c-Jun N-terminal kinase activation is required for the inhibition of neovascularization by thrombospondin-1


Thrombospondin-1 (TSP-1) is a potent inhibitor of angiogenesis that acts directly on endothelial cells via the CD36 surface receptor molecule to halt their migration, proliferation, and morphogenesis in vitro and to block neovascularization in vivo. Here we show that inhibitory signals elicited by TSP-1 did not alter the ability of inducers of angiogenesis to activate p42 and p44 mitogen-activated protein kinase (MAPK). Rather, TSP-1 induced a rapid and transient activation of c-Jun N-terminal kinases (JNK). JNK activation by TSP-1 required engagement of CD36, as it was blocked by antagonistic CD36 antibodies and stimulated by short anti-angiogenic peptides derived from TSP-1 that act exclusively via CD36. TSP-1 inhibition of corneal neovascularization induced by bFGF was severely impaired in mice null for JNK-1, pointing to a critical role for this stress-activated kinase in the inhibition of neovascularization by TSP-1.


Angiogenesis, the formation of new blood vessels from pre-existing ones, is a tightly regulated process controlled by the balance between positive and negative stimuli in the environment of endothelial cells (Bouck et al., 1996; Hanahan and Folkman, 1996). Identification of the molecules that regulate this complex process has been the first step towards understanding the mechanisms controlling angiogenesis (Carmeliet, 2000). Additional insight comes from the discovery that a large proportion of known inhibitors of angiogenesis, including thrombospondin-1, angiostatin, endostatin, 2-methoxyestradiol and canstatin, have the ability to induce apoptosis in endothelial cells thereby precluding the cells from responding to a wide variety of pro-angiogenic factors that signal through different pathways (Yue et al., 1997; Yeh et al., 1998; Claesson-Welsh et al., 1998; Lucas et al., 1998; Dhanabal et al., 1999; Jiménez et al., 2000; Kamphaus et al., 2000). In their turn, inducers of angiogenesis, besides being able to activate endothelial cell proliferation, migration and capillary morphogenesis, promote endothelial cell survival (Tran et al., 1999; Fujio and Walsh, 1999; Nor et al., 1999; O'Connor et al., 2000). Thus the balance between inducers and the apoptosis-inducing inhibitors of angiogenesis seems to be interpreted within the endothelial cell as a balance between survival and apoptosis pathways and to be critical for the regulation of angiogenesis (Jiménez et al., 2000). Defining the signaling pathways within the cell that are triggered by the molecules that regulate angiogenesis will be essential if these mechanisms are to be understood on a molecular level.

We have recently shown that the anti-angiogenic activity of TSP-1 requires the sequential activation of CD36, p59fyn, caspase-3-like proteases and p38 mitogen-activated protein kinases (Jiménez et al., 2000). Here we have investigated the contribution of c-Jun N-terminal kinase-1 (JNK-1) and p42/p44 mitogen-activated protein kinases (MAPKs) to the inhibition of neovascularization by TSP-1.

Results and discussion

Several inducers of angiogenesis, including basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), platelet derived growth factor (PDGF) and tumor necrosis factor α (TNFα), are known to rely on activation of p42 and p44 MAPK in order to transduce their stimulatory effects within the endothelial cell (D'Angelo et al., 1995; Modur et al., 1996; Pedram et al., 1998). The activation of p42 and p44 by bFGF (Figure 1a), VEGF, and PDGF (data not shown) in capillary endothelial cells was sustained for over 60 min, far exceeding the rapid and transient activation of these kinases that is commonly associated with the induction of proliferative responses (Marshall, 1995). This prolonged activation suggests that these kinases may also contribute to other essential components of the complex angiogenic response of endothelial cells, including migration and survival.

