The N-terminal Region of NTAK/Neuregulin-2 Isoforms Has an Inhibitory Activity on Angiogenesis*

NTAK (neural- and thymus-derived activator for ErbB kinases), also known as neuregulin-2, is a member of the epidermal growth factor (EGF) family, which binds directly to ErbB3 and ErbB4 and transactivates ErbB2. Because ErbB signaling has been implicated in various angiogenic mechanisms, the effect of NTAK (which has at least nine isoforms due to alternative splicing) in angiogenesis is explored. One isoform, NTAK (cid:1) , inhibited cell growth in terms of DNA synthesis and cell numbers in vascular endothelial cells specifically, whereas NTAK (cid:2) and (cid:3) had no activity. On the other hand, NTAK (cid:1) secreted by transfected MDA-MB-231 cells inhibited endothelial cell growth, and NTAK (cid:1) expressed in endothelial cells by adenovirus infection suppressed cell growth in a dose-dependent manner. The EGF-like domain of NTAK (cid:1) did not have this activity. The NTAK (cid:4) isoform, which had the Ig-like domain but not the EGF-like domain, inhibited proliferation of endothelial cells. NTAK (cid:4) prevented hyper-phosphorylation of the retinoblastoma tumor suppressor protein and caused G 1 arrest in endo- thelial cells. Both NTAK (cid:1) and (cid:4) isoforms displayed antiangiogenic activity in the chick embryo chorioallantoic membrane in vivo . These results suggest that the active site of NTAK is localized outside of the EGF-like domain but within the N-terminal region, including the Ig-like domain, of NTAK. activity assayed using chorioallantoic incubated at 37 °C g was within a sterilized on CAM. were incubated at 37 °C for 48 A emulsion chorioallantois and the vascular networks in the CAMs and photographed using a is specific

The human NTAK gene comprises 12 exons spanning Ͼ55 kilobases (13). Among the products of alternative splicing, the ␣ isoform of the NTAK gene is expressed in all tissues including the brain, and the ␤ isoform is restricted to the brain. The ␥ isoform is expressed in a rat pheochromocytoma cell line, PC-12. NTAK␦ is an isoform missing the EGF-like domain and is expressed in a human neuroblastoma cell line, SK-N-SH. NTAK␣ and ␤ preferentially induce phosphorylation in ErbB3 and ErbB4, respectively, transactivate ErbB2, and stimulate growth of human breast cancer cells (14). However, differences in the biological roles and the functions of the various NTAK isoforms are still unknown.
The roles of ErbB overexpression in cellular transformation and tumor metastasis have been elucidated by a line of in vitro and clinical studies. Furthermore, targeted deletion of ErbB2, ErbB3, ErbB4, or NRG1 in mice leads to developmental abnormalities that are severe in the nervous system and to lethality due to cardiovascular system failure (5,6). The cardiac abnormalities include aborted development of the endocardial cushion, which is dependent on mesenchymal cell growth and development of the endocardial endothelium.
Angiogenesis is the process of new vascular formation from preexisting blood vessels and is tightly regulated by the balance of angiogenic factors and inhibitors (15). Under normal conditions, vascular endothelial cells are quiescent due to the dominance of angiogenic inhibitory factors, including angiostatin (16), endostatin (17), and NK4 (18). Angiogenesis occurs during pathological events such as solid tumor growth and metastasis, diabetic retinopathy, atherosclerosis, and rheumatoid arthritis. Angiogenic inhibitors are capable of preventing tumor growth and metastasis, and, in fact, a number of angiogenic inhibitors are being tested in clinical trials for cancer treatment.
ErbB signaling has also been implicated in angiogenesis. Neutralizing antibodies against ErbB1 and ErbB2 down-regulate VEGF and inhibit tumor growth and angiogenesis in vivo (19). NRG1 has been reported to activate ErbBs in endothelial cells and induce angiogenesis (20). NRG1 binds to heparan sulfate proteoglycan (HSPG) via the Ig-like domain, and NRG1-HSPG interaction potentiates ErbB phosphorylation by the EGF-like domain of NRG1 (21). Targeted deletion of the Ig-like domain of NRG1 in mice leads to the embryonic lethality associated with a deficiency of ventricular myocardial trabec-ulation and impairment of cranial ganglion development (22). The Ig-like domain of NTAK is 38.2% identical to the corresponding domain of NRG1 (7). However, the function of the Ig-like domain of NTAK remains unknown.
