Induction of vascular endothelial growth factor by nitric oxide ... - Nature

Oncogene (1997) 15, 437 ± 442  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Induction of vascular endothelial growth factor by nitric oxide in human glioblastoma and hepatocellular carcinoma cells Keisho Chin1, Yukiko Kurashima1, Tsutomu Ogura1, Hisao Tajiri2, Shigeaki Yoshida2 and Hiroyasu Esumi1 1

Investigative Treatment Division, National Cancer Center Research Institute East and 2Division of Internal Medicine, National Cancer Center Hospital East, Kashiwa, Chiba, Japan

We evaluated the eect of nitric oxide (NO) on vascular endothelial growth factor (VEGF) gene expression in human A-172 glioblastoma cells and human HepG2 hepatocellular carcinoma cells. The mRNA level of VEGF increased in response to S-Nitroso-N-acetlyD,L-penicillamine (SNAP) in both cell lines, and increased in mRNA level well coincided with VEGF protein production in A-172 cells. SNAP at 0.5 mM induced maximal stimulation of 4.4 and 3.7 kb VEGF mRNA expression after 6 h about 11 and 8 fold increase, respectively above control level. Similar VEGF mRNA accumulation was observed also with NOR3, another chemical NO generator. To evaluate the eect of SNAP on VEGF mRNA stability, half-lives of VEGF mRNA were measured in A-172 cells cultured with or without 0.5 mM SNAP and treated with actinomycin D (25 mg/ ml). Half-life for VEGF mRNA was found to be prolonged about 2.4 fold by SNAP. VEGF expression induced by SNAP was inhibited by guanylate cyclase inhibitors, methylene blue (10 mM) and LY-83583 (1 mM), and by the protein synthesis inhibitor, cycloheximide (25 mg/ml). These results suggest that induction of VEGF gene expression by NO is mediated through guanylate cyclase activity and requires on-going protein synthesis. Keywords: vascular endothelial growth factor; nitric oxide; gene expression

Introduction Vascular endothelial growth factor (VEGF) is an angiogenesis regulatory factor and endothelial cellspeci®c mitogen (Leung et al., 1989; Keck et al., 1989). Many evidences indicate that VEGF is regulated by hypoxia in vivo and in vitro (Shweiki et al., 1992; Minchenko et al., 1994). The hypoxia-induced increase of VEGF gene expression is due to both transcriptional activation of corresponding gene and by increased stability of VEGF mRNA (Ikeda et al., 1995). In addition to VEGF, hypoxia is known to upregulate the expression of several genes such as those encoding erythropoietin (Epo) (Krantz, 1991), phosphoglycerate kinase 1, and lactate dehydrogenase A gene (Frith et al., 1994). Epo and VEGF gene expressions are signi®cantly up-regulated not only during hypoxia but also, by cobalt chloride, while it Correspondence: H Esumi Received 4 November 1996; revised 16 April 1997; accepted 16 April 1997

is inhibited by carbon monoxide (Goldberg and Schneider, 1994). These results suggest similarities between oxygen-sensing mechanisms that regulate expression of both these genes (Goldberg and Schneider, 1994). Computer analysis has revealed two 10-base pair sequences located 5' to the human VEGF gene transcription start site which are 90% homologous to a sequence element located within the human Epo 3' hypoxia-responsive enhancer (Beck et al., 1991; Blanchard et al., 1992; Semenza and Wang, 1992; Madan and Curtin, 1993). Furthermore, the rat VEGF 5'-¯anking sequence shares elements homologous to the hypoxia-inducible factor 1 (HIF-1) binding site (Levy et al., 1995), a regulatory element that is recognized by protein(s) that speci®cally bind to an enhancer element located 3' to the Epo gene in a hypoxia-regulated fashion (Semenza and Wang, 1992; Wang and Semenza, 1993). Taken together, these observations suggest that expression of Epo and VEGF genes are likely to be responsive to common oxygen-sensor and oxygen-sensing pathways. NO is known to be a diversely functioning physiological mediator and generated from L-arginine (L-Arg) by nitric oxide synthase (NOS) in many dierent cell types. Ohigashi et al. (1993) demonstrated that (L)-NG-nitroarginine methyl ester (L-NAME), a NOS inhibitor, inhibited both cyclic 3', 5'-guanosine monophosphate (cGMP) generation and Epo production in response to hypoxia, and suggested that NO synthesis is essential for hypoxic stimulation of Epo production. But little is known about the relationship between NO and VEGF. Starting from these observations, we hypothesized that NO could regulate also expression of the VEGF gene. Glioblastoma multiforme, a malignant human primary intracranial brain neoplasm, is known to present elevated expression of VEGF in the periphery of its necrotic area (Plate et al., 1992; Shweiki et al., 1992) and VEGF is essential for its growth (Millauer et al., 1994). In this report we evaluated the eect of NO on VEGF gene expression in A-172 human glioblastoma cells and human HepG2 hepatocellular carcinoma cells as models.