Figure 1

MAPK was activated by bFGF and required for induction of angiogenesis in vitro but not inhibited by TSP-1. (a) Kinetics of MAPK activation by bFGF was measured in bovine adrenal capillary endothelial cells BP10T8 (BCECs) (a gift of Dr J Folkman. Children's Hospital, Harvard Medical School, Boston, MA, USA). Cells were treated with 10 ng/ml bFGF for the indicated periods of time. Extracts were assayed for MAPK by assessing the shift to a higher mobility of the activated enzyme using 20 cm 10% SDS–PAGE gels with 0.2% SDS, a concentration doubled compared to the standard one, and acrylamide/bis-acrylamide ratio 35:0.6 in order to improve resolution of the non-phosphorylated and phosphorylated forms of the MAPKs. (b) Quiescent BCECs were stimulated 10 min with 10 ng/ml bFGF (lanes 5–9) or 20% FCS (lanes 10–14). Cells were pretreated 30 min with 1 μM PD98059 (lanes 6 and 11), 10 μM PD98059 (lanes 7 and 12), 50 μM PD98059 (lanes 8 and 13), 100 μM PD98059 (lanes 9 and 14), or the corresponding vehicule for 10 μM PD98059 (lane 2), 50 μM PD98059 (lane 3) or 100 μM PD98059 (lane 4). Unstimulated quiescent cells (lane 1). MAPK activity was assessed by immunocomplex kinase assay using myelin basic protein (MBP) as a substrate. (c) Endothelial cells were pre-treated for 2 h with 50 μM PD98059 (Calbiochem) and ability of the cells to migrate towards a variety of inducers tested. Background migration of endothelial cells in the presence of 0.1% BSA is indicated (BSA). Angiogenic factors used included from Sigma (St. Louis, MO, USA) β1-estradiol (E2) used at 1 μM, platelet activating factor (PAF) used at 10 μM, prostaglandin E1 (PGE1) used at 250 pg/ml and lysophosphatidic acid (LYS) used at 10 ng/ml; from R&D Systems (Minneapolis, MN, USA) basic fibroblast growth factor (bFGF) used at 10 ng/ml, transforming growth factor beta (TGFβ1) used at 15 pg/ml, scatter factor (SF) used at 5 ng/ml; and from Genzyme Diagnostics (Cambridge, MA, USA), vascular endothelial cell growth factor (VEGF) used at 100 pg/ml. bFGF indicates the migration towards 10 ng/ml of bFGF. Standard errors are shown. * Indicates instances where the drug caused significant inhibition, P<0.001. Endothelial cell migration assay was performed in an inverted modified Boyden Chamber as in Volpert et al., 1998. All samples were tested in quadruplicate. (d) and (e) Quiescent BCECs were preincubated for 3 h with the indicated concentrations of TSP-1 before a 10 min stimulation of the cells with 10 ng/ml bFGF in the absence or presence of the indicated concentrations of TSP-1. TSP-1 was purified from human platelets (Bornstein et al., 1994). MAPK activation was assessed by EMSA (d) or immunocomplex kinase assay using myelin basic protein (MBP) as a substrate (e). Immunocomplex kinase assays were done as described in (Jiménez et al., 2000) using 1 μg/ml anti-MAPK (Santa Cruz) and 5 μg of MBP (Sigma) for each individual kinase assay

The activation of MAPK was indeed critical for these growth factors to induce endothelial cell migration. PD98059, a peptide inhibitor that blocks MEK (Dudley et al., 1995) effectively inhibited MAPK activation in endothelial cells (Figure 1b). Pretreatment with PD98059 was sufficient to block endothelial cell migration towards multiple inducers of angiogenesis including VEGF, PDGF and bFGF (Figure 1c). As might be expected, PD98056 did not block migration towards TGFβ, which signals through a MAPK- independent pathway (Wrana, 2000).

The inhibitory effect of TSP-1 was not due to interference with the prolonged, inducer-dependent activation of MAPK. TSP-1 was not able to block p42/p44 MAPK activation by bFGF, nor was it able to independently activate MAPKs as measured both by mobility shift assay (EMSA) (Figure 1d) and by the more sensitive and quantitative immunocomplex kinase assay (Figure 1e). A similar lack of interference was seen when MAPK was activated by either VEGF or PDGF instead of bFGF (data not shown). This observation is in keeping with our previous results where we have demonstrated that the ability of TSP-1 to induce apoptosis in endothelial cells is also independent of either p42 or p44 MAPKs (Jiménez et al., 2000).

In contrast to the non-essential MAPKs, the stress-activated kinases are crucial to the inhibitory activity of TSP-1. We have shown previously that p38 MAPK is activated by TSP-1 and is essential for TSP-1 anti-angiogenic activity both in vitro and in vivo (Jiménez et al., 2000). Here we demonstrate that TSP-1 can also activate a second stress activated kinase, c-Jun N-terminal kinase (JNK). When used at doses that are inhibitory in the endothelial cell migration assay (Tolsma et al., 1993), TSP-1 stimulated JNK with a rapid kinetics typical for an early step in the signaling cascade (Figure 2a).

