- Email: [email protected]
Angiopoietin-1 induces sprouting angiogenesis in vitro
Brief Communication
529
Angiopoietin-1 induces sprouting angiogenesis in vitro Thomas I. Koblizek*, Cornelia Weiss*, George D. Yancopoulos†, Urban Deutsch* and Werner Risau* Sprouting of new capillaries from pre-existing blood vessels is a hallmark of angiogenesis during embryonic development and solid tumor growth [1]. In addition to the vascular endothelial growth factor (VEGF) and its receptors, the Tie receptors and their newly identified ligands, the angiopoietins, have been implicated in the control of blood vessel formation [2,3]. Although ‘knockouts’ of the gene encoding the Tie2 receptor, or its activating ligand angiopoietin-1 (Ang1), result in embryonic lethality in mice due to an absence of remodeling and sprouting of blood vessels [4,5], biological activity in vitro has not yet been described for this receptor–ligand system. In an assay in which a monolayer of endothelial cells were cultured on microcarrier beads and embedded in three-dimensional fibrin gels, recombinant Ang1 (0.5–10 nM) induced the formation of capillary sprouts in a dose-dependent manner that was completely inhibited by soluble Tie2 receptor extracellular domains. In contrast with VEGF, which also induced sprouting of capillaries, Ang1 was only very weakly mitogenic for endothelial cells. Suboptimal concentrations of VEGF and Ang1 acted synergistically to induce sprout formation. Thus, the biological activity of Ang1 in vitro is consistent with the specific phenotype of mice deficient in Tie2 or Ang1. The data suggest that, like in other developmental systems, blood vessel formation requires a hierarchy of master-control genes in which VEGF and angiopoietins, along with their receptors, are amongst the most important regulators. Addresses: *Department of Molecular Cell Biology, Max-PlanckInstitute for Physiological and Clinical Research, Kerckhoff Institute, Parkstrasse 1, 61231 Bad Nauheim, Germany. †Regeneron Pharmaceuticals, Inc., 777 Old Saw Mill River Road, Tarrytown, New York 10591-6707, USA. Correspondence: Werner Risau E-mail: [email protected] Received: 23 January 1998 Revised: 24 February 1998 Accepted: 11 March 1998 Published: 13 April 1998 Current Biology 1998, 8:529–532 http://biomednet.com/elecref/0960982200800529 © Current Biology Ltd ISSN 0960-9822
Results and discussion Previous attempts to detect biological activity of the newly identified angiopoietin ligands for the Tie2 receptors in
vitro have failed [5–7]. Because one of the most important phenotypes in Tie2-deficient mice is the absence of capillary sprouts in the neural tube [4], we have taken advantage of an in vitro sprouting assay [8] to investigate the biological activities of the angiopoietins. Adrenal-cortexderived microvascular endothelial (ACE) cells were cultivated on microcarrier (MC) beads until they had formed a confluent monolayer. These beads were then embedded into fibrin gels containing the various factors or controls. Initially, cell culture supernatants of a chinese hamster ovary (CHO) cell line stably transfected with an expression plasmid for Ang1, and, subsequently, purified recombinant Ang1* [7], were used to study the activity of Ang1 in vitro. Biological activity of recombinant Ang1 and Ang1* was confirmed by their high-affinity binding to, and stimulation of the tyrosine kinase activity of, the Tie2 receptor in vitro [5–7]. Ang1* is a variant of Ang1 that is used primarily because it is much easier to produce than Ang1, which cannot be made easily in quantities sufficient for experimental uses (S. Davis, T.H. Aldrich, N. Papadopoulos, T.J. Daly, J. Goldberg, V. Jain and G.D.Y., unpublished observations); in addition, Ang1* is also not susceptible to inhibition with Ang2. Ang1* consists of the first 73 residues of human Ang2 —beginning with the initiator methionine and ending with the sequence DAPLEY (in the single-letter amino-acid code) — fused to the portion of human Ang1 beginning at residue 77 with the sequence DFSSQ and extending to the carboxyl terminus. In addition, the cysteine corresponding to position 265 of human Ang1 is replaced by a serine. Ang1-containing supernatants, and also purified Ang1*, induced the formation of capillary sprouts, whereas supernatants of control cells did not (Figure 1). Endothelial cells migrated from the confluent monolayer