Results Signi®cant increase of VEGF mRNA expression in A-172 and HepG2 cells following treatment with NO donors We detected 3.7 kb and 4.4 kb VEGF message in RNA prepared from A-172 and HepG2 cells. In A-172 cells, as shown in Figure 1a, both VEGF mRNA species

Induction of VEGF by NO K Chin et al

438

were dramatically induced following treatment with 0.5 mM SNAP. Increased mRNA levels were observed as early as 2 h after addition of SNAP, and reached the maximum levels after 4 to 8 h of treatment to decline to about 60% of the maximum by 24 h (Figure 1a). When we examined the eect of SNAP on VEGF mRNA expression in HepG2 cells, results were comparable. Maximal induction of VEGF mRNA was detected in this case after 12 to 24 h of SNAP treatment (Figure 1b). In A-172 cells, maximal VEGF mRNA expression was observed after 6 h with 0.5 mM SNAP by about 11 and 8 fold respectively (Figure 2), although as little as 0.1 mM SNAP was sucient to induce accumulation of VEGF mRNA. Higher doses of SNAP, at and above 1.0 mM, were less eective in inducing VEGF mRNA expression. VEGF mRNA increase was also observed following A-172 cell

a

0.5% DMSO 0

2

4

6

0.5mM SNAP

8 12 24 2

8 12 24 hr

4 6

VEGF

0

2

6

12

24 2

4

6

8

12

24 hr

VEGF β-actin Figure 1 Time course of induction of VEGF mRNA by SNAP in A-172 and HepG2 cells. Cells were incubated for the number of hours indicated after addition of 0.5 mM SNAP under condition described in Materials and methods. Total RNA (20 mg/lane) was analysed by Northern analysis using probes for VEGF (top panel) and b-actin (bottom panel) in A-172 cells (a) and HepG2 cells (b)

a 6

0 SNAP

Involvement of guanylate cyclase in NO-mediated increase of VEGF gene expression To investigate the signal transduction pathway(s) mediating the increase of VEGF mRNA expression occurring in response to SNAP in A-172 cells, we examined the eects of herbimycin A, methylene blue, LY-83583 and staurosporine co-treatment on this parameter. VEGF mRNA induction by SNAP was completely inhibited by 10 mM methylene blue (Figure 4a) as well as by 1 mM LY-83583 (Figure 4b), but not by staurosporine (1 ± 50 nM) or herbimycin A (0.001 ± 0.1 mg/ml) (Figure 4c and d). These results indicate that guanylate cyclase activity is involved in NOmediated induction of VEGF gene expression, whereas protein kinase C or tyrosine kinase activities are not. VEGF mRNA accumulation in response to NO is inhibited by cycloheximide

β-actin b

stimulation with NOR3, another chemical NO generator, but not with acetylpenicillamine, a non-NO generating analogue of SNAP (Figure 3). These results indicate that NO itself is likely to induce VEGF mRNA expression in both A-172 and HepG2 cells.

