Vorinostat suppresses hypoxia signaling by modulating nuclear translocation of hypoxia inducible factor 1 alpha

Abstract

Histone deacetylase inhibitors (HDACis) are a potent class of tumor-suppressive agents traditionally believed to exert their effects through loosening tightly-wound chromatin resulting in de-inhibition of various tumor suppressive genes. Recent literature however has shown altered intratumoral hypoxia signaling with HDACi administration not attributable to changes in chromatin structure. We sought to determine the precise mechanism of HDACi-mediated hypoxia signaling attenuation using vorinostat (SAHA), an FDA-approved class I/IIb/IV HDACi. Through an in-vitro and in-vivo approach utilizing cell lines for hepatocellular carcinoma (HCC), osteosarcoma (OS), and glioblastoma (GBM), we demonstrate that SAHA potently inhibits HIF-a nuclear translocation via direct acetylation of its associated chaperone, heat shock protein 90 (Hsp90). In the presence of SAHA we found elevated levels of acetyl-Hsp90, decreased interaction between acetyl-Hsp90 and HIF-a, decreased nuclear/cytoplasmic HIF-α expression, absent HIF-α association with its nuclear karyopharyin Importin, and markedly decreased HIF-a transcriptional activity. These changes were associated with downregulation of downstream hypoxia molecules such as endothelin 1, erythropoietin, glucose transporter 1, and vascular endothelial growth factor. Findings were replicated in an in-vivo Hep3B HRE-Luc expressing xenograft, and were associated with significant decreases in xenograft tumor size. Altogether, this study highlights a novel mechanism of action of an important class of chemotherapeutic.

Keywords: HDACi; HIF; Hsp90; SAHA; hypoxia.

Figure 1. SAHA suppresses HIF-1α and HIF-2α induction in response to hypoxia

A. Hep3B cells were exposed to hypoxic conditions (1% O₂) for 16 hr in the presence of 0.5 or 1 μM of SAHA (+, and ++, respectively). Representative western blot with normalized densitometric values (protein/actin loading control) show decreases in HIF-1α and HIF-2α expression upon exposure to SAHA. B. Luciferase reporter assay demonstrates significant decreases (p < 0.0001) in HRE-associated luciferase activity in response to SAHA under hypoxic conditions. C. Effects of SAHA on hypoxia related gene expression in Hep3B cells exposed to SAHA for 16 hr under conditions of 21% or 1% O₂ analyzed by qRT-PCR, showing significant suppression of EDN1, EPO, GLUT1, and VEGFA in response to both 0.5 and 1 μM SAHA with no significant change in HIF1A or HIF2A expression. D. Tumor cell lines U87 MG, U2OS, and MG63 were exposed to 0.5 μM SAHA for 16 hr under 21% or 1% O₂, with resulting HIF-1α and HIF-2α suppression similar to that observed in Hep3B cells. Representative western blot with normalized densitometric values (protein/actin loading control) are shown. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 2. SAHA interferes with nuclear translocation of HIF-1α and HIF-2α

A. Hep3B cells were cultured under normoxic or hypoxic conditions for 16 hr. Representative western blot with normalized densitometric values (protein/actin loading control) show no change in total cellular HIF-1/2α expression at low concentrations of SAHA (0.1 μM, +), and moderate decrease at 2 μM (++). B. Representative western blot with normalized densitometric values (protein/histone loading control) show nuclear proteins isolated from Hep3B cells exposed to 0.1 μM (+) SAHA for 16 hr under indicated conditions with decreased nucleated HIF-1α and HIF-2α. C. Luciferase reporter assay demonstrates a significant decrease in HRE-associated luciferase activity in response to SAHA under hypoxic conditions. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 3. Effect of HDACi and Hsp90 inhibitor on HIFs expression