Figure 2

TSP-1 activated JNK in a CD36-dependent manner. (a) Human dermal microvascular endothelial cells (HMVECs) were treated for the indicated times and concentrations with TSP-1 and JNK immunocomplexes were obtained as indicated in (Jiménez et al., 2000) using 1 μg/ml anti-JNK-1 antibodies (Pharmingen), and assayed for ability to phosphorylate 1 μg of GST-c-jun (1–79) as substrate. TNFα was used as a positive control. (bd) HMVECs were pre-incubated, where indicated, for 1 h with 10 μg/ml anti-CD36 antibody FA6-152 (Immunotech) and then treated with TSP-1 (times and concentrations as indicated) (b); 10 ng/ml TNFα (c) or with anti-angiogenic peptides Col overlap or Mal III at 50 μM (d) (peptides synthesized, purified and dialyzed as described in Tolsma et al., 1993). Subsequently JNK immunocomplexes were obtained and assayed for ability to phosphorylate GST-c-jun (1–79)

TSP is known to bind a number of cellular receptors (Roberts, 1996) but it is CD36 that mediates its anti-angiogenic effects both in vitro (Dawson et al., 1997) and in vivo (Jiménez et al., 2000). To link JNK activation more tightly to the anti-angiogenic activity of TSP-1, its dependence on CD36 was tested. JNK activation by TSP-1 was blocked in the presence of FA6 (Figure 2b), an antibody that interferes with TSP-1 access to CD36 and prevents its acting as an anti-angiogenic agent (Dawson et al., 1997). The FA6 inhibitory effect was specific since TNFα activation of JNK was unaffected by the presence of these antibodies (Figure 2c). Additionally, the anti-CD36 monoclonal antibody SMΦ that is agonistic for CD36 and able to mimic TSP-1 activity on endothelial cells (Dawson et al., 1997), also caused activation of JNK in microvascular endothelial cells (data not shown).

The anti-angiogenic activity of TSP-1 resides in the 50 kDa central stalk of the molecule (Tolsma et al., 1993), where two independent anti-angiogenic sub-regions have been identified. From these regions two peptides were derived, Mal III and Col overlap, that share a common motif and act directly on endothelial cells via the CD36 receptors to mimic the anti-angiogenic properties of the whole TSP-1 molecule (Tolsma et al., 1993; Dawson et al., 1997). Both Mal III and Col overlap were able to activate JNK to the same extent as TSP-1 (Figure 2d). For both peptides this activation was effectively blocked by FA6 monoclonal antibodies, supporting the essential role of CD36 signaling for JNK activation by Mal III and Col overlap, and therefore by TSP-1.

To further link JNK activation by TSP-1 to the inhibition of angiogenesis a corneal neovascularization assay was performed on mice null for JNK-1 (JNK-1−/−). Null mice (Sabapathy et al., 1999b) were verified by Western blots of the whole cell extracts of embryonic fibroblasts derived from E 13.3 embryos (Figure 3a, upper panel). As a loading control the membrane was re-probed with an antibody that recognizes both isoforms of JNK-1 and JNK-2 (Figure 3a, lower panel). Null animals were bred to wild-type C56/Bl6J, the F1 backcrossed to the nulls, and the null F2 mice identified by their lack of the 46 kDa isomer of JNK-1. Progeny of these animals were used for cornea assays. JNK-1−/− mice showed normal response to inducers of angiogenesis, while their response to the inhibitory action of TSP-1 in the corneal neovascularization assay was substantially impaired (Figure 3 and Table 1). This result was further confirmed in an in vitro assay where endothelial cells were induced to migrate out of corneal explants in the presence of VEGF. For this assay, corneas of JNK-1−/− and C57Bl6 control mice were excised and placed in Matrigel depleted of most growth factors. Three days later active cell migration from the explants and capillary morphogenesis was found in the presence of VEGF (Figure 4). Migrating cells stained positively for the endothelial marker CD31 (T Ferguson, Washington University, St. Louis, MD, USA, personal communication). TSP-1 was able to halt the induction of cord formation by this corneal endothelium when it was derived from wild-type animals, but it was ineffective when applied to the corneas from JNK-1 −/− mice (Figure 4).