and extended long filopodia. Long sprouts consisted of several cells forming a lumen as described before [9]. This process closely resembles sprout formation in vivo. Depletion of Ang1 from the CHO-conditioned medium by preadsorption to a Tie2–immunoglobulin G (IgG) column (data not shown), or incorporation of the Tie2–IgG into the fibrin gel, completely inhibited sprout formation. Endothelial cells lacking the Tie2 receptor (derived from Tie2-receptor-deficient mice) did not respond to Ang1 by sprouting, in contrast to wild-type endothelial cells (data not shown). Taken together, these results clearly demonstrated the specificity of Ang1 action. Quantification of sprout formation revealed a more than 10-fold increase in the number of sprouts with a length exceeding the diameter of the bead, which was used as an internal standard
530
Current Biology, Vol 8 No 9
Figure 1 Assay for sprout formation in vitro. ACE cells on MC beads in three-dimensional fibrin gels incubated with control (a,b) or Ang1containing (c,d) samples. (a,c) Phase-contrast photomicrographs; (b,d) nuclear staining with Hoechst 33324 showing the number of endothelial cells that make up individual sprouts in (c). The assay for angiogenesis in vitro was carried out as described by Nehls and Drenckhahn [8] with slight modifications. Briefly, ACE cells were grown to confluence on MC beads (Sigma) and placed in a 2.5 mg/ml fibrinogen gel (Sigma) containing Ang1, VEGF or controls, and 200 U/ml Trasylol (Bayer). Clotting was started by the addition of 2.5 U/ml thrombin (Sigma). Fibrin gels were incubated in DMEM containing the appropriate concentration of factors, 0.1% calf serum and 200 U/ml Trasylol. After two days, gels were fixed with 4% paraformaldehyde, and the number of capillary sprouts with length exceeding the diameter of the MC bead (150 µm) was determined for every 50 MC beads counted. The scale bar represents 150 µm.
(Figure 2). Figure 3 shows the dose response of sprout formation to purified recombinant Ang1*. Significant stimulation was seen at a concentration of 220 ng/ml (equivalent to 1.7 nM with Ang1 as a dimer). It is not clear why previous assays in vitro (such as tube-formation assays [5,6]) have failed to reveal any biological activity. Differences in the cells, cell confluency (which affects receptor
expression), assay conditions and/or extracellular matrix are of considerable importance in these in vitro assays and may determine the particular response. Apart from those differences, there is a principle difference between sprouting/invasion assays that require active migration and possibly proliferation, and tube-formation assays that Figure 3
35 30 25 20 15 10 5 0 Control
30 25 20 15 10 5 0
0
Ang1
20
74
220
670 2000 670 + Tie2
Ang1* concentration (ng/ml)
Current Biology Current Biology
Sprout formation assays were performed using media conditioned by control or Ang1-producing cell lines. In each case, the number of capillary sprouts per 50 MC beads was counted. As described in the legend to Figure 1, only those capillary sprouts with length exceeding the diameter of the MC bead were counted. Values are mean ± s.e.m. (n = 6).
p = 0.0002
40
p = 0.0438
Number of capillary sprouts per 50 MC beads
Number of capillary sprouts per 50 MC beads
Figure 2
1
10
VEGF concentration (ng/ml)
Dose-response of sprout formation in purified Ang1* or VEGF. Values are mean ± s.e.m. (n = 6). Significance was determined by the Student t test. Recombinant Tie2–IgG fusion protein (20 µg/ml) was used to inhibit Ang1* activity. Irrelevant IgG fusion proteins did not inhibit sprout formation in response to Ang1 (data not shown).
531
Brief Communication
Figure 4
Figure 5
30
20
p = 0.0063
p = 0.0045
10000
p = 0.0304
p = 0.0330
p = 0.0051 40
Number of cells per well
Number of capillary sprouts per 50 MC beads
15000 50
5000
10
0 0
VEGF (1 ng/ml )
Ang1* (220 ng/ml )
Expected additive effect Ang1 + VEGF
Experimental Ang1* + VEGF
0
7
20
74 220
74
0.2
1
5
+ Tie2
Ang1* concentration (ng/ml)
Current Biology
VEGF concentration (ng/ml) Current Biology
Synergistic effect of Ang1 and VEGF. The combined stimulation using the suboptimal doses (see Figure 3) of VEGF (1 ng/ml) and Ang1* (220 ng/ml) results in a larger increase in the number of sprouts than expected for a purely additive action. Values are mean ± s.e.m. (n = 6); significance was determined by the Student t test.
