In vivo topoisomerase I inhibition attenuates the expression of hypoxia-inducible factor 1α target genes and decreases tumor angiogenesis

Abstract

Topoisomerase I is a privileged target for widely used anticancer agents such as irinotecan. Although these drugs are classically considered to be DNA-damaging agents, increasing evidence suggests that they might also influence the tumor environment. This study evaluates in vivo cellular and molecular modifications induced by irinotecan, a topoisomerase I-directed agent, in patient-derived colon tumors subcutaneously implanted in athymic nude mice. Irinotecan was given intraperitoneally at 40 mg/kg five times every 5 d, and expression profiles were evaluated at d 25 in tumors from treated and untreated animals. Unexpectedly, the in vivo antitumor activity of irinotecan was closely linked to a downregulation of hypoxia-inducible factor-1α (HIF1A) target genes along with an inhibition of HIF1A protein accumulation. The consequence was a decrease in tumor angiogenesis leading to tumor size stabilization. These results highlight the molecular basis for the antitumor activity of a widely used anticancer agent, and the method used opens the way for mechanistic studies of the in vivo activity of other anticancer therapies.

Figure 1

Growth inhibition of human colon tumor xenografts by irinotecan. Subcutaneous xenograft-bearing mice were treated with irinotecan intraperitoneally at 40 mg/kg q5dx5 on the days indicated by arrows. Tumor growth was evaluated by plotting the mean relative tumor volume ± SEM (n = 8–10) over time. Two representative examples of xenografts that exhibited different levels of in vitro sensitivity to SN-38 (the active metabolite of irinotecan) are shown (see Table 2). The control (C), treated (T) and regrowth (R) samples used for transcriptome analysis are indicated in the corresponding curves. Four to five mice, each bearing two individual tumor xenografts, were included in each group. Results were confirmed in two independent experiments.

Figure 2

Human gene expression profiles of irinotecan-treated colon tumor xenografts. The results of unsupervised hierarchical clustering of 19,818 human probe sets showing significant expression in xenograft samples are summarized in a dendogram in panel A. Numbers in red above the clusters in the tree indicate the percentage of times these clusters were observed in 3,999 iterations during bootstrap analysis. Panel B shows the results of unsupervised hierarchical clustering of 130 human probe sets, for which variation of expression was more specifically influenced by irinotecan than intertumor variability. Numbers in red above the clusters in the tree indicate the percentage of times these clusters were observed in 3,999 iterations during bootstrap analysis. Red indicates high relative expression; green indicates low relative expression.

Figure 3

Irinotecan inhibits HIF1A protein accumulation and angiogenesis in colon tumor xenografts. Western blot analysis of HIF1A protein accumulation in human colon tumor xenografts is shown in panel A. Tumors from mice treated with irinotecan (T) show a complete inhibition of HIF1A protein accumulation compared with control tumors from untreated mice (C). The inhibition of HIF1A protein accumulation is reversed in xenografts that have started to grow again (R) after the end of the treatment period. Results were confirmed in two independent experiments. Panel B shows the detection of intratumor hypoxia in human colon tumor xenografts of mice treated with irinotecan. Mice were injected in-traperitoneally with 100 mg/kg pimonidazole 1 h before euthanasia. Pimonidazole adducts in hypoxic areas of the tumors were identified using hypoxyprobe monoclonal antibodies on paraffin-embedded sections. Bars represent 250 μm. Results were confirmed in two independent experiments. Panel C shows histological examinations of human colon tumor xenografts in mice treated with irinotecan. H&E and trichrome staining show a decrease in tumor cellularity and an increase in extracellular matrix deposition, respectively, in tumors from irinotecan-treated mice (bars represent 2.5 mm). Cellular proliferation was evaluated with Ki-67 labeling within clusters of tumor cells (bars represent 500 μm). CD31 staining shows well-formed vessels (arrows) in the control tumors of untreated mice, whereas the vascular network is completely disorganized in the tumors of irinotecan-treated mice. This effect is reversed in tumors that have started to grow again after the end of the treatment period (bars represent 1 mm). Results were confirmed in two independent experiments. Panel D shows the quantification of the impact of irinotecan treatment on cell proliferation as assessed by Ki-67 staining. The mean values for the average percentage of Ki-67–positive cells were calculated along with SEM. No significant difference was observed in tumor cells from control (C), irinotecan-treated (T) and regrowth (R) samples.