0

0

0.1

0.25

0.5

hr 1.0

2.0

(mM)

VEGF β-actin b

To clarify whether newly synthesized proteins are necessary for stimulation of VEGF expression, the eects of the protein synthesis inhibitor, cycloheximide, was examined in A-172 cells. Cycloheximide at 25 mg/ ml was sucient to inhibit VEGF mRNA accumulation in response to SNAP (Figure 5), indicating that new protein synthesis is required for NO-mediating induction of VEGF mRNA. NO increases VEGF mRNA stability To evaluate the eect of SNAP on VEGF mRNA stability, we measured the half-lives of VEGF mRNA in A-172 cells in the presence or absence of NO donor. Following 6 h treatment of the cultures with or without 0.5 mM of SNAP, 25 mg/ml of actinomycin D was added to the medium to block further gene transcription and the rate of VEGF mRNA decay was then measured by Northern blot analysis. As shown in Figure 6, half-life values for VEGF mRNA were as follows: 4.3 h without SNAP, vs 10.4 h in the presence of 0.5 mM SNAP. This result indicate that the NO donor increased by about 2.4 fold VEGF mRNA stability in A-172 cells. 0

6 DMSO SNAP AP 0.5% 0.5 0.5

hr NOR3 0.5 1.0

(mM)

VEGF

β-actin Figure 2 Dose response of VEGF expression to SNAP in A-172 cells. A-172 cells were incubated for the 6 h after addition of SNAP with indicated concentrations. (a) Total RNA (20 mg/lane) was analysed by Northern analysis using probes for VEGF (top panel) and b-actin (bottom panel). (b) The corrected density was plotted as a ratio of control

Figure 3 Induction of VEGF mRNA by NOR3 in A-172 cells. A-172 cells were incubated for 6 h after addition of 0.5 mM of SNAP, NOR3 or acetylpenicillamine (AP). Total RNA (20 mg/ lane) was analysed by Northern analysis using probes for VEGF (top panel) and b-actin (bottom panel)


NO increases VEGF protein level in A-172 cells In order to con®rm that increase in VEGF mRNA levels in A-172 cells caused by NO treatment coincided with increase in VEGF protein production, we examined VEGF protein levels in A-172 cells and its culture medium. The antibody against VEGF recognized three bands in treated culture medium with molecular masses of about 24, 22, and 20.5 kDa,

a MB

0.5% DMSO 0

1

0.5mM SNAP

10 50 100

0

1

10

50 100

(µM)

respectively as shown in Figure 7. Trace amount of VEGF protein was detected in untreated cell culture medium but the amount of VEGF remarkably increased upon NO treatment (Figure 7). Compared to the amount of VEGF in the culture medium, cellular content of VEGF protein were much less throughout the treatment period indicating relatively rapid secretion of VEGF (data not shown). VEGF protein isolated from various cell sources has been reported to a SNAP (0.5 mM) Act D (25 µg/ml)

+ + –6

VEGF

0

4

– + 8

16

0

4

8

16

hr

VEGF

β-actin

β-actin

b LY

0.5% DMSO 0

1

b

0.5mM SNAP

10 50 100

0

1

10

50 100

(µM)

VEGF β-actin

c ST

0.5% DMSO 0

1

0.5mM SNAP

10

50

0

1

10

50

(nM)

VEGF β-actin

d HA

0.5% DMSO 0

0.01

0.1

0.5mM SNAP 1.0

0

0.01

0.1

1.0

(µg/ml)

VEGF β-actin Figure 4 Involvement of guanylate cyclase in NO induced VEGF expression. A-172 cells were incubated for 6 h after addition of 0.5 mM SNAP with or without indicated concentration of methylene blue (MB) (a), LY-83583 (LY) (b), staurosporine (ST) (c), and herbimycin A (HA) (d). Total RNA (20 mg/lane) was analysed by Northern analysis using probes for VEGF and b-actin

SNAP (0.5 mM) Cycloheximide (µg/ml)

–

+

–

–

– 25

+ 50

25

Figure 6 Eect of SNAP on VEGF mRNA half-life. (a) A-172 cells were exposed to 0.5 mM of SNAP 6 h before addition of actinomycin D (25 mg/ml), and then total RNA samples were collected at the indicated times after actinomycin D administration and analysed by Northern analysis. The exposure time to the imaging plate was dierent between SNAP(+) and SNAP(7) group. (b) Comparison of VEGF mRNA levels in the cells at each time point was achieved by normalization to b-actin. The corrected density was plotted as a percentage of 0 h value against time

hr Kd 34.0 —

0

2

4

6

8

12

24

27.3 — 50

VEGF

β-actin Figure 5 Cycloheximide inhibition of VEGF mRNA accumulation by NO. After A-172 cells were treated with 0, 25, or 50 mg/ml cycloheximide for 4 h, 0.5 mM of SNAP was added and cultured for 6 h. Total RNA (20 mg/lane) was analysed by Northern analysis using probes for VEGF (top panel) and b-actin (bottom panel)