just require a reorganization of cells. Ang1 and Ang1* have repeatedly failed to show activity in tube-formation assays. Of course, we cannot exclude the possibility that some variation of the assay or some factor combination would reveal some activity. VEGF also induced sprout formation (Figure 3) that was not inhibited by Tie2–IgG (data not shown). When tested at suboptimal doses, VEGF (1 ng/ml) and Ang1 (220 ng/ml) acted synergistically to induce sprout formation (Figure 4). Unlike VEGF, Ang1 possesses only very weak mitogenic activity for endothelial cells (Figure 5). This suggests that the combined activity of these factors may induce synergistic signaling events in endothelial cells, leading to the formation of new capillaries by sprouting from pre-existing vessels. Signal transduction pathways that are involved in the sprouting of capillaries have not yet been identified. Filopodia formation triggered by small GTPases of the cdc42 family, migration and localized proteolytic activity provided by the plasmin or metalloprotease systems all seem to participate in this complex process. It will be important to determine the contribution of the VEGF and angiopoietin receptor systems in the regulation of these activities. In any case, the specific early embryonic lethalities in mice deficient for the Ang1 ligand and the Tie2 receptor have demonstrated the essential role of this ligand–receptor pair in sprouting angiogenesis (although less dramatic in the case of Ang1 than Tie2) and remodeling of pre-existing vascular plexus [4,5]. In addition, pericyte and smooth muscle cell recruitment has been
Ang1 is only a weak mitogen for endothelial cells. Third-passage human umbilical vein endothelial cells (5000 per well) were seeded into a 48-well plate (Nunc). The next day, medium containing 5% FCS and the indicated amounts of Ang1* and VEGF were added, and the cells were incubated for 3 days. Cells were counted using a Casy cell counter (Schärfe System). Ang1* induced a 30% increase in cell number, which was inhibited by Tie2–IgG (20 µg/ml), whereas VEGF induced a more than twofold increase.
proposed to be regulated by this ligand–receptor system [3,10]. Whether the remodeling and/or recruitment processes involve similar or different activities as the sprouting process, or require the action of additional ligands such as Ang2 or receptors such as Tie1, remains to be determined. Our results show that Ang1 is a potent inducer of sprouting angiogenesis. Inhibition of this activity may have implications for the therapeutic intervention of pathological angiogenesis, such as in solid tumors.
Acknowledgements We are indebted to Volker Nehls (Universität Würzburg, Germany) for showing us the sprouting assay and for helpful discussions. We thank Silvia Hennig and Carmen Fangmann for skillful technical assistance; Herbert Weich for the gift of purified recombinant VEGF; Peter Maisonpierre, Samuel Davis and the Protein Sciences Division of Regeneron, particularly Terence Ryan, Thomas Daly and Nick Papadopoulos, for generation of CHO lines producing Ang1, and for recombinant Ang1* and Tie2–Fc. This study was supported in part by the Deutsche Krebshilfe to W.R.
References 1. Risau W: Mechanisms of angiogenesis. Nature 1997, 386:671674. 2. Breier G, Risau W: The role of vascular endothelial growth-factor in blood-vessel formation. Trends Cell Biol 1996, 6:454-456. 3. Hanahan D: Signaling vascular morphogenesis and maintenance. Science 1997, 277:48-50. 4. Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, Fujiwara Y, Gendron-Maguire M, et al.: Distinct roles of the receptor tyrosine kinases tie-1 and tie-2 in blood-vessel formation. Nature 1995, 376:70-74. 5. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, Davis S, et al.: Requisite role of angiopoietin-1, a ligand for the tie2 receptor, during embryonic angiogenesis. Cell 1996, 87:1171-1180.
532
Current Biology, Vol 8 No 9
6. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, et al.: Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell 1996, 87:1161-1169. 7. Maisonpierre PC, Suri C, Jones PF, Bartunkova S, Wiegand SJ, Radziejewski C, et al.: Angiopoietin-2, a natural antagonist for tie2 that disrupts in vivo angiogenesis. Science 1997, 277:55-60. 8. Nehls V, Drenckhahn D: A microcarrier-based cocultivation system for the investigation of factors and cells involved in angiogenesis in 3-dimensional fibrin matrices in vitro. Histochem Cell Biol 1995, 104:459-466. 9. Nehls V, Drenckhahn D: A novel, microcarrier-based in vitro assay for rapid and reliable quantification of 3-dimensional cellmigration and angiogenesis. Microvasc Res 1995, 50:311-322. 10. Folkman J, D'Amore PA: Blood-vessel formation — what is its molecular basis. Cell 1996, 87:1153-1155.