We scrutinized the angiogenic effects of NTAK isoforms and report herein that the N-terminal region of NTAK, including the Ig-like domain but not the EGF-like domain, inhibits angiogenesis and that NTAK␦ causes G 1 arrest in vascular endothelial cells.
Preparation of Recombinant Proteins from Escherichia coli-The 1.2kbp cDNA fragments corresponding to the extracellular regions of NTAK␣2a, NTAK␤, NTAK␥, and NTAK␦ were inserted into pASK-IBA6 vectors (Genosys Biotechnologies, The Woodlands, TX). The resulting plasmids were then used to transform E. coli DH5␣, and the recombinant proteins were purified using a StrepTactin affinity column according to the manufacturer's instructions (Genosys Biotechnologies).
Establishment of NTAK␥ Gene-transfected MDA-MB-231 Cells-An expression vector of NTAK␥ was constructed by inserting the full sequence of NTAK␥ cDNA into pRc/CMV. The vector was transfected into MDA-MB-231 cells by the electroporation method, and selection was performed with G418 (Sigma-Aldrich). The conditioned media from each clone were concentrated using heparin-Sepharose (Amersham Biosciences), and NTAK␥ expression was examined by Western blotting analyses using the anti-NTAK antibody, #N1-1 (7). Three highly expressing clones, 231␥-1, -2, and -3, were independently selected.
Cell Growth Assay-For the cell number assay, cells were resuspended in the maintenance medium and then seeded onto collagen-coated 24-well microplates (2 ϫ 10 3 cells/1 ml/well). The plates were incubated for 12 h at 37°C and then re-fed medium containing 2% FCS. After 12 h, the samples to be tested were added. After 2 days of incubation, the cells were harvested and counted using a Coulter counter (Coulter Electronics, Luton, England). For the assay of DNA synthesis, cells were resuspended in the maintenance medium and then seeded onto collagen-coated 96-well microplates (2 ϫ 10 3 cells/200 l/ well). The plates were incubated for 24 h at 37°C and then re-fed medium containing 5% FCS. After 3 h, the samples or growth factors to be tested for inhibitory activity were added. After a 24-h incubation, 10 l (37 kBq) of [ 3 H]thymidine was added, and incubation was continued for another 6 h. The [ 3 H]thymidine incorporation into DNA was determined by liquid scintillation counting (1450 MicroBeta TRILUX; PerkinElmer Life Sciences). To determine HUVEC growth, HUVECs were plated at a density of 1 ϫ 10 4 cells/1 ml/well in collagen-coated 12-well microplates. After 12 h, the plates were re-fed MCDB-131/FCS (5%)/PS, followed by the addition of 20 ng/ml NTAK␣ or NTAK␦ (Day 0). The plates were incubated for 48 h (Day 2) and then re-fed fresh MCDB-131/FCS (5%)/PS without NTAKs.
Inhibitory Effect of Adenoviral Vectors-We used an adenovirus expression vector kit (Takara, Kusatsu, Shiga, Japan) that carries a cytomegalovirus-driven cDNA for full-length NTAK␥. For the cell growth assay, we measured DNA synthesis in HUVECs infected by adenovirus for 24 h, as described above.
Immunoprecipitation and Western Blotting Analyses of Retinoblastoma Protein-To synchronize the cells in a quiescent state (G 0 ), the cells were cultured in starvation medium containing 0.5% FCS for 16 h and re-fed fresh medium with or without 50 ng/ml NTAK␦. Cells were harvested and lysed with lysis buffer (1% Triton X-100, 1 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 5 mM EDTA, 0.1 mM aprotinin, and 1 mM phenylmethylsulfonyl fluoride). After centrifugation at 15,000 rpm for 10 min, the supernatant was incubated with anti-Rb antibody for 2 h at 4°C and then with 20 l of protein A-Trisacryl (50% suspension; Pierce) for 2 h at 4°C. The samples were analyzed by electrophoresis on a 6% polyacrylamide gel. Proteins in the gel were transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA) in 150 mM CAPS buffer, pH 10.5, containing 20% methanol. The membrane was blocked with 5% Asterisk, COOH terminus. Gaps are inserted in the sequence to provide optimal alignment. skimmed milk in PBS overnight at 4°C and incubated with anti-Rb antibody in 5% skimmed milk/PBS for 2 h at room temperature. The membrane was washed three times at 10-min intervals with 0.05% Tween 20 in PBS and then incubated with biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA) for 1 h at room temperature. The membrane was washed again three times at 10-min intervals with 0.05% Tween 20 in PBS and then incubated with avidin-HRP (Vectastatin, Vector Laboratories) for 30 min at room temperature. The membrane was washed five times with 0.05% Tween 20 in PBS, and the horseradish peroxidase activity was detected using an ECL kit (Amersham Biosciences) followed by autoradiography.