16.9 — Figure 7 Induction of VEGF protein by NO. A-172 cells were treated with or without SNAP (0.5 mM) for indicated times and 5 ml of conditioned medium from the cell culture was harvested. Heparin-binding proteins were enriched by their capacity to bind to heparin-Sepharose beads and analysed by immunoblotting for the presence of VEGF protein. Arrows indicate three forms of VEGF having 24, 22, and 20.5 kd molecular mass

439


440

have a subunit molecular mass between 18 and 24 kDa being similar to those in A-172 cells. Discussion One of the main biological roles of NO is maintaining relaxation of blood vessel but the full eects of NO on angiogenesis have not been fully understood to date although some reports indirectly suggest that NO may be involved in angiogenesis. The NO-generating drugs nitroglycerin (NTG) and L-arginine facilitate healing of experimental gastric ulcer induced by acetic acid (Konturek et al., 1993). During this process, angiogenesis appears to be increased by administration of exogenous NO donor. Ziche et al. (1993) reported that 10 mM sodium nitroprusside treatment increases [3H]thymidine incorporation in endothelial cells isolated from bovine coronary postcapillary venules, while methylene blue treatment selectively reduces this eect. These results suggest that NO production can exert a proliferative eect on endothelial cells at microvascular level, mediated by guanylate cyclase activation. In this report we show that treatment with a NO donor SNAP increases both VEGF mRNA expression and proteins in A-172 and HepG2 cells (Figures 1, 2 and 7). The induction of VEGF mRNA was achieved not only by SNAP but also by a structurally dierent NOR3 but not by the non-NO-generating SNAP analogue acetylpenicillamine (Figure 3). These results indicate that NO can induce increase of VEGF gene expression in A-172 cells. This is the ®rst demonstration that VEGF expression is induced by NO. Moreover, the evidence that VEGF induction by SNAP can be observed in two cell lines indicate that this is possibly general response of various cell types to NO. The NO-induced increase VEGF mRNA level in A172 cells was accompanied by increased mRNA stability (Figure 6). However, while VEGF mRNA accumulation following 6 h exposure to SNAP was quite signi®cant about 10 times higher than in the controls, increase in mRNA stability under the same conditions was limited (only 2.4 fold). This dierence could suggest that transcriptional activation of the VEGF gene is also involved (Figures 1 and 6). When oxygen demand in a given tissue is higher than its supply, blood ¯ow must increase locally to match the increased need. Vasodilatation must be one of the mechanisms that follow acute increases of blood ¯ow. Endothelium derived relaxing factor (EDRF) has been suggested to play a coordinating role in local blood ¯ow regulation and is NO itself (Palmer et al., 1987). Pohl and Busse (1989) reported that a reduction in the partial pressure of oxygen stimulated the rapid release of EDRF from native and cultured endothelial cells, and this can, temporarily, reduce tissue hypoxia. But if tissue hypoxia is not fully corrected and continues in spite of vasodilatation, compensatory angiogenesis must occur. Angiogenesis is stimulated by hypoxic conditions (Adir et al., 1990). In the regulation of this process, several growth factors such as ®broblast growth factors, transforming growth factor-b (TGF-b) and VEGF are considered to be involved (Folkman and Klagsbrun, 1987; Risau, 1990).

