JNK1 Mediates Degradation HIF-1α by a VHL-Independent Mechanism that Involves the Chaperones Hsp90/Hsp70

Hypoxia-inducible factor-1 (HIF-1) is a heterodimeric transcription factor that plays a key role in cellular adaptations to hypoxia (deficiency of oxygen supply) by controlling the expression of a series of genes involved in angiogenesis, oxygen transport, and glucose metabolism (1). Hypoxia occurs in many pathologic processes, including ischemical heart disease, stroke, cancer, chronic lung disease, and congestive heart failure (2). Thus, HIF-1 has attracted a great deal of attention in the past few decades and serves as a pharmaceutical and gene therapeutic target for many diseases (3). HIF-1 consists of an oxygen-regulated HIF-1α subunit and a constitutively expressed hydrocarbon receptor nuclear translocator (also called HIF-1β or ARNT). Whereas both HIF-1α and HIF-1β are required for formation of HIF heterodimer, HIF-1α is a key regulatory subunit responsible for HIF transcriptional function (4). In normoxia, HIF-1α is a very unstable protein with a half-life of <5 min. This rapid turnover is mediated by the ubiquitin-proteasome protein degradation system and requires hydroxylation of prolyl residues (P402 and P564) in the conserved oxygen-dependent degradation domain of HIF-1α. Such hydroxylation facilitates HIF-1α binding to a von Hippel-Lindau tumor suppressor gene product (VHL), which leads to the recognition of HIF-1α by the E3 ligase complex for proteasomal degradation (5). This hydroxylation of HIF-1α is executed by the mammalian prolyl hydroxylase domain enzymes (PHD1–PHD3), which require oxygen, ferrous ion, and 2-oxoglutarate for their activity (6). Therefore, deprivation of oxygen or treatment of cells with transition metals (cobalt and nickel), iron chelators (deferoxamine), or the 2-oxoglutarate inhibitor dimethyloxalylglycine (DMOG) could impair the activity of PHDs leading to inhibition of HIF-1α hydroxylation and subsequently result in HIF-1α protein accumulation (7). In the present studies, we used hypoxia and several chemical-mimicked hypoxia conditions to increase HIF-1α protein levels by blocking the PHDs/VHL-dependent degradation pathway and found that, under the above conditions, the HIF-1α protein accumulation was impaired when c-Jun NH₂ terminal kinase 1 (JNK1) expression was deficient. Therefore, our results revealed that JNK1 participated in the VHL-independent HIF-1α regulating mechanism.

The JNKs are implicated in several physiologic processes, including proliferation, apoptosis, and differentiation (8). Although JNK1 or JNK2 mutations are not prevalent in cancer, many tumor cell lines have been reported to possess constitutively active JNKs (9, 10). The expression of JNK1 was markedly increased in breast cancer tissue compared with normal samples (11). Recent studies from Chen's laboratory also indicated that JNK1 played a pivotal role in the expression of the key signature genes and the prognostic outcomes of human hepatocellular carcinoma (12). The molecular mechanisms underlying the oncogenic role of JNK1 need to be further exploited. Our previous studies have shown that JNK1 is responsible for the nickel-related cyclooxygenase-2 induction, a key molecule involved in inflammatory response and tumor development (13). Our present study further definitely shows that JNK1 is also implicated in HIF-1α stabilization in nickel mimicked hypoxia conditions, which adds another evidence for the oncogenic role of JNK1 in nickel-related carcinogenesis.

Beas-2B cells, HaCat cells, 293T cells, and mouse embryonic fibroblasts (MEF) were maintained at 37°C in a 5% CO₂ incubator with DMEM supplemented with 10% fetal bovine serum (FBS) (14); A549 cells were cultured with Ham's F-12K medium supplemented with 10% FBS; HeLa and Cl41 cells were cultured in MEM with 10% FBS and 5% FBS, respectively. Cycloheximide, deferoxamine, trichostatin A, leupeptin, rapamycin, MG132, SP600125, and novobiocin were purchased from Calbiochem. Nickel chloride was from Sigma-Aldrich. DMOG was from Frontier Scientific. Topotecan, UBEI-41, and tamoxifen were from Alexis Biochemicals Corporation. Antibodies against HIF-1α, PHD1, or PHD3 were purchased from Novus Biologicals, Inc. Anti-VHL antibody was from Santa Cruz Biotechnology, Inc. Antibody against heat shock protein 90 (Hsp90) was from Stressgene. Anti-JNK1 antibody was from Invitrogen. Anti-pan lysine acetylation, Hsp70, ATF2, total p70S6 kinase, phosphorylated p70S6K (Thr³⁸⁹ and Thr⁴²¹), and histone deacetylase 6 (HDAC6) antibodies were from Cell Signaling Technology. α-Tubulin and β-actin antibodies were purchased from Sigma.