In Vivo Angiogenesis Assay-Antiangiogenic activity in vivo was assayed using the chick chorioallantoic membrane (CAM) method (24). Briefly, fertilized white Leghorn chicken eggs were incubated at 37°C for 5 days, and a methyl cellulose disk containing 100 g of the samples to be tested was placed within a 5-mm-round sterilized silicon ring on CAM. The eggs were incubated at 37°C for 48 h. A white fat emulsion (Intralipos; Yoshitomi, Osaka, Japan) was injected into the chorioallantois and the vascular networks in the CAMs and photographed using a digital camera.

Growth Inhibition of Endothelial Cells Stimulated by
NTAK-NTAK has at least nine alternatively spliced isoforms. The differences in biological properties among NTAK isoforms remain unclear. To examine the effect of NTAK on angiogenesis, we examined three isoforms, NTAK␣, ␤, and ␥, whose structures differ in their EGF-like domain (Fig. 1). We analyzed the DNA synthesis of BAECs stimulated by recombinant NTAKs. NTAK␣ and ␤ dose-dependently induced phosphorylation of ErbBs in MDA-MB-453 and T47D cells within the 1-10 ng/ml range (14). At these levels, NTAK␣ and ␤ had neither stimulatory nor inhibitory effects on [ 3 H]thymidine incorporation by BAECs, but NTAK␥ inhibited it in a dose-dependent manner ( Fig. 2A). Similarly, NTAK␣ and ␤ isoforms had no effects on the proliferation of BAE cells, whereas NTAK␥ was inhibitory (Fig. 2B). The inhibitory activity of NTAK␥ was comparable with that of a typical angiogenic inhibitor, TNF-␣ (25). However, NTAK␥ had no effects on the growth of smooth muscle cells (SMCs), T47D cells, MDA-MB-453 cells, or SK-N-SH cells (Fig. 2C). 2 Thus, it is suggested that the growth inhibitory activity of NTAK␥ is specific for vascular endothelial cells.
Growth Inhibition of Endothelial Cells Co-cultured with NTAK␥ Transfectants-Human breast cancer-derived MDA-MB-231 cells are undifferentiated cells that are invasive by nature and secrete angiogenic factors such as VEGF and FGF-2 (26,27). We examined whether NTAK␥ was capable of antagonizing these angiogenic factors. We transfected a plasmid containing the full-length NTAK␥ cDNA and the neomycin resistance gene into MDA-MB-231 cells. Three cell lines, 231␥-1, -2, and -3, secreting high levels of NTAK␥, were obtained and co-cultured with HUVECs. These transfectants secreted VEGF and FGF-2 as verified by an enzyme-linked immunosorbent assay (data not shown). The [ 3 H]thymidine incorporation into HUVECs in the lower chambers was increased by the angiogenic factors secreted byMDA-MB-231cells in the upper chambers. When three transfectants were cocultured in the upper chambers, [ 3 H]thymidine incorporation into HUVECs was reduced to the basal level (Fig. 3A), and the expected increase in HUVEC number was inhibited (Fig. 3B). Thus, NTAK␥ was able to reduce the growth of endothelial cells to the basal level, probably by antagonizing angiogenic factors.
