Insulin-like growth factor binding protein-3, induced by hypoxia inducible factor-1α, determines esophageal cancer cell adaptation to the microenvironment

347

Introduction: Insulin-like growth factor binding protein (IGFBP)-3, a key regulator of the bioactivity of insulin-like growth factors (IGFs), is overexpressed frequently in esophageal cancer with concordant overexpression of epidermal growth factor (EGF) receptor Takaoka M et al. Cancer Res. 2004;64:7711-23), in spite of the fact that EGF potently represses IGFBP-3 mRNA transcription through the Ras-MAPK signaling pathway (Takaoka M et al. Am J Physiol. 2006; 290: G404-16). However, the mechanism underlying and the role of IGFBP-3 induction in esophageal tumor biology remain elusive. Methods: hTERT-immortalized human esophageal cells were retrovirally transduced with wild-type (WT) and dominant-negative (DN) hypoxia inducible factor (HIF)-1α. Small hairpin RNA directed against HIF-1α was stably transfected. Ha-Ras^V12-transformed human esophageal cell derivatives (Kim SH et al. Cancer Research 2006; 66:10415-424) were stably transduced with either WT or I56G/L80G/L81G mutant IGFBP-3, the latter incapable of binding IGFs. Xenograft transplantation was done to monitor tumor growth by in vivo bioluminescence imaging. Immunohistochemistry (IHC) determined protein adducts of reductively-activated pimonidazole to assess tissue hypoxia. Gene expression was determined by real-time RT-PCR, Western blotting, and IHC. Results: IGFBP-3 was barely detectable in vitro under normoxic conditions in the presence of EGF or constitutively activated Ras, while it was robustly induced by hypoxia with concomitant stabilization of HIF-1α. Hypoxic induction of IGFBP-3 was augmented by WT-HIF-1α and sharply suppressed by either DN-HIF-1α or RNA interference, implying a role of HIF-1α in transcriptional activation of IGFBP-3. In a xenograft transplantation model with Ras-transformed cells, the hypoxic microenvironment was permissive for induction of endogenous IGFBP-3, thereby leading to transient tumor regression as detected by in vivo bioluminescence imaging with induction of apoptosis as evaluated by caspase-3 cleavage. WT-IGFBP-3 induced apoptosis to a greater extent, and completely abrogated tumor formation, while mutant IGFBP-3 permitted tumor growth without detectable caspase-3 cleavage in tumor tissues. Conclusions: IGFBP-3 is induced by HIF-1α under hypoxic conditions and is not subjected to negative regulation by either EGF receptor signaling or Ras activation in esophageal tumors. Our innovative xenograft transplantation model coupled with in vivo bioluminescence imaging reveals a requirement of IGF-1 receptor signaling for tumor cells to undergo adaptation to the microenvironment, while IGFBP-3 may antagonize a prosurvival effect of IGFs in vivo, a novel mechanistic finding.