Hypoxia response element (HRE) and vascular endothelial growth factor (VEGF) luciferase reporters were as described previously (15, 16) and transfected into wild-type (WT) MEFs and JNK1−/− cells. Hsp90 expression vector (17) or Hsp70 expression vector (18) was transfected into JNK1−/− cells, and the transfectants were established and named as JNK1−/− (Hsp90) or JNK1−/− (Hsp70). The dominant-negative mutant of HIF-1α (DN-HIF-1α; refs. 19, 20) was transfected into HaCat cells. A catalytically inactive mutant of HDAC6 construct (HDAC6-DC, H216/611A; ref. 21) was transfected into WT MEFs. The shRNA-JNK1 set (RHS4531) was purchased from Open Biosystems (Thermo Fisher Scientific).

Total RNA was extracted using Trizol reagent (Invitrogen) after various treatment. Total cDNAs were synthesized by ThermoScript™ reverse transcription-PCR (RT-PCR) system (Invitrogen). The mRNA amount present in the cells was measured by semiquantitative RT-PCR. The primers were 5′-AGC CCT AGA TGG CTT TGT GA-3′ and 5′-TAT CGA GGC TGT GTC GAC TG-3′ for mouse hif-1α, 5′-GTG TGC AAC AGC TGA AGG AA-3′ and 5′-ACA GCA GCA CTG GTG TCA TC-3′ for mouse hsp90, 5′-CGA CCT GAA CAA GAG CAT CA-3′ and 5′-ATG ACC TCC TGG CAC TTG TC-3′ for mouse hsp70, and 5′-CCC CAA TCT AGC GGA GGT AAA-3′ and 5′-CAT GAG TGC ATC TAC CAG CCG-3′ for mouse hdac6. The mouse β-actin was used as control (22). The PCR products were separated on 2% agarose gels and stained with ethidium bromide. The results were imagined with Alpha Innotech SP image system (Alpha Innotech Corporation).

The cells stably transfected with HRE or VEGF luciferase reporters were seeded into 96-well plates (8 × 10³ per well) and subjected to the various treatments. Luciferase activities were determined using a luminometer (Wallac 1420 Victor 2 multilabel counter system) as described previously (23).

Cells were exposed to 1 mmol/L NiCl₂ for 24 h. Nickel-containing medium was then removed by rinsing each well with PBS thrice. Nickel-treated cells were cultured in fresh 10% FBS DMEM for 2 d. The cultures were split and subjected to another round of treatment. In total, nickel exposure was repeated twice a week for 8 wk. For the SP600125 or topotecan coexposure groups, cells were pretreated with either SP600125 (50 μmol/L) or topotecan (1 μmol/L) for 4 h before each of nickel exposure. After completion of 8-wk nickel exposure, nickel-induced anchorage-independent growth capability was evaluated in the soft agar assay as described previously (24). The colonies were scored under microscopy.

WT and JNK1−/− MEFs were seeded into 10-cm dishes and treated with 0.5 mmol/L NiCl₂ for 12 h. The nuclear proteins were extracted according to the protocol of nuclear/cytosol fractionation kit (Biovision, Inc.).

After various treatments, cells were extracted in a cell lysis buffer and total protein was quantified with a DC protein assay kit (Bio-Rad). Western blotting was carried out as described previously (25).

Cells (1 × 10⁶) were seeded into 10-cm dishes and cultured for 24 h. Cells were exposed to nickel (0.5 mmol/L) for 12 h and then incubated with methionine-cysteine–free DMEM (Life Technologies-Bethesda Research Laboratories) containing 2% FBS for 1 h. In the pulse assay, [³⁵S]-labeled methionine-cysteine (Trans ³⁵S-Label; ICN, 250 μCi/dish) was added and cultured for indicated time periods and the cells were collected for immunoprecipitation. In the pulse-chase assay, after the cells were pulsed for 45 min, cells were washed thrice in PBS, and cold DMEM with 2% FBS (containing 100 μg/mL l-methionine and 500 μg/mL l-cystein) was then added. Cells were harvested at different time points in cell lysis buffer [1% Triton X-100, 150 mmol/L NaCl, 10 mmol/L Tris (pH 7.4), 1 mmol/L EDTA, 1 mmol/L EGTA, 0.2 mmol/L Na₃VO₄, 0.5% NP40, and complete protein inhibitors mixture tablet] on ice. Total lysate (0.5 mg) was incubated with 2 μg of anti-HIF-1α monoclonal antibody (Sigma) for 2 h at 4°C. Then Protein A/G plus agarose beads (Santa Cruz Biotechnology, Inc.) that were precleared by 20 mg/mL bovine serum albumin for 2 h was added into the mixture and incubated with agitation for an additional 2 h at 4°C. The immunoprecipitated samples were washed with the cell lysis buffer and heated at 100°C for 5 min. Radiolabeled HIF-1α protein and the input cell lysate were assessed using SDS-PAGE analysis.