Growth Inhibition of Adenovirus-infected HUVEC-We then evaluated potential therapeutic actions via NTAK delivery. HUVECs were infected with adenoviral vectors encoding NTAK␥ or a control. The expressions of vectors were verified histochemically by ␤-galactosidase activity. The DNA synthesis of HUVECs infected by the adenoviral vector encoding NTAK␥ was inhibited in a dose-dependent manner, and the half-maximal effectiveness was ϳ5 ϫ 10 2 multiplicities of infection (Fig.  4). The cells showed no morphological changes, and the control vector had no effect on HUVEC proliferation. Thus, growth inhibition of HUVECs was not due to the cytotoxic effect of viral infection but to NTAK␥ expressed by adenoviral vectors. This suggested that NTAK␥ has a potential therapeutic use for angiogenesis-related diseases. The Active Site of Angiogenic Inhibition in NTAK-NTAK␣ and ␤ isoforms have complete EGF-like domain motifs, which have consensus sequences containing three disulfide bonds. On the contrary, the NTAK␥ isoform has an incomplete EGF-like domain, which does not have the third cysteine loop (Fig. 1B). To investigate the active site of NTAK involved in angiogenic inhibition, we focused on the EGF-like domain of NTAK␥ and synthesized a 33-amino acid-long peptide of ␥EGF corresponding to the region. In the assays of cell number and DNA synthesis, this ␥EGF peptide had no effect on the growth of HUVECs and BAECs at concentrations up to 3.3 nM (Fig. 5A and data not shown), whereas the NTAK␥ protein showed an inhibitory activity at 10 ng/ml or 0.22 nM, as shown in Fig. 2, A and B. Detectable levels of ErbB phosphorylation were not induced by the NTAK␥ or the ␥EGF peptide in Western blotting analysis. 2 It was suggested, therefore, that the active site for the angiogenic inhibitory effect of NTAK␥ existed in a region other than the EGF-like domain. Subsequently, we examined the angiogenic inhibitory effect of NTAK␦, which has the same N-terminal part, including the Ig-like domain, as the other isoforms but is missing the EGF-like domain (Fig. 1A). NTAK␦ decreased [ 3 H]thymidine incorporation into HUVECs in a dose-dependent manner (Fig. 5B) with no accompanying ErbB phosphorylation (data not shown). It is suggested that the anti-angiogenic activity of NTAK resides in the N-terminal region of NTAK, including the Ig-like domain, but not in the EGF-like domain of NTAK␥.
NTAK␦ Reduces Retinoblastoma Protein Phosphorylation in HUVEC-We investigated the mechanism underlying the angiogenic inhibitory effects of NTAK␦. Retinoblastoma protein (pRb) is the most important protein regulating cell cycle progression into the S phase, and hyper-phosphorylation of pRb inactivates of its growth inhibitory function (28). We investigated whether NTAK␦ affected pRb hyper-phosphorylation. When serum-starved HUVECs were re-fed fresh medium, the hyper-phosphorylation of pRb was increased. However, treatment of HUVECs with NTAK␦ inhibited hyper-phosphorylation of pRb, suggesting that the G 1 arrest was induced by NTAK␦ (Fig. 6).
To determine whether the inhibition of angiogenesis by NTAK␦ is reversible, we analyzed HUVEC growth in response to with NTAK␣ or ␦. HUVECs treated with 20 ng/ml NTAK␦ became spindle-shaped on Day 1 and cell growth was reduced, whereas HUVECs treated with NTAK␣ proliferated normally (Fig. 7). The NTAK␦-treated cells did not show apoptotic or necrotic changes, and growth was restored when the medium was replaced, on Day2, with medium not containing NTAK␦. Thus, it was suggested that NTAK␦ blocks cell cycle progression and causes G 1 arrest in HUVECs.
In Vivo Anti-angiogenic Activity of NTAK-The in vivo an-giogenic inhibitory activity of NTAKs was examined using the CAM assay. NTAK␣ and control bovine serum albumin neither induced nor inhibited angiogenesis of chick microvessels on CAMs (Fig. 8, A and B). In contrast, the microvascular network formation was inhibited by NTAK␥ and, more markedly, by NTAK␦ (Fig. 8, C and D). Avascular areas were clearly formed surrounding the disks containing NTAK␥ or NTAK␦. Thus, NTAK␥ and NTAK␦, but not NTAK␣, inhibited angiogenesis in vivo as well as in vitro.

DISCUSSION
NTAK represents at least nine alternative splicing isoforms derived from the same gene as NRG2 (7)(8)(9)(10)13), and belongs to the EGF family. In our previous study, NTAK␣ and ␤ bound directly to the ErbB3 and ErbB4 receptors, but not to ErbB1 or ErbB2 (7). It has been reported that targeted deletion of ErbB2, ErbB3, ErbB4, or NRG1 in mice is lethal due to developmental abnormalities involving the cardiovascular system. Effects of NTAK in angiogenesis have been suggested, because endothelial cells express ErbB2, ErbB3, and ErbB4 (20). Recombinant NTAK␣ and ␤ isoforms using an E. coli expression system stimulated breast tumor cell growth and differentiation at physiological concentrations ranging from 1 to 10 ng/ml (14). However, NTAK␣ and ␤ neither stimulated nor inhibited endothelial cell growth. NTAK␥ did not stimulate the phosphorylation of ErbBs in MDA-MB-453 and T47D cells (data not shown) but did inhibit endothelial cell growth in terms of DNA synthesis and increase of cell number (Fig. 2, A and B). Furthermore, NTAK␥, as well as NTAK␣ and ␤, had no effect on the proliferation of SMCs, breast cancer cell lines, and neuro-

FIG. 3. Growth of HUVECs co-cultured with NTAK␥-transfectants.
