Anti-angiogenic action of hyperthermia by suppressing gene expression and production of tumour-derived vascular endothelial growth factor in vivo and in vitro

Abstract

Vascular endothelial growth factor is an important angiogenic factor for tumour progression because it increases endothelial-cell proliferation and remodels extracellular matrix in blood vessels. We demonstrated that hyperthermia at 42 degrees C, termed heat shock, suppressed the gene expression and production of vascular endothelial growth factor in human fibrosarcoma HT-1080 cells and inhibited its in vitro angiogenic action on human umbilical vein endothelial cells. The gene expression of alternative splicing variants for vascular endothelial growth factor, VEGF121, VEGF165 and VEGF189, was constitutively detected in HT-1080 cells, but the VEGF189 transcript was less abundant than VEGF121 and VEGF165. When HT-1080 cells were treated with heat shock at 42 degrees C for 4 h and then maintained at 37 degrees C for another 24 h, the gene expression of all vascular endothelial growth factor variants was suppressed. In addition, HT-1080 cells were found to produce abundant VEGF165, but much less VEGF121, both of which were inhibited by heat shock. Furthermore, the level of vascular endothelial growth factor in sera from six cancer patients was significantly diminished 2-3 weeks after completion of whole-body hyperthermia at 42 degrees C (49.9+/-36.5 pg x ml(-1), P<0.01) as compared with that prior to the treatment (177.0+/-77.5 pg x ml(-1)). On the other hand, HT-1080 cell-conditioned medium showed vascular endothelial growth factor-dependent cell proliferative activity and the augmentation of pro-matrix metalloproteinase-1 production in human umbilical vein endothelial cells. The augmentation of endothelial-cell proliferation and pro-matrix metalloproteinase-1 production was poor when human umbilical vein endothelial cells were treated with conditioned medium from heat-shocked HT-1080 cells. These results suggest that hyperthermia acts as an anti-angiogenic strategy by suppressing the expression of tumour-derived vascular endothelial growth factor production and thereby inhibiting endothelial-cell proliferation and extracellular matrix remodelling in blood vessels.

Figure 1

Structure of human VEGF mRNA. Exons are represented by box and numbered. Arrows indicate the specific primers for VEGF variants as shown in Table 1.

Figure 2

Characterisation of gene expression of VEGF splicing variants in human fibrosarcoma HT-1080 cells. Isolated RNA (1 μg) was subjected to RT–PCR analysis with 25 (lanes 1, 4 and 7), 27 (lanes 2, 5 and 8) and 29 cycles (lanes 3, 6 and 9) using specific primers for respective VEGF splicing variants; VEGF₁₂₁, VEGF₁₆₅ and VEGF₁₈₉ as indicated in Figure 1 and Table 1. Two independent experiments were reproducible and typical data were shown. Lanes 1–3, VEGF₁₂₁; lanes 4–6, VEGF₁₆₅ and lanes 7–9, VEGF₁₈₉.

Figure 3

Heat shock suppresses gene expression of VEGF variants in HT-1080 cells. Confluent HT-1080 cells were treated with or without heat shock at 42°C for 4 h and then incubated for another 24 h. Isolated RNA was subjected to RT–PCR analysis with 27 cycles for VEGF₁₂₁ and VEGF₁₆₅ and with 29 cycles for VEGF₁₈₉ as described in Figure 2. The relative amounts of VEGF mRNA were quantified by densitometric scanning followed by normalising against that of GAPDH mRNA and expressed taking the untreated HT-1080 cells as 100. Three independent experiments were reproducible and typical data were shown.

Figure 4

Heat shock suppresses production of VEGF in HT-1080 cells. Confluent HT-1080 cells were treated with or without heat shock as described in Figure 3. The harvested culture medium was subjected to Western blot analysis for VEGF as described in Materials and Methods. The relative amounts of VEGF₁₆₅ were quantified by densitometric scanning and expressed taking the untreated HT-1080 cells as 100. Three independent experiments were reproducible and typical data were shown. Lane 1, untreated HT-1080 cells; lane 2, heat-shocked HT-1080 cells; lane 3, recombinant human VEGF₁₆₅ (10 ng) and lane 4, recombinant human VEGF₁₂₁ (20 ng).

Figure 5

Proliferation of HUVECs by HT-1080 cell-derived conditioned medium. HUVECs (500 cells well^-1) were treated with control medium (filled circles), with the HT-1080 cell-conditioned medium (filled triangles) or with the heat-shocked HT-1080 cell-conditioned medium (filled squares). The proliferation of HUVECs was monitored by alamer Blue assay as described in Materials and Methods. The data are the mean±s.d. of values from six wells at each point. ***Significantly different from HUVECs treated with control medium (P<0.001). Two independent experiments were reproducible and typical data were shown.

Figure 6

Characterisation of HT-1080 cell-derived factor for endothelial-cell proliferation. Conditioned medium from HT-1080 cells was pretreated with an antibody against VEGF (50 μg ml⁻¹) or bFGF (50 μg ml⁻¹) and then HUVECs (500 cells well⁻¹) were treated with or without the conditioned medium. After 2 days treatment, the proliferation of HUVECs was monitored by alamer Blue assay as described in Figure 5. The data are the mean±s.d. of values from six wells. Two independent experiments were reproducible and typical data were shown. Lane 1, HUVECs cultured in control medium; lane 2, HUVECs treated with the HT-1080 cell-conditioned medium; lane 3, HUVECs treated with the HT-1080 cell-conditioned medium pretreated with VEGF antibody and lane 4, HUVECs treated with the HT-1080 cell-conditioned medium pretreated with bFGF antibody. ***Significantly different from HUVECs treated with the HT-1080 cell-conditioned medium (P<0.001).

Figure 7

HT-1080 cell-derived VEGF stimulates HUVECs to produce proMMP-1. (A) Conditioned medium from HT-1080 cells and the culture medium supplemented with human recombinant VEGF₁₆₅ (20 ng ml⁻¹) were pretreated with or without an antibody against VEGF (50 μg ml⁻¹) and then HUVECs were treated with these conditioned media for 24 h. The harvested culture medium was subjected to Western blot analysis for proMMP-1 as described in Materials and Methods. Lane 1, untreated HUVECs; lane 2, HUVECs treated with the HT-1080 cell-conditioned medium; lane 3, HUVECs treated with the HT-1080 cell-conditioned medium pretreated with VEGF antibody; lane 4, HUVECs treated with recombinant human VEGF₁₆₅ and lane 5, HUVECs treated with recombinant human VEGF₁₆₅ pretreated with VEGF antibody. (B) Confluent HUVECs were treated with or without conditioned medium from untreated or heat-shocked HT-1080 cells. Three independent experiments were reproducible and typical data were shown. Lane 1, untreated HUVECs; lane 2, HUVECs treated with the HT-1080 cell-conditioned medium and lane 3, HUVECs treated with the heat-shocked HT-1080 cell-conditioned medium. The relative amounts of proMMP-1 production were quantified by densitometric scanning and expressed taking the untreated HUVECs as 100.