VEGF has been recognized as an important angiogenesis factor during hypoxia-induced angiogenesis (Dvorak et al., 1995; Shweiki et al., 1992) and Ku et al. (1993) showed that VEGF induces EDRFdependent relaxation in coronary arteries. It seems quite reasonable that NO and VEGF cooperate to correct and prevent hypoxia. Tuder et al. (1995) reported that sodium nitroprusside (1074 M) decreased VEGF and VEGF receptor mRNA levels in isolated lungs following sodium nitroprusside perfusion under normoxic and hypoxic condition, result in apparent contrast with the hypothesis reported above. However, it is possible that the NO released from sodium nitroprusside might have induced vasodilatation, rapidly improving the blood ¯ow and reducing tissue hypoxia, which all would result in down-regulation of VEGF mRNA. In addition, hypoxic responses in lung and renal vessel are often opposite to those of other vessels. Ohigashi et al. (1993) demonstrated that LNAME inhibited cGMP generation and Epo production in response to hypoxia, and suggested that NO synthesis is essential for hypoxic stimulation of Epo synthetic pathway. Furthermore, Melillo et al. (1995) provided evidence that inducible NOS itself may be a hypoxia inducible gene, regulated by the HIF-1 transcription factor. But mRNA levels of HIF-1 a and b, the subunits of HIF-1, which have been known to be induced in hypoxic condition (Wang and Semenza, 1995), were not signi®cantly increased in A172 cells under the conditions reported in this work (unpublished), suggesting that NO might not induce all of the hypoxia inducible genes but might partly mimic to hypoxic stimulation. Although VEGF is known to be induced by various conditions such as hypoxia and some cytokines including interleukin-1b (IL-1b) (Li et al., 1995), interleukin-6 (IL-6) (Cohen et al., 1996), TGF-b (Pertovaara et al., 1994), and prostaglandin E2 and E1 (Harada et al., 1994), and protein kinase C has been shown to be involved in signal transduction of VEGF expression induced by PDGF (Finkenzeller et al., 1992), details of the mechanisms of VEGF gene regulation remains to be elucidated. Our results showed that guanylate cyclase is likely to be important for NO-mediated VEGF accumulation (Figure 4). In this sense, VEGF induction by NO appears to be similar to the pathway of Epo gene induction by NO (Ohigashi et al., 1993). This is consistent with the similarity in induction of Epo and VEGF genes during hypoxia (Goldberg and Schneider, 1994; Levy et al., 1995). However Mukhopadhyay et al. (1995) reported that VEGF induction is dependent on c-Src activation and this observation is inconsistent with our present result. NO might take a shorter way the signal transduction pathway to VEGF gene or activate VEGF gene through dierent pathway.

Materials and methods Chemicals Chemical NO donor, S-Nitroso-N-acetyl-D,L-penicillamine (SNAP) and (+)-(E)-4-Ethyl-2-[(E)-hydroxyimino]-5-nitro3-hexenamide (NOR3), were purchased from DOJINDO (Kumamoto, Japan). N-acetyl-D,L-penicillamine (acetylpe-


nicillamine) was obtained from Sigma (St. Louis, MO, U.S.A.). Methylene blue and staurosporine which are respectively inhibitors of guanylate cyclase and protein kinase C, were purchased from Wako (Tokyo, Japan). Herbimycin A, tyrosine kinase inhibitor, cycloheximide and actinomycin D were obtained from Sigma. LY-83583 (6-Anilino-5,8-quinolinequinone), another guanylate cyclase inhibitor, was purchased from Calbiocem-Novabiochem (CA, USA). Cell line and culture conditions A human glioblastoma cell line, A-172, and a human hepatocellular carcinoma cell line, HepG2, were obtained from the Japanese Cancer Research Resources Bank. A172 cells were maintained in Dulbecco's modi®ed Eagle's medium (GIBCO) supplemented with 10% fetal bovine serum (MultiSer; Cytosystems, Australia), 1% nonessential amino acids (GIBCO), and 3.7 g/L of sodium bicarbonate (Wako) at 378C under 5% CO2 in air. HepG2 cells were cultured in Dulbecco's modi®ed Eagles medium: Nutrient Mixture F-12 (Ham) (GIBCO) supplemented with 10% fetal bovine serum (MultiSer; Cytosystems, Australia) and 2.438 g/L of sodium bicarbonate (Wako) at 378C under 5% CO2 in air. A-172 cells were passaged at 1.36104 cells/ cm2 before 12 h and then agents such as SNAP, NOR3, acetylpenicillamine, staurosporine, herbimycine A, methylene blue, and LY-83583 were added into the culture medium. In another A-172 cells, 4 h of 25 or 50 mg/ml cycloheximide treatment were performed before administration of 0.5 mM of SNAP. HepG2 cells were passaged at 2.76104 cells/cm2 before 24 h and then SNAP were administrated to the culture medium. RNA extraction and Northern blot analysis Total RNA from A-172 and HepG2 cells were isolated by the guanidinium thiocyanate extraction method (Chomczynski and Sacchi, 1987). Equal amount of total RNAs (20 mg/lane) were electrophoresed in a 1% agarose gel containing 6% formaldehyde and transferred to a nylon membrane (Hybond N; Amersham). After u.v. crosslinking, the ®lter was prehybridized for 6 h at 428C in hybridization buer containing 50% deionized formamide, 0.65 M NaCl, 5 mM EDTA (pH 7.6), 0.1% sodium dodecyl sulfate (SDS), 0.1 M piperazine-N',N'-bis (2-ethanesulfonic acid) (pH 6.8), 56Denhardt's solution [0.1% Ficoll 400/ 0.1% PVP/0.1% BSA (fraction V)], and 100 mg/ml denatured salmon sperm DNA, and then hybridized with a probe for 12 h at 428C. As a probe, a 610 bp fragment of VEGF cDNA cloned from human whole brain mRNA was labeled with [a-32P] dCTP (3000 Ci/mmol), using a random