529
Angiopoietin-1 induces sprouting angiogenesis in vitro Thomas I. Koblizek*, Cornelia Weiss*, George D. Yancopoulos†, Urban Deutsch* and Werner Risau* Sprouting of new capillaries from pre-existing blood vessels is a hallmark of angiogenesis during embryonic development and solid tumor growth [1]. In addition to the vascular endothelial growth factor (VEGF) and its receptors, the Tie receptors and their newly identified ligands, the angiopoietins, have been implicated in the control of blood vessel formation [2,3]. Although ‘knockouts’ of the gene encoding the Tie2 receptor, or its activating ligand angiopoietin-1 (Ang1), result in embryonic lethality in mice due to an absence of remodeling and sprouting of blood vessels [4,5], biological activity in vitro has not yet been described for this receptor–ligand system. In an assay in which a monolayer of endothelial cells were cultured on microcarrier beads and embedded in three-dimensional fibrin gels, recombinant Ang1 (0.5–10 nM) induced the formation of capillary sprouts in a dose-dependent manner that was completely inhibited by soluble Tie2 receptor extracellular domains. In contrast with VEGF, which also induced sprouting of capillaries, Ang1 was only very weakly mitogenic for endothelial cells. Suboptimal concentrations of VEGF and Ang1 acted synergistically to induce sprout formation. Thus, the biological activity of Ang1 in vitro is consistent with the specific phenotype of mice deficient in Tie2 or Ang1. The data suggest that, like in other developmental systems, blood vessel formation requires a hierarchy of master-control genes in which VEGF and angiopoietins, along with their receptors, are amongst the most important regulators. Addresses: *Department of Molecular Cell Biology, Max-PlanckInstitute for Physiological and Clinical Research, Kerckhoff Institute, Parkstrasse 1, 61231 Bad Nauheim, Germany. †Regeneron Pharmaceuticals, Inc., 777 Old Saw Mill River Road, Tarrytown, New York 10591-6707, USA. Correspondence: Werner Risau E-mail: [email protected] Received: 23 January 1998 Revised: 24 February 1998 Accepted: 11 March 1998 Published: 13 April 1998 Current Biology 1998, 8:529–532 http://biomednet.com/elecref/0960982200800529 © Current Biology Ltd ISSN 0960-9822
Results and discussion Previous attempts to detect biological activity of the newly identified angiopoietin ligands for the Tie2 receptors in
vitro have failed [5–7]. Because one of the most important phenotypes in Tie2-deficient mice is the absence of capillary sprouts in the neural tube [4], we have taken advantage of an in vitro sprouting assay [8] to investigate the biological activities of the angiopoietins. Adrenal-cortexderived microvascular endothelial (ACE) cells were cultivated on microcarrier (MC) beads until they had formed a confluent monolayer. These beads were then embedded into fibrin gels containing the various factors or controls. Initially, cell culture supernatants of a chinese hamster ovary (CHO) cell line stably transfected with an expression plasmid for Ang1, and, subsequently, purified recombinant Ang1* [7], were used to study the activity of Ang1 in vitro. Biological activity of recombinant Ang1 and Ang1* was confirmed by their high-affinity binding to, and stimulation of the tyrosine kinase activity of, the Tie2 receptor in vitro [5–7]. Ang1* is a variant of Ang1 that is used primarily because it is much easier to produce than Ang1, which cannot be made easily in quantities sufficient for experimental uses (S. Davis, T.H. Aldrich, N. Papadopoulos, T.J. Daly, J. Goldberg, V. Jain and G.D.Y., unpublished observations); in addition, Ang1* is also not susceptible to inhibition with Ang2. Ang1* consists of the first 73 residues of human Ang2 —beginning with the initiator methionine and ending with the sequence DAPLEY (in the single-letter amino-acid code) — fused to the portion of human Ang1 beginning at residue 77 with the sequence DFSSQ and extending to the carboxyl terminus. In addition, the cysteine corresponding to position 265 of human Ang1 is replaced by a serine. Ang1-containing supernatants, and also purified Ang1*, induced the formation of capillary sprouts, whereas supernatants of control cells did not (Figure 1). Endothelial cells migrated from the confluent monolayer and extended long filopodia. Long sprouts consisted