2ME2 inhibits HIF-1 mRNA translation and its association with microtubules: Loss of microtubule-dependent HIF regulation in renal carcinoma cells

3265

We have previously shown that the microtubule destabilizing drug 2ME2 down-regulates the HIF-1 α protein (HIF-1α) and inhibits its transcriptional activity at concentrations that disrupt interphase microtubules. We have further shown that 2ME2 inhibits the de novo synthesis of HIF-1α by at least 10-fold, based on immunoprecipitation of newly synthesized HIF-1 α. To investigate the molecular mechanism by which 2ME2 inhibits HIF-1 α synthesis, we examined the polyribosome association profile of the HIF-1 mRNA by linear sucrose gradient fractionation (LSGF). Using LSGF followed by Northern blotting, we showed that the HIF-1 α mRNA was associated with the translationally active components in untreated cells, while upon 2ME2 treatment there was a complete shift towards translation-inactive fractions. This effect was specific for HIF-1α as no shift in the actin and HIF-1β translation was observed. To test the hypothesis that the HIF-1α mRNA requires association with microtubules for efficient translation, we used fluorescent molecular beacons (MB) complimentary to the HIF-1 α mRNA. Transfection of HIF-1α MBs into MCF-7 cells followed by confocal microscopy revealed colocalization between the HIF-1α mRNA and microtubules in untreated cells. Upon 2ME2 treatment extensive microtubule depolymerization was observed leading to lack of colocalization between HIF-1α mRNA and microtubules. These preliminary results support our hypothesis that 2ME2 inhibits HIF-1α translation by affecting the localization of the HIF-1α mRNA complexes with the microtubule cytoskeleton, thus impairing efficient HIF-1α translation. We next investigated whether cells with constitutively high levels of HIF-1 protein due to lack of its degradation (VHL−/− cells) would retain sensitivity to 2ME2. Surprisingly we found that 2ME2 destabilized microtubules but nevertheless had no effect on HIF-1α in renal carcinoma cells lacking VHL and therefore expressing constitutively high HIF-1. In these RCC cells the HIF-1α half life was 4 hr, which is much longer than wt VHL harboring cancer cells (10 min), but nevertheless shorter than the time of exposure to 2ME2 (20 hr). This indicates that RCC cells have lost microtubule control of HIF-1α translation. Next we used the isogenic RCC2/VHL and RCC4/VHLcells, in which wild-type VHL has been stably re-introduced. Despite restoration of VHL, 2ME2 failed to downregulate HIF-1α under hypoxia. Under hypoxia, the half-life of HIF-1α was 4 hr so this could not be responsible for the lack of downregulation of HIF-1 during the 20 hr treatment with 2ME2. We conclude that, in addition to stabilization of HIF-1α due to loss of VHL, RCC cells also acquire microtubule-independent regulation of HIF-1α expression that could not be restored by VHL. We hypothesize that in RCC cells HIF-1α translation is independent of microtubules and contributes to HIF-1 dyregulation.