Total lysate (0.5 mg) from WT and JNK1−/− MEFs was incubated with 2 μg anti-Hsp90 antibody for 2 h at 4°C. Protein A/G plus agarose (40 μL) was added into the mixture and incubated with agitation for an additional 4 h at 4°C. The immunoprecipitated samples were washed with the cell lysis buffer and subjected to the Western blot assay with the anti-pan-lysine acetylation and anti-Hsp90 antibodies.

Exposure of WT MEFs to nickel chloride at concentrations ranging from 0.125 to 0.5 mmol/L induced an accumulation of HIF-1α protein in a dose-dependent manner, and the induction of HIF-1α was sustained at all time points tested. The stabilization of HIF-1α protein by nickel was reproducible in different cell lines, including mouse epidermal cell JB6 Cl41, human normal keratinocyte HaCat, human bronchial epithelial cell Beas-2B, and human cervix adenocarcinoma HeLa cells (Fig. 1A), indicating that HIF-1α induction by nickel exposure is not cell type specific. Nickel exposure elevated HIF-1–dependent transcriptional activity significantly in various cell types by using HRE luciferase reporter (Fig. 1B). Accordingly, VEGF, a HIF-1α downstream target gene, was also induced by nickel exposure in a VEGF luciferase reporter assay (Fig. 1B). In addition, a HIF-1α chemical inhibitor, topotecan (26, 27), and a HIF-1α dominant negative mutant plasmid, DN-HIF-1α (19, 20), were used to further confirm that VEGF induction by nickel was directly dependent on HIF-1α transactivation (Fig. 1C). Then, we detected the involvement of HIF-1α in nickel carcinogenesis by evaluating anchorage-independent growth capability in response to nickel exposures. As shown in Fig. 1D, repeated exposure of cells to nickel for 8 weeks led to cell transformation in soft agar assay. Furthermore, coincubation of cells with topotecan significantly reduced colony formation induced by nickel, strongly suggesting the involvement of HIF-1α in the carcinogenic effect of nickel exposure.

More importantly, HIF-1α protein accumulation was impaired in JNK1−/− cells at both time points tested compared with that in WT cells (Fig. 2A). Moreover, the deficiency of HIF-1α accumulation in JNK1−/− cells was confirmed under the treatments with either hypoxia or PHD enzyme inhibitors, deferoxamine (28) and DMOG (ref. 29; Fig. 2A). In addition, HIF-1–dependent transcriptional activity and VEGF transcription upon nickel exposure were both impaired in JNK1−/− cells (Fig. 2A). Accordingly, nickel-induced cell transformation was also blocked by coincubation of cells with JNK inhibitor, SP600125 (Fig. 2B). And ectopic expression of JNK1 in JNK1−/− cells restored HIF-1α protein accumulation and HIF-1–dependent transcriptional activity in response to nickel exposure compared with those in parental JNK1−/− cells (Fig. 2C). To confirm the contribution of JNK1 to nickel-induced HIF-1α protein accumulation in other types of cells, shRNA-JNK1 was introduced into HeLa and A549 cells to knock down endogenous JNK1 expression. In agreement with above studies, HIF-1α accumulation by nickel exposure was reduced in these JNK1 knockdown cells (Fig. 2D). Taken together, our results show that JNK1 is an important player in the regulation of HIF-1α protein accumulation, HIF-1–dependent transactivation, and its downstream target gene transcription in cellular response to either hypoxia or chemical-mimicked hypoxia conditions.