HUVECs were seeded onto collagencoated 24-well microplates at a density of 2 ϫ 10 4 cells/well. After a 6-h incubation, the cells were re-fed medium containing 5% FCS, and the inner cups containing MDA-MB-231 cells or transfectants (1 ϫ blastoma cell lines. These observations suggest that the cellular effect of NTAK␥, but not that of NTAK␣ and ␤, is specific for vascular endothelial cells. MDA-MB-231 cells secrete the angiogenic factors VEGF and FGF-2, which promote the growth of co-cultured endothelial cells. All three NTAK␥-transfected MDA-MB-231 cells, 231␥-1, -2 and -3, secreted these angiogenic factors (data not shown) and showed reduced [ 3 H]thymidine incorporation and proliferation of HUVECs to the basal level. NTAK␥ was able to antagonize the angiogenic activities such as VEGF and FGF-2 (Fig.  3). Furthermore, NTAK␥ expression by adenoviral vector in HUVECs impaired DNA synthesis in a dose-dependent manner (Fig. 4). The angiogenic inhibitory activities of the NTAK␥ and ␦ isoforms were demonstrated in vivo by CAM assays. NTAK␣ had neither stimulatory nor inhibitory effects on angiogenesis on CAMs. NTAK␥ and NTAK␦ inhibited new vessel formation (Fig. 8). Thus, NTAK␥ and ␦ exhibited anti-angiogenic activities in vivo as well as in vitro, suggesting that these NTAKs may be possible candidate agents for suppression of tumor growth and metastasis.
All members of the EGF family have the EGF-like domain with the consensus sequence containing three disulfide bonds. The EGF-like domain binds to receptor ErbB family members and promotes cell growth, differentiation, and survival. NTAK␥ inhibited endothelial cell growth, whereas NTAK␣ and ␤ had no effect. The structural differences among these NTAK isoforms are found only in the EGF-like domain. NTAK␣ and ␤ isoforms have the full EGF-like domain motif, whereas NTAK␥ has a truncated EGF-like domain missing the third cysteine loop. Unexpectedly, the ␥EGF peptide, corresponding to the EGF-like domain of NTAK␥, had no effect on the growth of HUVECs and BAECs at concentrations up to 3.3 nM, although NTAK␥ exerted an inhibitory effect at 10 ng/ml (0.22 nM). Otherwise, NTAK␦, which has the Ig-like domain but not the EGF-like domain, inhibited angiogenesis. The ␥EGF peptide and the NTAK␥ and ␦ isoforms did not stimulate ErbB phosphorylation (data not shown). The results suggest that the EGF-like domain of NTAK␥ is not involved in the anti-angiogenic activity and that the active site is in the Ig-like domain.
The function of the Ig-like domain of NTAK remains unknown. NTAK is structurally homologous to NRG1, and the Ig-like domain of NRG1 contains a heparin-binding site and increases ligand-receptor affinity (29). Targeted deletion of the Ig-like domain of NRG1 leads to embryonic lethality. The Iglike domain of NTAK may also interact with glycosaminoglycans on the cell surface and affect the interaction of NTAK with ErbBs. The structures of both the Ig-like and EGF-like domains of NTAK may be essential for ErbB interaction and the FIG. 5. Growth inhibition of endothelial cells stimulated by NTAK. A, HUVECs were seeded onto collagen-coated 24-well plates at a density of 2 ϫ 10 3 cells/well, and, after a 12-h incubation, the cells were re-fed medium containing 2% FCS in the presence of the ␥EGF peptide. After 2 days of incubation, the cells were harvested and counted. B, HUVECs were seeded onto collagen-coated 96-well plates at a density of 2 ϫ 10 3 cells/well, and, after a 24-h incubation, the cells were re-fed medium containing 2% FCS in the presence of NTAK␦. Each point is the mean Ϯ S.D. of triplicate measurements.