of several cells forming a lumen as described before [9]. This process closely resembles sprout formation in vivo. Depletion of Ang1 from the CHO-conditioned medium by preadsorption to a Tie2–immunoglobulin G (IgG) column (data not shown), or incorporation of the Tie2–IgG into the fibrin gel, completely inhibited sprout formation. Endothelial cells lacking the Tie2 receptor (derived from Tie2-receptor-deficient mice) did not respond to Ang1 by sprouting, in contrast to wild-type endothelial cells (data not shown). Taken together, these results clearly demonstrated the specificity of Ang1 action. Quantification of sprout formation revealed a more than 10-fold increase in the number of sprouts with a length exceeding the diameter of the bead, which was used as an internal standard
530
Current Biology, Vol 8 No 9
Figure 1 Assay for sprout formation in vitro. ACE cells on MC beads in three-dimensional fibrin gels incubated with control (a,b) or Ang1containing (c,d) samples. (a,c) Phase-contrast photomicrographs; (b,d) nuclear staining with Hoechst 33324 showing the number of endothelial cells that make up individual sprouts in (c). The assay for angiogenesis in vitro was carried out as described by Nehls and Drenckhahn [8] with slight modifications. Briefly, ACE cells were grown to confluence on MC beads (Sigma) and placed in a 2.5 mg/ml fibrinogen gel (Sigma) containing Ang1, VEGF or controls, and 200 U/ml Trasylol (Bayer). Clotting was started by the addition of 2.5 U/ml thrombin (Sigma). Fibrin gels were incubated in DMEM containing the appropriate concentration of factors, 0.1% calf serum and 200 U/ml Trasylol. After two days, gels were fixed with 4% paraformaldehyde, and the number of capillary sprouts with length exceeding the diameter of the MC bead (150 µm) was determined for every 50 MC beads counted. The scale bar represents 150 µm.
(Figure 2). Figure 3 shows the dose response of sprout formation to purified recombinant Ang1*. Significant stimulation was seen at a concentration of 220 ng/ml (equivalent to 1.7 nM with Ang1 as a dimer). It is not clear why previous assays in vitro (such as tube-formation assays [5,6]) have failed to reveal any biological activity. Differences in the cells, cell confluency (which affects receptor
expression), assay conditions and/or extracellular matrix are of considerable importance in these in vitro assays and may determine the particular response. Apart from those differences, there is a principle difference between sprouting/invasion assays that require active migration and possibly proliferation, and tube-formation assays that Figure 3
35 30 25 20 15 10 5 0 Control
30 25 20 15 10 5 0
0
Ang1
20
74
220
670 2000 670 + Tie2
Ang1* concentration (ng/ml)
Current Biology Current Biology
Sprout formation assays were performed using media conditioned by control or Ang1-producing cell lines. In each case, the number of capillary sprouts per 50 MC beads was counted. As described in the legend to Figure 1, only those capillary sprouts with length exceeding the diameter of the MC bead were counted. Values are mean ± s.e.m. (n = 6).
p = 0.0002
40
p = 0.0438
Number of capillary sprouts per 50 MC beads
Number of capillary sprouts per 50 MC beads
Figure 2
1
10
VEGF concentration (ng/ml)
Dose-response of sprout formation in purified Ang1* or VEGF. Values are mean ± s.e.m. (n = 6). Significance was determined by the Student t test. Recombinant Tie2–IgG fusion protein (20 µg/ml) was used to inhibit Ang1* activity. Irrelevant IgG fusion proteins did not inhibit sprout formation in response to Ang1 (data not shown).
531
Brief Communication
Figure 4
Figure 5
30
20
p = 0.0063
p = 0.0045
10000
p = 0.0304
p = 0.0330
p = 0.0051 40
Number of cells per well
Number of capillary sprouts per 50 MC beads
15000 50
5000
10
0 0
VEGF (1 ng/ml )
Ang1* (220 ng/ml )
Expected additive effect Ang1 + VEGF
Experimental Ang1* + VEGF
0
7
20
74 220
74
0.2
1
5
+ Tie2
Ang1* concentration (ng/ml)
Current Biology
VEGF concentration (ng/ml) Current Biology
Synergistic effect of Ang1 and VEGF. The combined stimulation using the suboptimal doses (see Figure 3) of VEGF (1 ng/ml) and Ang1* (220 ng/ml) results in a larger increase in the number of sprouts than expected for a purely additive action. Values are mean ± s.e.m. (n = 6); significance was determined by the Student t test.
