NADH elevation during chronic hypoxia leads to VHL-mediated HIF-1α degradation via SIRT1 inhibition

Abstract

Background: Under conditions of hypoxia, cancer cells with hypoxia inducible factor-1α (HIF-1α) from heterogeneous tumor cells show greater aggression and progression in an effort to compensate for harsh environmental conditions. Extensive study on the stability of HIF-1α under conditions of acute hypoxia in cancer progression has been conducted, however, understanding of its involvement during the chronic phase is limited.

Methods: In this study, we investigated the effect of SIRT1 on HIF1 stability in a typical chronic hypoxic conditon that maintains cells for 24 h under hypoxia using Western blotting, co-IP, measurement of intracellular NAD + and NADH levels, semi-quantitative RT-PCR analysis, invasion assay, gene knockdown.

Results: Here we demonstrated that the high concentration of pyruvate in the medium, which can be easily overlooked, has an effect on the stability of HIF-1α. We also demonstrated that NADH functions as a signal for conveyance of HIF-1α degradation via the SIRT1 and VHL signaling pathway under conditions of chronic hypoxia, which in turn leads to attenuation of hypoxically strengthened invasion and angiogenic activities. A steep increase in the level of NADH occurs during chronic hypoxia, leading to upregulation of acetylation and degradation of HIF-1α via inactivation of SIRT1. Of particular interest, p300-mediated acetylation at lysine 709 of HIF-1α is recogonized by VHL, which leads to degradation of HIF-1α via ubiquitin/proteasome machinary under conditions of chronic hypoxia. In addition, we demonstrated that NADH-elevation-induced acetylation and subsequent degradation of HIF-1α was independent of proline hydroxylation.

Conclusions: Our findings suggest a critical role of SIRT1 as a metabolic sensor in coordination of hypoxic status via regulation of HIF-1α stability. These results also demonstrate the involvement of VHL in degradation of HIF-1α through recognition of PHD-mediated hydroxylation in normoxia and p300-mediated HIF-1α acetylation in hypoxia.

Keywords: Angiogenesis; Chronic hypoxia; HIF-1α degradation; Invasion; NADH elevation; SIRT1; VHL.

Fig. 1

NADH controls HIF-1α decay via involvement of SIRT1 during chronic hypoxia. (A) The timecourse of HIF-1α protein levels and intracellular NAD⁺ and NADH concentrations was determined in HeLa cells over 24 h of hypoxia. (B ~ E) Both the extent of NADH oxidation and the HIF-1α maintenance or recovery levels induced by pyruvate were determined by measurement of NAD⁺ and NADH concentrations, and HIF-1α protein, all under hypoxic conditions. Pyruvate (B) or NADH (C,D) or lactate (E) was added to confirm the existence of NADH-sensitive HIF-1α rescue. Statistical significance (p-value) was determined using the ANOVA-t-test. * : p < 0.05, ** : p < 0.01. NADH and pyruvate were added at a concentration of 1 mM

Fig. 2

SIRT1-specific regulatd HIF-1α stability under hypoxic conditions. (A and B) Both the extent of NADH oxidation and the HIF-1α maintenance or recovery levels induced by SIRT1 were determined by measurement of NADH concentrations and HIF-1α protein in the presence of SIRT1-siRNA (A) or SIRT1- shRNA (B). (C) Association of SIRT1 activity with pyruvate-mediated recovery of HIF-1α was also validated in HeLa cells incubated with or without the SIRT1 inhibitors NAM (20 mM), EX-527 (1 µM), or sirtinol (25 µM); these inhibitors were present from commencement of hypoxia to 24 h, and 1 mM pyruvate was added 6 h before harvesting. Chronic decay of HIF-1α was monitored in HeLa cells cotransfected with Flag-tagged HIF-1α and either Myc-tagged wild type-SIRT1 or empty vector (-) (D); or in HeLa cells expressing wild-type SIRT1 (WT), dominant-negative SIRT1 (DN), or empty vector (V) (E). Statistical significance (p-value) was determined using the ANOVA-t-test. * : p < 0.05