Figure 3

Inhibition of cornea neovascularization by TSP-1 was severely impaired in JNK-1 null mice. (a) Western blot analysis of JNK-1 isoforms in whole cell extracts of embryonic fibroblasts derived from wild-type (1, 2) and JNK-1−/− (3, 4) E13.3 embryos as in Loo and Cotman, 1994. Western blotting was performed using 100 μg of whole cell lysates obtained as in Hibi et al., 1993. The membrane was first probed with a monoclonal antibody (mAb) to JNK-1 (Pharmingen) (upper panel) and was reprobed with a mAb against JNK-1 and JNK-2 (Pharmingen) for a loading control (lower panel). (b) The cornea assay was performed as previously described (Volpert et al., 1998). Sucralfate/Hydron pellets were formulated with indicated test substances and implanted into the avascular cornea of anesthetized wild-type (C57/Bl6 from Jackson Labs) or jnk-1 null mice 0.5–1.0 mm from the vascular limbus. Where indicated the pellets contained 50 ng bFGF alone or in combination with purified human TSP-1 (200 ng/pellet). Photos were taken on day 5 post implantation and the results are presented schematically as a diagram

Table 1 Diminution of TSP-1 effect in mice lacking JNK-1
Figure 4

Defective inhibition of morphogenesis by TSP-1 in corneal endothelial cells from mice null for JNK-1. The assay was developed by Dr P Stuart at Washington University, St. Louis, MO, USA (PM Stuart and TA Ferguson, personal communications). The cornea of JNK-1−/− mice and their wild-type background counterparts were removed and the central area excised so that no limbal vessels with accessory vascular smooth muscle cells would be incorporated in the assay. The resulting fragments were placed in 48-well tissue culture plates into polymerized growth factor depleted Matrigel (Beckton-Dickinson) diluted 1 : 1 with endothelial cell basal medium (EBM). The explants were refed daily with EBM supplemented where indicated with VEGF (200 pg/ml) or VEGF and TSP-1 (5 nM). Tissue explants were observed daily and photographed on day 4 of experiment where the difference in response became evident

These results demonstrate that JNK-1 is required for the anti-angiogenic effect of TSP-1. Studies of mice that lack distinct members of the JNK family, JNK-1, JNK-2, JNK-3 reveal that although these kinases belong to a common family, they can play highly specific roles in T cell differentiation (Dong et al., 1998; Yang et al., 1998; Chu et al., 1999 and Sabapathy et al., 1999a) and apoptosis (Kuan et al., 1999). Data presented in this work support that JNK-1 requirement for TSP-1 anti-angiogenic action cannot be adequately compensated for by other MAPK family members. This suggests that independent targets exist in endothelial cells for each of the JNK kinases, and that JNK-1 plays a critical role in the inhibition of neovascularization by TSP-1. Thus two stress-activated kinases, JNK-1 and p38 MAPK seem to be integral parts of the signaling network that leads from CD36 to the apoptosis-dependent inhibition of angiogenesis by TSP-1.


  1. Bornstein P, Sage EH . 1994 Methods Enzymol. 245: 62–84

  2. Bouck N, Stellmach V, Hsu S . 1996 Adv. Cancer Res. 69: 135–174

  3. Carmeliet P . 2000 Nat. Med. 6: 389-395

  4. Chu WM, Ostertag D, Li ZW, Chang L, Chen Y, Hu Y, Williams B, Perrault J, Karin M . 1999 Immunity 11: 721–31

  5. Claesson-Welsh L, Welsh M, Ito N, Anand-Apte B, Soker S, Zetter B, O'Reilly MS, Folkman J . 1998 Proc. Natl. Acad. Sci. USA 95: 5579–5583

