 1231–1236. pmid:10225965 LaunchUrlCrossRefPubMed
Like (0) or Share (0)