prime labeling kit (Amersham). The ®lter was washed at 578C with 26SSC/0.1% SDS for 1.5 h and exposed to an imaging plate. Radioactivity was then analysed with a Bioimage analyser BAS 2000 (Fuji Photo Film Co., Japan). To evaluate the amount of RNA analysed, the ®lter was rehybridized with a32P-labeled cDNA probe of 1.5 kb for human b-actin containing the entire coding region. mRNA stability assay The half-life of VEGF mRNA was determined by treating A-172 cells with actinomycin D. A-172 cells were cultured for 6 h with or without SNAP (0.5 mM), then actinomycin D was added to the culture medium (25 mg/ml) to block transcription and promptly returned to the same conditions. The cells were harvested at 0, 4, 8 and 16 h after addition of actinomycin D. Total RNA was prepared and Northern analysis was performed as described. The amount of VEGF mRNA and b-actin mRNA at each time point was quantify after Northern analysis with a Bio-image analyser BAS 2000. The amounts of VEGF mRNA was corrected for loading dierences by amount of the b-actin mRNA. Western blot analysis of VEGF protein A-172 cells were seeded onto a 15 cm petridish 16 h before SNAP treatment and added 75 ml of stock solution of SNAP in DMSO or DMSO alone. After incubation of A172 cells with SNAP, 5 ml of conditioned medium from the cell cultures was harvested and centrifuged to remove cell debris. Heparin-binding proteins were precipitated from the supernatant with 80 ml of heparin-Sepharose (1:1 slurry; Pharmacia) for 2 h at 48C. Heparin-Sepharose beads were precipitated by centrifugation and washed three times with washing buer containing 20 mM TrisHCl, pH 7.5, 300 mM NaCl, 200 mM PMSF, 1 mM DTT, and 1 mM leupeptin. Heparin-Sepharose-bound proteins were extracted by a 3-min incubation in Laemmli's sample buer at 1008C and separated on a 12% sodium dodecyl sulfate-polyacrylamide gel. After transfer to Immobilon membrane (Millipore), VEGF proteins were detected using polyclonal rabbit anti-human VEGF antibody (Santa Cruz Biotechnology) and the bands were visualized by ECL system (Amersham).

Acknowledgements This work was partly supported by a Grant from Ministry of Health and Welfare for the Second-term Comprehensive 10 year Strategy for Cancer Control.

