Hypoxia-inducible factor 1 upregulation of both VEGF and ANGPTL4 is required to promote the angiogenic phenotype in uveal melanoma

Abstract

Purpose: Expression of the hypoxia-inducible factor (HIF)-1-regulated gene product, vascular endothelial growth factor (VEGF), correlates with tumor vascularity in patients with uveal melanoma (UM). While the relationship between HIF-1 and VEGF in cancer is well-studied, their relative contribution to the angiogenic phenotype in UM has not previously been interrogated. Here we evaluate the contribution of HIF-1, VEGF, and a second HIF-1-regulated gene product, angiopoietin-like 4 (ANGPTL4), to angiogenesis in UM.

Experimental design: UM cells were examined for expression of HIF-1α, VEGF, and ANGPTL4. Their contribution to the angiogenic potential of UM cells was assessed using the endothelial cell tubule formation and directed in vivo angiogenesis assays. These results were corroborated in tissue from UM animal models and in tissue from patients with UM.

Results: Inhibition of VEGF partially reduced tubule formation promoted by conditioned medium from UM cells. Inhibition of ANGPTL4, which was highly expressed in hypoxic UM cells, a UM orthotopic transplant model, a UM tumor array, and vitreous samples from UM patients, inhibited the angiogenic potential of UM cells in vitro and in vivo; this effect was additive to VEGF inhibition.

Conclusions: Targeting both ANGPTL4 and VEGF may be required for the effective inhibition of angiogenesis in UM.

Keywords: angiogenesis; angiopoietin-like 4 (ANGPTL4); choroidal melanoma; hypoxia inducible factor-1α; vascular endothelial growth factor (VEGF).

Figure 1. HIF-1α expression is increased in UM cells and in UM patient biopsies

A., C., E. Immunoblot assays were performed to determine HIF-1α protein levels in UM cell lines (OMM1, OCM1 and 92.1) following exposure to DMOG (300 μM), hypoxia (1% O₂) or hypoxia and digoxin (dig; 100-300 nM) for 8 or 24 hours and compared to control conditions (20% O₂). B., D., F. Representative images are shown from immunofluorescence analysis of HIF-1α in UM cell lines following exposure to hypoxia (1% O₂ for 8 or 24 hours) or DFO (100 μM for 8 or 24 hours). G. Representative images are shown from immunohistochemical analysis of HIF-1α expression in tumors formed following intravitreal injection of OCM1 cells into mice. Similar results were observed in 3/3 tumors analyzed. H. Representative images are shown from immunofluorescence analysis of HIF-1α protein accumulation and nuclear localization in a human UM tumor biopsy. Similar results were observed in 6/6 UM biopsies examined.

Figure 2. HIF-1α is necessary for the angiogenic potential of UM cells

A. Formation of tubules by human microvascular endothelial cells (HMECs) treated with conditioned medium from UM cell lines cultured under serum starved conditions (1% FBS) and exposed to 20% O₂, hypoxia (1% O₂), hypoxia and 100 nM digoxin, or prolyl hydroxylase inhibitor (300 μM DMOG or 100 μM DFO) in 20% O₂. Treatment with 10% FBS (serum) is used as a positive control for the tubule formation assay. B., C. Expression and nuclear localization of HIF-1α were analyzed by immunoblot B. and immunofluorescence C. assays in parental 92.1 cells and subclones stably expressing either a short hairpin RNA (shRNA) targeting HIF-1α (HIF KD) or a scrambled control (scr) shRNA. D. Aliquots of conditioned medium from parental 92.1 cells exposed to 20% O₂, or subclones expressing control shRNA (scr) or shRNA targeting HIF-1α (HIF KD) and exposed to 1% O₂, were incubated with HMECs and the effect on tubule formation was determined. *P < 0.05; **P < 0.01

Figure 3. HIF-1α-dependent VEGF expression contributes to the angiogenic potential of UM cells