  6. D'Angelo G, Struman I, Maltral J, Weiner R . 1995 Proc. Natl. Acad. Sci. USA 92: 6374–6378

  7. Dhanabal M, Ramchandran R, Waterman MJ, Knebelmann B, Segal M, Sukhatme VP . 1999 J. Biol. Chem. 274: 11721–11726

  8. Dawson DW, Pearce SF, Zhong R, Silverstein RL, Frazier WA, Bouck NP . 1997 J. Cell Biol. 138: 707–717

  9. Dong C, Yang DD, Wysk M, Whitmarsh AJ, Davis RJ, Flavell RA . 1998 Science 282: 2093–2095

  10. Dudley DT, Pang L. Decker SJ, Bridges AJ, Saltiel AR . 1995 Proc. Natl. Acad. Sci. USA 92: 7686–7689

  11. Fujio Y, Walsh K . 1999 J. Biol. Chem. 274: 16349–16354

  12. Hanahan D, Folkman J . 1996 Cell 86: 353–364

  13. Hibi M, Lin A, Smeal T, Minden A, Karin A . 1993 Genes Dev. 7: 2135–2148

  14. Jiménez B, Volpert OV, Crawford SE, Febbraio M, Silverstein RL, Bouck N . 2000 Nat. Med. 6: 41–48

  15. Kamphaus GD, Colorado PC, Panka DJ, Hopfer H, Ramchandran R, Torre A, Maeshima Y, Mier JW, Sukkatma VP, Kalluri R . 2000 J. Biol. Chem. 275: 1209–1215

  16. Kuan CY, Yang DD, Samanta-Roy DR, Davis RJ, Rakia P, Flavell RA . 1999 Neuron 22: 667–676

  17. Loo D, Cotman C . 1994 Cell Biology: a laboratory handbook. Celis JE (ed). San Diego, New York, Boston, London, Sydney, Tokyo, Toronto pp. 45–53

  18. Lucas R, Holmgren L, Garcia I, Jimenez B, Mandriota SJ, Borlat F, Sim BK, Wu Z, Grau GE, Shing Y, Soff GA, Bouck N, Pepper MS . 1998 Blood 92: 4730–4741

  19. Marshall CJ . 1995 Cell 80: 179–185

  20. Modur V, Zimmerman GA, Precott JM, McIntyre TM . 1996 J. Biol. Chem. 271: 13094–13102

  21. Nor JE, Christensen J, Mooney DJ, Polverini PJ . 1999 Am. J. Pathol. 154: 375–384

  22. O'Connor DS, Schechner JS, Adida C, Mesri M, Rothermel AL, Li F, Nath AK, Pober JS, Altieri DC . 2000 Am. J. Pathol. 156: 393–398

  23. Pedram A, Razandi M, Levin ER . 1998 J. Biol. Chem. 273: 26722–26728

  24. Roberts DD . 1996 FASEB J. 10: 1183–1191

  25. Sabapathy K, Hu Y, Kallunki T, Schreiber M, David JP, Jochum W, Wagner EF, Karin M . 1999a Curr. Biol. 9: 116–125

  26. Sabapathy K, Jochum W, Hochedlinger K, Chang L, Karin M, Wagner EF . 1999b Mech. Dev. 89: 1115–1124

  27. Tolsma SS, Volpert OV, Good DJ, Frazier WA, Polverini PJ, Bouck N . 1993 J. Cell Biol. 122: 497–511

  28. Tran J, Rak J, Sheehan C, Saibil SD, LaCasse E, Korneluk RG, Kerbel RS . 1999 Biochem. Biophys. Res. Commun. 264: 781–788

  29. Volpert OV, Lawler J, Bouck NP . 1998 Proc. Natl. Acad. Sci. USA 95: 6343–6348

  30. Wrana JL . 2000 Cell 100: 189–192

  31. Yang DD, Conze D, Whitmarsh AJ, Barret T, Davis RJ, Rincon M, Flavell RA . 1998 Immunity 9: 575–585

  32. Yeh CH, Peug HC, Huang TF . 1998 Blood 92: 3268–3276

  33. Yue TL, Wang X, Louden CS, Gupta S. Pilarisetti K, Gu JL, Hart TK, Lysko PG, Feuerstein GZ . 1997 Mol. Pharmacol. 51: 951–962

Download references


This work was funded by NIH grants CA52750 and CA64239 to N Bouck and by American Heart Association grant AHA SGD 0030023N to OV Volbert and by Plan Nacional de I+D grant SAF 98-0060 to A Muñoz and Comunidad de Madrid 08.1/0010/2000 to B Jiménez. L Chang was supported by a fellowship from Bank of America-Ginnini Foundation. M Karin is the Frank and Else Schilling American Cancer Society Research Professor.

Author information

Correspondence to Benilde Jiménez.

Rights and permissions

Reprints and Permissions

About this article


  • thrombospondin
  • anti-angiogenic signalling
  • CD36
  • tumor angiogenesis
  • JNK

Further reading