References Adair TH, Gay WJ and Montani JP. (1990). Am. J. Physiol., 259, R393 ± R404. Beck I, Ramirez S, Weinmann R and Caro J. (1991). J. Biol. Chem., 266, 15563 ± 15566. Blanchard KL, Acquviva AM, Galson DL and Bunn HF. (1992). Mol. Cell. Biol., 12, 5373 ± 5385. Cohen T, Nahari D, Cerem LW, Neufeld G and Levi BZ. (1996). J. Biol. Chem., 271, 736 ± 741. Chomczynski P and Sacchi N. (1987). Anal. Biochem., 162, 156 ± 159. Dvorak HF, Brown LF, Detmar M and Dvorak AM. (1995). Am. J. Pathol., 146, 1029 ± 1039. Finkenzeller G, Marme M, Weich HA and Hug H. (1992). Cancer Res., 52, 4821 ± 4823. Firth JD, Ebert BL, Pugh CW and Ratclie PJ. (1994). Proc. Natl. Acad. Sci. USA, 91, 6496 ± 6500.

Folkman J and Klagsbrun M. (1987). Science, 235, 442 ± 447. Goldberg MA and Schneider TJ. (1994). J. Biol. Chem., 269, 4355 ± 4359. Harada S, Nagy JA, Sullivan KA, Thomas KA, Endo N, Rodan GA and Rodan SB. (1994). J. Clin. Invest., 93, 2490 ± 2496. Ikeda E, Achen MG, Breier G and Risau W. (1995). J. Biol. Chem., 270, 19761 ± 19766. Keck, PJ, Hauser SD, Krivi G, Sanzo K, Warren T, Feder J and Connolly DT. (1989). Science, 246, 1309 ± 1312. Konturek SJ, Brzozowski T, Majka J, Pytko-Polonczyk J and Stachura J. (1993). Eur. J. Pharmacol., 239, 215 ± 217. Krantz SB. (1991). Blood, 77, 419 ± 434. Ku DD, Zaleski JK, Liu S and Brock TA. (1993). Am. J. Physiol., 265, H586 ± H592.

441


442

Leung DW, Cachianes G, Kuang WJ, Goeddel DV and Ferrara N. (1989). Science, 246, 1306 ± 1309. Levy AP, Levy NS, Wegner S and Goldberg MA. (1995). J. Biol. Chem., 270, 13333 ± 13340. Li J, Perrella MA, Tsai JC, Yet SF, Hsieh CM, Yoshizumi M, Patterson C, Endege WO, Zhou F and Lee ME. (1995). J. Biol. Chem., 270, 308 ± 312. Madan A and Curtin PT. (1993). Proc. Natl. Acad. Sci. USA, 90, 3928 ± 3932. Melillo G, Musso T, Sica A, Taylor LS, Cox GW and Varesio L. (1995). J. Exp. Med., 182, 1683 ± 1693. Millauer B, Shawver LK, Plate KH, Risau W and Ullrich A. (1994). Nature, 367, 576 ± 579. Minchenko A, Bauer T, Salceda S and Caro J. (1994). Lab. Invest., 71, 374 ± 379. Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS and Sukhatme VP. (1995). Nature, 375, 577 ± 581. Ohigashi T, Brookins J and Fisher JW. (1993). J. Clin. Invest., 92, 1587 ± 1591. Palmer RM, Ferrige AG and Moncada S. (1987). Nature, 327, 524 ± 526.

Pertovaara L, Kaipainen A, Mustonen T, Orpana A, Ferrara N, Saksela O and Alitalo K. (1994). J. Biol. Chem., 269, 6271 ± 6274. Plate KH, Breier G, Weich HA and Risau W. (1992). Nature, 359, 845 ± 848. Pohl U and Busse R. (1989). Am. J. Physiol., 256 (Heart Circ. Physiol. 25), H1590 ± H1600. Risau W. (1990). Prog. Growth Factor Res., 2, 71 ± 79. Semenza GL and Wang GL. (1992). Mol. Cell. Biol., 12, 5447 ± 5454. Shweiki D, Itin A, Soer D and Keshet E. (1992). Nature, 359, 843 ± 845. Tuder RM, Flook BE and Voelkel NF. (1995). J. Clin. Invest., 95, 1798 ± 1807. Wang GL, Jiang BH, Rue EA and Semenza GL. (1995). Proc. Natl. Acad. Sci. USA, 92, 5510 ± 5514. Wang GL and Semenza GL. (1993). J. Biol. Chem., 268, 21513 ± 21518. Ziche M, Morbidelli L, Masini E, Granger H, Geppetti P and Ledda F. (1993). Biochem. Biophys. Res. Commun., 192, 1198 ± 1203.