Hypoxic induction of an HIF-1alpha-dependent bFGF autocrine loop drives angiogenesis in human endothelial cells

Abstract

Hypoxia is a major pathophysiological condition for the induction of angiogenesis, which is a crucial aspect of growth in solid tumors. In mammalian cells, the transcriptional response to oxygen deprivation is largely mediated by hypoxia-inducible factor 1 (HIF-1), a heterodimer composed of HIF-1alpha and HIF-1beta subunits. However, the response of endothelial cells to hypoxia and the specific involvement of HIF-alpha subunits in this process are still poorly understood. We show that human umbilical vein endothelial cells (HUVECs) cultured in the absence of growth factors survive and form tubelike structures when cultured under hypoxic, but not normoxic, conditions. HUVECs expressed both HIF-1alpha and HIF-2alpha when cultured under hypoxic conditions. Transfection of HIF-1alpha, but not HIF-2alpha, siRNA to HUVECs completely abrogated hypoxic induction of cords. Neutralizing antibodies to bFGF, but not IGF-1, VEGF, or PDGF-BB, blocked survival and sprouting of HUVECs under hypoxic conditions, suggesting the existence of an autocrine loop induced by low oxygen levels. Notably, bFGF-dependent induction of cord formation under normoxic conditions required HIF-1alpha activity, which was also essential for hypoxic induction of bFGF mRNA and protein expression. These results uncover the existence of an HIF-1alpha-bFGF amplification pathway that mediates survival and sprouting of endothelial cells under hypoxic conditions.

Figure 1.

Hypoxia induces cord formation in HUVECs. (A) HUVECs were seeded on growth factor–reduced Matrigel in GF⁺ or in GF^– medium. Cells were then cultured under 21% O₂ (normoxia) or 1% O₂ (hypoxia) for 16 hours. Cord formation was examined by phase-contrast microscopy at × 10 magnification. Results are representative of 3 independent experiments. (B) HUVECs were cultured as described in panel A. Every well was divided in 3 different parts, and the length of cords was counted in each part using the Bioquant Image Analysis System.

Figure 2.

HIF-1α, but not HIF-2α, mediates hypoxic induction of cord formation in HUVECs. HUVECs were seeded in a 10 mm dish and transfected with a negative control siRNA (NC) or siRNA specific for HIF-1α or HIF-2α. Cells were then treated in the absence or presence of DFO (100 μM) for 6 hours. HIF-1α and HIF-2α mRNA expression (A) was analyzed by real-time PCR, and HIF-1α and HIF-2α protein accumulation (B) was measured by immunoblotting.(C) HUVECs transfected with the indicated siRNA were subjected to hypoxic conditions for 16 hours in GF^– medium, and cord formation was examined. Results are representative of 4 independent experiments. (D) Length of cords from the experiment shown in panel C was quantified using the Bioquant Image Analysis System. Results are presented as mean of 3 different areas examined per well.

Figure 3.

Topotecan inhibits HIF-1α but not HIF-2α in endothelial cells. (A) HUVECs cultured in GF^– medium were incubated under normoxic or hypoxic conditions in the presence or absence of topotecan at the indicated concentrations for 16 hours. HIF-1α and HIF-2α protein accumulation was performed by immunoblotting; β-actin is shown as loading control. (B) HUVECs were seeded on Matrigel in GF⁺ medium under normoxic condition or in GF^– medium under hypoxic conditions for 16 hours in the absence or presence of topotecan (1000 nM to 1 nM) or TNP-470 (100 nM), and cord formation was examined. Results are representative of 3 independent experiments.(C) Length of cords from the experiment shown in panel B was quantified using the Bioquant Image Analysis System.

Figure 4.

bFGF is involved in hypoxic induction of cords in HUVECs. (A) HUVECs seeded on Matrigel in GF^– medium were incubated under hypoxic conditions for 16 hours in the presence or absence of specific neutralizing antibodies for IGF-1, VEGF, bFGF, and PDGF-BB (100 ng/mL). (B) HUVECs were treated with 20 ng/mL IGF-1, VEGF, bFGF, or PDGF-BB, in the presence or absence of the specific neutralizing antibodies (100 ng/mL) for 16 hours, and cord formation was examined. Results are representative of 4 independent experiments.

Figure 5.

bFGF neutralizing antibody inhibits HIF-1α protein accumulation at a late time. (A) HUVECs were seeded in GF^– medium and then cultured under normoxic or hypoxic conditions for 4, 8, 16 and 24 hours. HIF-1α protein accumulation was assessed by immunoblotting (B-C) HUVECs were seeded in GF^– medium and then cultured under normoxic or hypoxic conditions for 6 or 16 hours in the presence or absence of specific neutralizing antibodies for IGF-1, VEGF, bFGF, and PDGF-BB (100 ng/mL). HIF-1α protein accumulation was assessed by immunoblotting. (D) HUVECs were seeded in GF^– medium and then cultured for 16 hours under normoxic conditions in the presence or absence of bFGF (20 ng/mL) or under hypoxic conditions. HIF-1α protein accumulation was assessed by immunoblotting; β-actin is shown as loading control.

Figure 6.

HIF-1α siRNA and TPT inhibit cord formation induced by bFGF. (A) HUVECs transfected with the indicated siRNA were seeded on Matrigel in GF^– medium and treated in the presence or absence of bFGF (20 ng/mL) under normoxic conditions for 16 hours, and cord formation was examined. (B) Length of cords from the experiment shown in panel A was quantified using the Bioquant Image Analysis System as described above. (C) HUVECs were cultured as indicated in panel A under normoxic conditions for 16 hours in the presence or absence of bFGF (20 ng/mL) or TPT (0.1 μM), and cord formation was examined. Results are representative of 4 independent experiments.

Figure 7.

HIF-1α siRNA and TPT inhibit bFGF mRNA and protein accumulation. (A-B) HUVECs transfected with the indicated siRNA were seeded on Matrigel in GF^– medium and cultured under normoxic or hypoxic conditions for 16 hours in the presence or absence of TPT (0.5 μM). Expression of bFGF mRNA (A) was determined by real-time PCR (expressed as fold increase relative to untreated cells), while bFGF protein production (B) was measured in supernatants as described in “Material and methods.” (C-D) HUVECs seeded as in panel A were cultured either under normoxic or hypoxic conditions for 8 and 16 hours (C) or in the presence or absence of increasing concentrations of TPT (1 to 1000 nM) for 16 hours (D). Production of bFGF protein was measured in the supernatants as described.

Figure 8.

Schematic representation of an HIF-1α–bFGF autocrine/paracrine loop in ECs.