Fig. 3

NAD⁺ synthesized via AMPK and NAMPT pathways stabilizes HIF-1α during hypoxia. (A and B) NAD⁺-sensitive HIF-1α accumulation. HIF-1α protein levels were determined under NAD⁺-reduced and -reconstituted conditions after hypoxic exposure for 9 h; NAD⁺ limitation was achieved via use of NAMPT- or AMPK-siRNA whereas NAD⁺ reconstitution involved external addition of 1 mM NAD⁺ (A) or nicotinic acid (NA) (B). (C) NAD⁺/NADH-sensitive recovery in HIF-1α levels. Pyruvate or NAD-mediated recovery in HIF-1α levels was measured under NAD⁺-reduced and NAD⁺-reduced-plus-NADH-upregulated conditions, upon hypoxic exposure for 9 and 24 h, respectively; tests employed AMPK-siRNA treatment. (D) The need for SIRT1 expression in terms of NAD⁺/NADH-sensitive HIF-1α recovery was evaluated in HeLa cells in which SIRT1 was or was not stably depleted, using SIRT1- or control-shRNA, respectively. Statistical significance (p-value) was determined using the ANOVA-t-test. * : p < 0.05, ** : p < 0.01

Fig. 4

HIF-1α acetylation is increased during chronic hypoxia. (A ~ D) HIF-1α acetylation was determined in immunoprecipitated HIF-1α preparations, followed by probing with anti-acetyl-K antibody, in HeLa cells exposed to hypoxia for 9 and 24 h, under condition of MG132 treatment. HIF-1α immunoprecipitates were prepared with anti-HIF-1α or anti-Flag-antibodies from the cells accumulating HIF-1α endogenously (A) or exogenously through transfection with Flag-tagged HIF-1α (B and C). The effect of chronic hypoxia condition (A), pyruvate (B), and Myc-tagged SIRT1 transfection (C) on the acetylation was monitored. (D) Pyruvate/NADH-sensitive recovery in HIF-1α acetylation. HIF-1α acetylation was determined in the presence of pyruvate or NADH under chronic hypoxia condition

Fig. 5

SIRT1 deacetylates Lys709 of HIF-1α to facilitate protein stabilization, even under conditions of protein dehydroxylation. (A) Overexpression of p300 regulated HIF-1a stability. HIF-1α protein levels were determined under hypoxic exposure for 9 h after overexpression of p300. (B) Deacetylation of HIF-1α by SIRT1. 293T cells were transiently transfected with Flag-tagged HIF-1α and HA-tagged p300 together with wt-SIRT1 (WT) or SIRT1/H363Y (DN). HIF-1α acetylation and HIF-1α-SIRT1 interaction in immunoprecipitated HIF-1α were assessed by probing with antibodies to acetylated lysine (AcK), and SIRT1, respectively. (C) 293T cells were transiently cotransfected with 2 µg of each of Flag-tagged HIF-1α, K709R or empty vector, and 2 µg of each of HA-tagged p300 and/or Myc-tagged SIRT1. (D) The stability of K709Q was compared to that of K709R and wt-HIF-1α in SIRT1-depleted HeLa cell clone either treated or not with MG132. (E) Acetylation and ubiquitination of immunoprecipitated WT-HIF-1α and K709R were assessed in in SIRT1-depleted HeLa cell clone

Fig. 6

The extent of HIF-1α acetylation and HIF-1α interaction with VHL is increased during chronic hypoxia via SIRT1 inactivation. (A ~ C) HeLa cells co-transfected with Flag-tagged HIF-1α and either HA-tagged VHL (HA-VHL) were exposed to hypoxic conditions for 9 and 24 h (A and B) or 6 h (C). The effect of pyruvate (A), Myc-tagged SIRT1 (B), and either SIRT1- or control-shRNA (-) transfection (C) on HIF-1α interaction to VHL (A ~ C), acetylation, (B and C) was monitored in immunoprecipitated HIF-1α preparations, followed by probing with anti-acetyl-K and/or anti-HA antibodies, respectively. (D ~ E) The contribution of an association of VHL with HIF-1α in terms of HIF-1α stability was determined in the presence of SIRT1- (+) or control-siRNA (-), or pyruvate in RCC4 and RCC4/VHL cells. (E) The extent of acetylation of immunoprecipitated HIF-1α was assessed in RCC4 and RCC4/VHL cells treated with NAM. (F-G) Acetylation and ubiquitination of immunoprecipitated HIF-1α were assessed in RCC4/VHL cells treated with pyruvate alone (F), or with concomitant addition of NADH 6 h before harvest (G). The concentration of MG132, pyruvate and NADH was 10 mM, 1 mM, and 1 mM, respectively

