TS-1 enhances the effect of radiotherapy by suppressing radiation-induced hypoxia-inducible factor-1 activation and inducing endothelial cell apoptosis

Abstract

The therapeutic effect of concurrent chemoradiotherapy with TS-1 has been confirmed in various solid tumors; however, the detailed mechanism of action has not yet been fully elucidated. In the present study, we identified hypoxia-inducible factor-1 (HIF-1) as one of the targets of TS-1 in chemoradiotherapy. In growth delay assays using a tumor xenograft of non-small-cell lung carcinoma, H441, TS-1 treatment enhanced the therapeutic effect of single gamma-ray radiotherapy (14 Gy) and significantly delayed tumor growth by 1.58-fold compared to radiotherapy alone (P < 0.01). An optical in vivo imaging experiment using a HIF-1-dependent 5HRE-luc reporter gene revealed that TS-1 treatment suppressed radiation-induced activation of HIF-1 in the tumor xenografts. The suppression led to apoptosis of endothelial cells resulting in both a significant decrease in microvessel density (P < 0.05; vs radiation therapy alone) and a significant increase in apoptosis of tumor cells (P < 0.01; vs radiation therapy alone) in tumor xenografts. All of these results indicate that TS-1 enhances radiation-induced apoptosis of endothelial cells by suppressing HIF-1 activity, resulting in an increase in radiosensitivity of the tumor cells. Our findings strengthen the importance of both HIF-1 and its downstream gene, such as vascular endothelial cell growth factor, as therapeutic targets to enhance the effect of radiotherapy.

Figure 1

TS‐1 improves the control of tumor growth after radiation therapy. (A) Treatment schedule of radiation therapy (RT) and TS‐1. Tumor‐bearing mice in the TS‐1 group and combination group were administered p.o. TS‐1 daily from days 0–4. The tumor xenografts in the RT group and combination group were exposed locally to ionizing radiation with a single dose of 14 Gy on day 2. Intratumoral hypoxia‐inducible factor‐1α (HIF‐1α) expression and HIF‐1 activity, apoptotic cells, and microvessel (MVD) density were analyzed on days 4, 5, and 10, respectively. (B) Tumor volume was measured with calipers and calculated as 0.523LW². Relative tumor volume indicates the ratio of the tumor volume on each day to the corresponding volume on day 0. Results are the means of eight independent tumors ± standard deviation.

Figure 2

TS‐1 suppresses radiation‐induced activation of hypoxia‐inducible factor‐1 (HIF‐1) in tumor xenografts. (A) Tumor‐bearing mice with H441/5HRE‐Luc were treated as mentioned in Fig. 1(A). The intratumoral HIF‐1 activity was monitored as HIF‐1‐dependent bioluminescence on day 4 (insets). Frozen sections of each tumor xenograft were subjected to immunohistochemical analysis with anti‐HIF‐1α (red fluorescence) and anti‐CD31 (green fluorescence) antibody (large panels). Scale line, 50 µm. (B) The H441/5HRE‐Luc tumor xenografts were surgically excised on day 4 and subjected to enzyme‐linked immunosorbent assay to quantify the intratumoral vascular endothelial growth factor (VEGF) level (blue). The VEGF level in each group was compared to that in sham‐treated group and represented as the relative VEGF level. The volume of HIF‐1α‐positive regions was quantified with NIH Image ver. 1.63 software, compared to the entire tumor, and represented as HIF‐1α‐active fraction (red). Results are the mean ± standard deviation. n = 3. *P < 0.05; **P < 0.01.

Figure 3

TS‐1 suppresses hypoxia‐inducible factor‐1 (HIF‐1) activity and consequently enhanced radiation‐induced apoptosis of endothelial cells in vitro. H441/5HRE‐Luc cells were treated with HIF‐1α small interfering RNA (siRNA) or scramble (Scr)‐siRNA or without both of them (–), and cultured under normoxic (N) or hypoxic (H) conditions in the presence (+) or absence (–) of TS‐1. (A) The H441/5HRE‐Luc lysates were subjected to western blotting for HIF‐1α (upper) or β‐Actin (middle) and to luciferase assay (lower). (B) The conditional mediums of the H441/5HRE‐Luc cells were subjected to enzyme‐linked immunosorbent assay for vascular endothelial growth factor (VEGF) level. The VEGF level after each treatment was compared to that after normoxic, TS‐1 (–), and siRNA (–) conditions. (C) Human umbilical vein endothelial cells (HUVEC) were X‐ray irradiated (2 Gy) and cultured in the above conditional mediums for 48 h. The apoptotic fraction was quantified as sub‐G₁ fraction. (D) H441/5HRE‐Luc cells were cultured under normoxic (N) or hypoxic (H) conditions in the presence or absence of TS‐1, 5‐fluorouracil (5‐FU) or 5‐chloro‐2,4‐dihydroxypyridine (CDHP) (see Materials and Methods for detail). The cell lysates were subjected to western blotting for HIF‐1α (upper) or β‐Actin (middle) and to luciferase assay (lower). Results are the mean ± standard deviation. n = 3. *P < 0.05; **P < 0.01.

Figure 4

TS‐1 enhances radiation‐induced apoptosis of endothelial cells. (A) Tumor xenografts with H441/5HRE‐Luc were treated as mentioned in Fig. 1A and surgically excised on day 5. Frozen sections of the solid tumors were stained with both anti‐CD31 antibodies (red fluorescence) and transferase deoxytidyl uridine end labeling (green fluorescence), and representative images are shown. Yellow arrows indicate the positions of apoptotic endothelial cells. Scale line, 10 µm. (B,C) Apoptotic endothelial cells (B) and apoptotic tumor cells (C) detected in Fig. 4A as yellow and green fluorescence, respectively, were counted. Results are the mean of 10 random fields in seven independent tumors ± standard deviation (SD). NS, not significant. *P < 0.05; **P < 0.01.

Figure 5

TS‐1 decreases microvessel density in combination with irradiation. (A) Tumor xenografts with H441/5HRE‐Luc were treated as mentioned in Fig. 1A and surgically excised on days 4, 5, 6 and 10. Frozen sections of the solid tumors were stained with anti‐CD31 antibodies (brown). Relative microvessel densities were calculated by comparing the number of microvessels on the indicated day to that on the first day. Results are the mean of nine random fields in seven independent tumors ± standard deviation (SD). *P < 0.05; **P < 0.01. (B) Microvessel densities of each group on day 10 were counted. Results are the mean of nine random fields in seven independent tumors ± SD. NS, not significant. *P < 0.05. (C) Representative images on day 10 are shown. Scale line, 20 µm.