A. Hep3B cells were exposed to the HDACis SAHA (1 μM), LB205 (2 μM), and the Hsp90 inhibitor geldanamycin (GA, 2 μM) under hypoxic culture conditions. Representative western blot with normalized densitometric values (protein/actin loading control) demonstrate a decrease in HIF-1α and HIF-2α levels in all treatment conditions. B. Luciferase reporter assay demonstrates a significant decrease in HRE-associated luciferase activity in response to 0.1 and 2 μM SAHA (+ and ++, respectively), 1 and 5 μM LB-205 (+ and ++, respectively), and 1 μM and 2 μM geldanamycin (+ and ++, respectively). C. Hypoxia related genes were suppressed by 2 μM GA, and with 1 and 5 μM of LB-205 (+ and ++, respectively). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4. SAHA reduces the interaction between HIF-1/2α and Hsp90

A., B. Representative western blot with normalized densitometric values (protein/HA-HIF loading control) after co-immunoprecipitation demonstrate diminished Hsp90 association with HIF-1α and HIF-2α in the presence of SAHA (1 μM). C. Hsp90 acetylation is increased by SAHA (1 μM), seen by immunoprecipitation with anti-Flag antibodies. Normalized densitometric values (protein/Hsp90-Flag loading control) are shown. D. Hsp90β recombinants that mimic acetylated (K286Q and K286A) and dominant-negative (D88N) forms of Hsp90β were cultured in Hep3B cells under hypoxic conditions, with resultant suppressed HIF-1α and HIF-2α expression compared to the WT Hsp90 (- recombinant) condition. Normalized densitometric values (protein/actin loading control) are shown. E., F. Co-immunoprecipitation and representative western blot shows reduced HIF-α-importin interaction in the presence of SAHA (1 μM) or GA (1 μM). Normalized densitometric values (protein/HIF loading control) are shown.

Figure 5. SAHA-induced degradation of HIF-1/2α is mediated by the ubiquitin and proteasome pathway

A. The proteasome inhibitor MG132 (5 μM) reverses SAHA (1 μM) mediated HIF-1α and HIF-2α downregulation in Hep3B cells exposed to hypoxic conditions for 8 hours. Normalized densitometric values (protein/actin loading control) are shown. B., C. Representative western blot with densitometric values after co-immunoprecipitation show exposure to SAHA (1 μM) or LB-205 (2 μM) resulting in increased HIF-ubiquitin association. Normalized densitometric values (protein/HIF loading control) are shown.

Figure 6. SAHA suppressed hypoxia signaling in Hep3B cell xenografts in vivo

A. In-vitro luciferase assay shows decreased HRE activity in Hep3B HRE-Luc cells in the presence of SAHA. B. Tumor growth curves of Hep3B HRE-Luc cell xenografts treated with SAHA show significantly decreased tumor size at 7- and 9-days post-implantation (p < 0.05, p < 0.01, respectively). C. Average tumor weight of excised xenografts is significantly reduced in SAHA-treated mice compared to control mice (p < 0.05) at study endpoint. D. Example luminescence of excised Hep3B HRE-Luc cell xenografts at study endpoint. E. HRE-luc luminescence is significantly decreased in SAHA treated mice compared to control (p < 0.05) at study endpoint. F., G. VEGFA and EPO expression is significantly decreased in tumors resected from SAHA treated mice at study endpoint. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 7. HDACis interfere with hypoxia signaling by affecting Hsp90 acetylation and HIF-α nuclear translocation

In normoxic conditions, HIF-1/2α are first modified by prolyl hydroxylase (PHD) for protein hydroxylation, and removed via the VHL associated proteasomal degradation pathway. In hypoxic conditions, HIF-α recruits Hsp70 and Hsp90 in the cytoplasm, and interacts with the karyopherin importin for nuclear translocation. Nucleated HIF-α further recruits other cofactors such as HIF-b and p300/CBP, and initiates gene transcription and hypoxia signaling. HDACis act to increase accumulation of the acetylated form of Hsp90. The reduction of Hsp90 chaperone activity and HIF-a recognition results in HIF-α nuclear translocation and transcriptional activation.