Fig. 7

Chronic degradation of HIF-1α is independent of proline hydroxylation. HeLa cells were transiently transfected with plasmids encoding HA-tagged wt-HIF-1α (WT), proline mutant P402A/P564A (Mutant) and empty vector (V), or treated with DMOG (0.5mM). (A ~ C) The levels of the exogenous wt- and mutant-HIF-1α proteins (A), and the DMOG-exposed endogenous HIF-1α were determined 9 and 24 h after commencement of hypoxia (B). The effect of pyruvate on the exogenous wt- and mutant-HIF-1α proteins HIF-1α levels was examined (C). The HIF-1α hydroxylation (OH-HIF-1α) (D), and the mutant acetylation and interaction to HA-tagged VHL in HIF-1α immunoprecipitates (E) were measured in the presence of MG132 under acute and chronic hypoxia condition. (F) At the indicated times after DMOG addition, HIF-1α levels were compared between HeLa cells with SIRT1- (+) or control-siRNA transfection (-) under normoxic condition. The concentration of pyruvate and MG132 was 1 mM and 10 mM, rescpectively

Fig. 8

Chronic degradation of HIF-1α attenuates invasive and angiogenesis properties of cancer cells under hypoxic conditions. (A-D) Invasion activity of HeLa and HT1080 cells growing under acute and/or chronic hypoxia over 9 and/or 24 h, respectively, was determined using a matrigel invasion chamber (A). Invasion activity under chronic hypoxia was determined in the absence and presence of either pyruvate alone or together with SIRT1-siRNA (B) or in the presence of Myc-tagged SIRT1(C) or sh-SIRT1 alone or in combination with HIF-1 (D). SIRT1 depletion was achieved with three different SIRT1-siRNAs (lower panel, B). Invasion activity of HeLa cells with sh-control (-) or sh-SIRT1 (+) plasmid was determined after transfection of either Flag-tagged HIF-1α (+) or empty vector (-). (E and F) Angiogenic activity of HeLa cells growing under acute and chronic hypoxia over 9 and 24 h, respectively, was determined by measuring tube length of HUVEC cells in the absence and presence of either pyruvate alone or together with SIRT1-siRNA (+) or control-siRNA (-) (F). (G-J) The amounts of MMP9 and VEGF proteins secreted into culture medium were determined in HeLa cells growing under hypoxia over 9- and 24-h in the absence and presence of pyruvate. The concentration of pyruvate was 0.05 mM. Statistical significance (p-value) was determined using the ANOVA-t-test. * : p < 0.05, ** : p < 0.01, *** : p < 0.001

Fig. 9

A schematic showing how sensing by the redox couple NAD⁺/NADH acetylation mediates HIF-1α degradation and stabilization, employing SIRT1 during acute and chronic hypoxia. During normoxia, oxygen-sensing HIF-1α prolines, Pro402 and Pro564, are hydroxylated by HIF hydroxylases, promoting HIF-1α degradation in an oxygen-sensitive manner. HIF-1α degradation occurred during chronic hypoxia, however, is controlled in an NADH-sensitive manner via acetylation of the protein. NADH, the level of which is upregulated during chronic hypoxia, functions as a destabilizing messenger for chronic HIF-1α decay. Surplus NADH inactivates SIRT1 (iSIRT1), and HIF-1α interaction with VHL that is attenuated after commencement of hypoxia is thus renewed. This leads to elevated ubiquitination and degradation of HIF-1α. However, during the acute phase of hypoxia, NAD⁺ functions as a stabilizing messenger; the NAD⁺-sensing protein SIRT1 (aSIRT1) protects HIF-1α from redox-sensitive acetylation, thereby facilitating dissociation of HIF-1α from VHL and initiating HIF-1α stabilization