Biglycan is a specific marker and an autocrine angiogenic factor of tumour endothelial cells

Abstract

Background: We isolated tumour endothelial cells (TECs), demonstrated their abnormalities, compared gene expression profiles of TECs and normal endothelial cells (NECs) by microarray analysis and identified several genes upregulated in TECs. We focused on the gene encoding biglycan, a small leucine-rich repeat proteoglycan. No report is available on biglycan expression or function in TECs.

Methods: The NEC and TEC were isolated. We investigated the biglycan expression and function in TECs. Western blotting analysis of biglycan was performed on sera from cancer patients.

Results: Biglycan expression levels were higher in TECs than in NECs. Biglycan knockdown inhibited cell migration and caused morphological changes in TECs. Furthermore, immunostaining revealed strong biglycan expression in vivo in human tumour vessels, as in mouse TECs. Biglycan was detected in the sera of cancer patients but was hardly detected in those of healthy volunteers.

Conclusion: These findings suggested that biglycan is a novel TEC marker and a target for anti-angiogenic therapy.

Figure 1

Characterisation of isolated TECs and NECs. (A) The binding of lectin BS1-B4 and expression of CD31, CD105 and CD144 (blue line) indicated the high purity of isolated TECs and NECs. The isotype control is shown as a red line. (B) Cultured TECs and NECs were positive for CD31, CD105, CD144, VEGFR-1 and VEGFR-2 by RT–PCR. Mouse tumour stromal CD31 (−) cells were also included in the samples. TECs and NECs were negative for the monocyte marker CD11b and haematopoietic marker CD45. Human HB–EGF expression was detected in human tumour cells but not in TECs or NECs. Abbreviations: NCT=negative control template. (C) Isolated and cultured ECs formed tubes on matrigel-coated plates. Scale bar, 10 μm.

Figure 2

Biglycan is specifically expressed in mouse TECs. (A) The relative expression of biglycan to that of GAPDH in TECs and NECs was measured using quantitative real-time RT–PCR (^**P<0.01). (B) Biglycan protein expression was analysed by fluorescent immunocytochemistry. Biglycan was detected in TECs but not in NECs. Blue: DAPI, Red: biglycan. Scale bar, 25 μm. (C) Western blotting revealed that biglycan protein expression was upregulated in TECs than in NECs. Representative data are shown from one of three experiments. The level of biglycan was normalised to that of β-actin and was analysed by scanning densitometry using Image J software from NIH (^**P<0.01). (D) Biglycan expression in tumour tissues dissected from mice xenografted human tumour cells (A375SM), normal dermal tissue and normal kidney tissue. Fluorescent immunohistochemical staining with the biglycan antibody revealed biglycan (green stain) predominantly in the tumour vessels of mice xenografted human tumour cells. Scale bar, 50 μm. (E) Western blotting revealed that biglycan expression was detected in the TEC supernatant but hardly detected in the NEC supernatant. Representative data are shown from one of three experiments.

Figure 3

Biglycan knockdown inhibited TEC migration and tube formation. (A) Silencing of biglycan mRNA was confirmed by quantitative real-time RT–PCR (^**P<0.01). (B and C) Silencing of the biglycan protein was confirmed by immunocytochemistry and western blotting (^**P<0.01). Scale bar, 25 μm. (D) Migration towards VEGF was significantly inhibited by biglycan siRNA in TECs (^**P<0.01). When biglycan knockdown TECs were treated with exogenous biglycan protein, migration towards VEGF was restored (^*P<0.05). Scale bar, 100 μm. (E) Tube formation was significantly inhibited by biglycan knockdown (^**P<0.01). Exogenous biglycan protein restored the length of tube in biglycan knockdown TECs (^*P<0.05). Representative figures are shown. Scale bar, 100 μm). (F) TEC proliferation was not affected by biglycan siRNA.

Figure 4

Biglycan knockdown caused morphological changes in TECs. (A) TECs with biglycan knockdown became more spread. Vinculin expression increased in TECs with biglycan knockdown. Scale bar, 10 μm. (B) The ratio of cell length vs cell width was (^*P<0.05) decreased in TECs with biglycan knockdown. (C) Vinculin expression was increased in TECs with biglycan knockdown (^*P<0.05).

Figure 5

Biglycan activated NEC migration and tube formation and ECs expressed biglycan receptors. (A) When NECs were treated with exogenous biglycan (20 nM), the number of cells migrating towards VEGF increased (^*P<0.05). Scale bar, 100 μm. (B) When NECs were treated with exogenous biglycan (20 nM), the length of tube increased (^*P<0.05). Scale bar, 100 μm. (C) Expressions of biglycan receptors, TLR2 and TLR4, were analysed in TECs by RT–PCR. Both receptors were expressed in in TECs and NECs. Abbreviation: NCT=negative control template.

Figure 6

Biglycan acts in an autocrine manner through TLR2 and TLR4. (A) TECs migration was inhibited by anti-TLR2 or anti-TLR4 antibodies (^**P<0.01). Scale bar, 100 μm. (B) In biglycan knockdown TECs, the biglycan-induced cell migration was inhibited by blocking anti-TLR2 or anti-TLR4 antibodies (^*P<0.05). Scale bar, 100 μm (C) TECs tube formation was inhibited in the presence of blocking anti-TLR2 or anti-TLR4 antibodies (^**P<0.01, ^*P<0.05). Scale bar, 100 μm. (D) In biglycan knockdown TECs, the biglycan-induced tube formation was suppressed by blocking anti-TLR2 or anti-TLR4 antibodies (^*P<0.05). Scale bar, 100 μm.

Figure 7

Human TECs expressed higher levels of biglycan in vitro and in vivo. (A) RT–PCR confirmed that biglycan was overexpressed in four of the six TEC samples compared with the corresponding NEC samples (n=6) (^*P<0.05). (B) Tumour vessels were double stained with anti-CD31 and anti-biglycan antibodies in human renal cancer, and biglycan was expressed in tumour blood vessels but not in normal kidney vessels (n=6). Scale bar, 50 μm. (C) Western blotting revealed that biglycan expression in the sera of cancer patients was upregulated compared with that in the sera of healthy volunteers. The representative data are shown. (D) Quantification of biglycan protein in human blood serum (n=13). The data are presented as fold change relative to control (Human foetal lung fibroblasts, HFL-1cell lysate 10 μg).