just require a reorganization of cells. Ang1 and Ang1* have repeatedly failed to show activity in tube-formation assays. Of course, we cannot exclude the possibility that some variation of the assay or some factor combination would reveal some activity. VEGF also induced sprout formation (Figure 3) that was not inhibited by Tie2–IgG (data not shown). When tested at suboptimal doses, VEGF (1 ng/ml) and Ang1 (220 ng/ml) acted synergistically to induce sprout formation (Figure 4). Unlike VEGF, Ang1 possesses only very weak mitogenic activity for endothelial cells (Figure 5). This suggests that the combined activity of these factors may induce synergistic signaling events in endothelial cells, leading to the formation of new capillaries by sprouting from pre-existing vessels. Signal transduction pathways that are involved in the sprouting of capillaries have not yet been identified. Filopodia formation triggered by small GTPases of the cdc42 family, migration and localized proteolytic activity provided by the plasmin or metalloprotease systems all seem to participate in this complex process. It will be important to determine the contribution of the VEGF and angiopoietin receptor systems in the regulation of these activities. In any case, the specific early embryonic lethalities in mice deficient for the Ang1 ligand and the Tie2 receptor have demonstrated the essential role of this ligand–receptor pair in sprouting angiogenesis (although less dramatic in the case of Ang1 than Tie2) and remodeling of pre-existing vascular plexus [4,5]. In addition, pericyte and smooth muscle cell recruitment has been
Ang1 is only a weak mitogen for endothelial cells. Third-passage human umbilical vein endothelial cells (5000 per well) were seeded into a 48-well plate (Nunc). The next day, medium containing 5% FCS and the indicated amounts of Ang1* and VEGF were added, and the cells were incubated for 3 days. Cells were counted using a Casy cell counter (Schärfe System). Ang1* induced a 30% increase in cell number, which was inhibited by Tie2–IgG (20 µg/ml), whereas VEGF induced a more than twofold increase.
proposed to be regulated by this ligand–receptor system [3,10]. Whether the remodeling and/or recruitment processes involve similar or different activities as the sprouting process, or require the action of additional ligands such as Ang2 or receptors such as Tie1, remains to be determined. Our results show that Ang1 is a potent inducer of sprouting angiogenesis. Inhibition of this activity may have implications for the therapeutic intervention of pathological angiogenesis, such as in solid tumors.
Acknowledgements We are indebted to Volker Nehls (Universität Würzburg, Germany) for showing us the sprouting assay and for helpful discussions. We thank Silvia Hennig and Carmen Fangmann for skillful technical assistance; Herbert Weich for the gift of purified recombinant VEGF; Peter Maisonpierre, Samuel Davis and the Protein Sciences Division of Regeneron, particularly Terence Ryan, Thomas Daly and Nick Papadopoulos, for generation of CHO lines producing Ang1, and for recombinant Ang1* and Tie2–Fc. This study was supported in part by the Deutsche Krebshilfe to W.R.
References 1. Risau W: Mechanisms of angiogenesis. Nature 1997, 386:671674. 2. Breier G, Risau W: The role of vascular endothelial growth-factor in blood-vessel formation. Trends Cell Biol 1996, 6:454-456. 3. Hanahan D: Signaling vascular morphogenesis and maintenance. Science 1997, 277:48-50. 4. Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, Fujiwara Y, Gendron-Maguire M, et al.: Distinct roles of the receptor tyrosine kinases tie-1 and tie-2 in blood-vessel formation. Nature 1995, 376:70-74. 5. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, Davis S, et al.: Requisite role of angiopoietin-1, a ligand for the tie2 receptor, during embryonic angiogenesis. Cell 1996, 87:1171-1180.
532
Current Biology, Vol 8 No 9
6. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, et al.: Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell 1996, 87:1161-1169. 7. Maisonpierre PC, Suri C, Jones PF, Bartunkova S, Wiegand SJ, Radziejewski C, et al.: Angiopoietin-2, a natural antagonist for tie2 that disrupts in vivo angiogenesis. Science 1997, 277:55-60. 8. Nehls V, Drenckhahn D: A microcarrier-based cocultivation system for the investigation of factors and cells involved in angiogenesis in 3-dimensional fibrin matrices in vitro. Histochem Cell Biol 1995, 104:459-466. 9. Nehls V, Drenckhahn D: A novel, microcarrier-based in vitro assay for rapid and reliable quantification of 3-dimensional cellmigration and angiogenesis. Microvasc Res 1995, 50:311-322. 10. Folkman J, D'Amore PA: Blood-vessel formation — what is its molecular basis. Cell 1996, 87:1153-1155.