A., B. VEGF mRNA expression A. and protein secretion B. by UM cells exposed to hypoxia (1% O₂) or prolyl hydroxylase inhibitor (300 μM DMOG or 100 μM DFO) in 20% O₂ were determined. C., D. VEGF mRNA expression C. and protein secretion D. were determined in parental 92.1 cells (Control) and in subclones expressing HIF-1α shRNA (HIF KD) or a scrambled shRNA (Scramble) that were exposed to 20% or 1% O₂. E., F. VEGF mRNA expression E. and protein secretion F. were determined in UM cells treated with a scrambled (scr) short interfering RNA (siRNA) or siRNA targeting VEGF. G. The effect on tubule formation of conditioned medium from parental cells (control), or cells transfected with scr or VEGF siRNA and exposed to 20% or 1% O₂ was determined. H. Aliquots of conditioned medium from UM cells, which were cultured under serum starved conditions (1% FBS) and exposed to 20% or 1% O₂ in the presence or absence of serum or bevacizumab (bvczmb), were tested for their effects on tubule formation. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4. ANGPTL4 is a HIF-1-regulated angiogenic factor expressed by UM cells

A., B. ANGPTL4 mRNA expression A. and protein secretion B. in UM cell lines (OMM1, OCM1 and 92.1) exposed to hypoxia (1% O₂), DMOG (300 μM), or DFO (100 μM) in 20% O₂ compared to control conditions (20% O₂) were determined. C., D. ANGPTL4 mRNA expression C. and protein secretion D. were determined in parental 92.1 cells (Control) and in subclones expressing HIF-1α shRNA (HIF KD) or a scrambled shRNA (Scramble) that were exposed to 20% or 1% O₂. E. Representative images are shown from hematoxylin and eosin (H&E) staining and immunohistochemical analysis of VEGF and ANGPTL4 expression in tumors formed following intravitreal injection of OCM1 cells into mice. Similar results were observed in 3/3 tumors analyzed. IgG was used as a negative control. Normal retina (black arrows) and UM tumor cells (red arrows) are labeled. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 5. ANGPTL4 and VEGF promote the angiogenic potential of UM cells

A. Recombinant human ANGPTL4 (1, 10 or 100 ng/mL) was tested for its effect on HMEC tubule formation. B., C. ANGPTL4 mRNA expression B. and protein secretion C. were determined in UM cells treated with a scrambled (scr) siRNA or siRNA targeting ANGPTL4. D. The effect on tubule formation of conditioned medium from parental cells (control), or cells transfected with scr or ANGPTL4 siRNA and exposed to 20% or 1% O₂, was determined. Treatment with 10% FBS was used as a positive control for the tubule formation assay. E. The effect of conditioned medium from cells transfected with scr or ANGPTL4 siRNA and exposed to 20% (control) or 1% O₂ on angiogenesis in vivo was determined using the directed in vivo angiogenesis assay. Representative angioreactors (left) and fold induction compared to control (right) are shown. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 6. ANGPTL4 and VEGF are expressed and are angiogenic in UM tissue

A. Core biopsies from 80 pathology-confirmed primary UMs (in quadruplicate) were used to generate a UM array. The percentage of tumor biopsies in which expression of VEGF (left) or ANGPTL4 (right) was detected by immunohistochemical analysis was determined. Expression was graded as weak (+), modest (++), or strong (+++). NS = no staining. B. Expression levels of VEGF (left) and ANGPTL4 (right) in the vitreous of eyes of patients with primary UM as compared to control patients without UM were determined. C.-G. VEGF and ANGPTL4 mRNA expression C. and protein secretion D., E. were determined in 92.1 cells treated with a scrambled (scr) short interfering RNA (siRNA) or siRNA targeting VEGF, ANGPTL4, or both VEGF and ANGPTL4 and exposed to hypoxia (1% O₂) as compared to control conditions (20% O₂). F., G. The effect on tubule formation of conditioned medium from 92.1 cells transfected with a scrambled (scr) siRNA or siRNA targeting VEGF, ANGPTL4, or both VEGF and ANGPTL4 and exposed to 20% or 1% O₂ was determined. Results are shown as % induction compared to untreated control F. or % inhibition of induction compared to scr siRNA. G. *P < 0.05; **P < 0.01; ***P < 0.001.