FIG. 6. Inhibition of hyper-phosphorylation of pRb in HUVECs stimulated by NTAK␦. HUVECs were incubated in starvation medium containing 0.5% FCS for 16 h and then re-fed fresh medium in the presence (؉) or absence (Ϫ) of 50 ng/ml of NTAK␦. HUVECs were harvested at the indicated times and then lysed. The lysates were immunoprecipitated with anti-Rb antibody and then treated with protein A-Trisacryl. The precipitated proteins were separated by 6% SDS-PAGE, analyzed by Western blotting with anti-Rb antibody, and detected using an ECL kit.

FIG. 7. Effect of NTAK on HUVECs.
HUVECs were seeded onto collagen-coated 12-well microplates (1 ϫ 10 4 cells/1 ml/well). After 12 h, the plates were re-fed medium containing 5% FCS, followed by the addition of 20 ng/ml of NTAK␣ (A, C, and E) or NTAK␦ (B, D, and F) (Day 0). The plates were incubated for 48 h (Day 2, C and D), and re-fed fresh media without NTAK. Panels A and B show cells on Day 1; E and F show cells on Day 3. subsequent stimulation of cell growth. NTAK isoforms have the Ig-like domain exclusively. NTAK␥ and ␦ isoforms showed angiogenic inhibitory activities, whereas the NTAK␣ and ␤ isoforms did not. NTAK␣ and ␤ isoforms presumably stimulate the growth of endothelial cells, because endothelial cells have ErbB2, ErbB3, and ErbB4. We can speculate that NTAK␣ and ␤ have both angiogenic activities because of ErbB phosphorylation via the EGF-like domain and anti-angiogenic properties associated with other regions, including the Ig-like domain, which results in no net effects on cell growth.
It is now well known that some of endogenous angiogenic inhibitors are fragments of proteins whose biological functions are unrelated to angiogenesis. For example, angiostatin and endostatin are proteolytic fragments of plasminogen (16) and collagen XVIII (17), respectively. NK4 is an internal fragment of hepatocyte growth factor (HGF) and is considered to be bifunctional, as it is an hepatocyte growth factor antagonist and an endogenous angiogenic inhibitor (18). NK4 inhibits angiogenesis through another pathway than the c-Met receptor. It is also possible that NTAK proteins with the full EGFlike domain have no effects on angiogenesis and that only fragments of NTAK with the truncated EGF-like domain have the anti-angiogenic properties.
The mechanism underlying the angiogenic inhibitory effects of NTAK isoforms remains unknown. NTAK␥ and ␦ isoforms did not induce ErbB phosphorylation (data not shown). We have searched for a ligand binding NTAK␥ on the endothelial cell surface but have not, to date, found it. NTAK␦ inhibited hyper-phosphorylation of pRb in HUVECs, thus providing molecular confirmation of G 1 arrest induced by NTAK␦. In addition, NTAK␦ did not induce apoptotic or necrotic cell death of endothelial cells. We also investigated the effect of NTAK␦ on other molecules involved in cell cycle regulation such as the p16, p21, and p27 proteins, but their expressions were unchanged on Western blotting analysis (data not shown).
The therapeutic potential of angiogenic inhibitors for angiogenic diseases, such as malignant tumors and diabetic retinopathy, have been proposed. Pathological angiogenesis could be triggered either by up-regulation of angiogenic factors or downregulation of endogenous angiogenic inhibitory factors. Angiogenic inhibitors have been shown to inhibit tumor growth and metastasis, and some angiogenic inhibitors are currently being tested in clinical trials for cancer treatment. We found that the NTAK␥ and ␦ isoforms had angiogenic inhibitory activities, and these isoforms of NTAK have therapeutic potential for angiogenic diseases. The EGF-ErbB family is involved in tumor growth, and EGF family members generally act as stimulators of tumor growth. This is the first report showing that molecules belonging to the EGF family inhibit angiogenesis.
In summary, NTAK␥ and ␦ isoforms, but not ␣ and ␤ isoforms, have angiogenic inhibitory activities in vitro and in vivo. The active sites are localized in the N-terminal region of NTAK containing the Ig-like domain but not the EGF-like domain. Further studies are needed to identify the target of NTAK␥ and ␦ on the endothelial cell surface to delineate the mechanism underlying anti-angiogenesis.