Our RT-PCR results showed no observable difference on hif-1α mRNA expression between WT and JNK1−/− cells (Fig. 3A). These results indicate that JNK1-mediated HIF-1α accumulation upon nickel exposure might occur at the posttranscriptional level. We then compared HIF-1α cellular distribution between WT and JNK1−/− cells. As shown in Fig. 3B, HIF-1α protein was mostly present in the nuclear fraction in both types of MEFs, and only a small part of HIF-1α protein resided in the cytosolic compartment, suggesting that JNK1 did not affect HIF-1α intracellular distribution. Thus, we determined the possible involvement of JNK1 in regulation of HIF-1α protein translation using [³⁵S]-labeled methionine and cysteine in pulse analysis. As shown in Fig. 3C, HIF-1α protein synthesis rates were almost identical between WT and JNK1−/− cells. Therefore, the potential role of JNK1 in regulation of HIF-1α protein turnover was taken into consideration. By applying the protein translation inhibitor cycloheximide (CHX) to prevent de novo HIF-1α protein synthesis after nickel exposure, we compared HIF-1α protein degradation rate between WT and JNK1−/− cells. As shown in Fig. 3D, in presence of CHX, nickel-stabilized HIF-1α protein underwent a gradual degradation from 2 to 5 hours in WT cells; however, in JNK1−/− cells, HIF-1α protein degradation was augmented and HIF-1α protein was almost absent 4 hours after CHX treatment. We further performed pulse-chase experiment and found that HIF-1α degradation rate was higher in JNK1−/− cells compared with that in WT cells (Fig. 3D). A similar finding was observed when JNK1 expression was knocked down in HeLa cells, although HIF-1α protein degradation rate was slower in parental HeLa cells (Fig. 3D). Interestingly, nickel-induced HIF-1α accumulation was enhanced in JNK1−/− cells by cotreatment of cells with MG132 and nickel, and the ubiquitination of HIF-1α was more abundant in JNK1−/− cells as indicated by the multiple retarded migration bands (Fig. 4A). These results suggest that JNK1 might be responsible for ubiquitin-mediated stabilization of HIF-1α protein in cellular response to nickel exposure. To investigate the involvement of VHL in HIF-1α accumulation, VHL conditional knockout MEFs, in which the expression level of VHL can be depleted by treatment of tamoxifen (30), were used. As shown in Fig. 4B, VHL expression was almost absent in VHL^f/dCre cells after tamoxifen treatment compared with that in VHL^f/+Cre cells. Knockout of VHL led to the increment in both basal and nickel-induced HIF-1α protein levels in VHL^f/dCre cells. These results suggest that VHL is an important player in mediation of HIF-1α protein accumulation. To further determine whether the reduced HIF-1α protein accumulation in JNK1−/− MEFs is due to the upregulation of VHL protein expression, we compared the protein levels of VHL in these two cell lines. However, our results indicated that the VHL protein expression level in JNK1−/− MEFs was not higher, but lower, than that in WT cells, revealing that VHL was not responsible for the reduced HIF-1α protein accumulation in JNK1−/− MEFs (Fig. 4C). Furthermore, PHD proteins (PHD1 and PHD3) show comparable expression levels in WT and JNK1−/− cells (Fig. 4D). Therefore, we anticipate that JNK1 regulates HIF-1α stability in a VHL-independent manner in cellular response to nickel exposure.

Hsp90 and Hsp70 are molecular chaperones required for the stability and function of a number of proteins implicated in cancer cell growth, survival, or both (31–34). When we pretreated cells with novobiocin, a Hsp90/Hsp70 inhibitor that can alter the interaction of Hsp90/Hsp70 with their clients (35, 36), we found that novobiocin reduced nickel-induced HIF-1α accumulation in WT MEFs (Fig. 5A). More importantly, expression levels of Hsp90 and Hsp70 in JNK1−/− cells were much lower than those in WT cells (Fig. 5B), and β-Actin loading controls were at the similar levels between the two cell lines (Fig. 2A). To provide evidence that the reduction of Hsp90/Hsp70 expression is responsible for the attenuated HIF-1α accumulation in JNK1−/− cells, exogenous Hsp90 and Hsp70 were ectopically expressed in JNK1−/− cells. We found that this intervention increased the accumulation of HIF-1α protein in response to nickel compared with that in the parental JNK1−/− cells (Fig. 5B). Our results strongly indicate that Hsp90 and Hsp70 are JNK1 downstream mediators for the regulation of HIF-1α stability in cellular response to nickel exposure.

To investigate the molecular mechanisms involved in JNK1 modulation of Hsp90 and Hsp70 expressions, hsp90 and hsp70 mRNA levels were assessed by RT-PCR. As shown in Fig. 5C, hsp90 mRNA levels were comparable between these two cells, whereas hsp70 mRNA was more abundant in JNK1−/− cells than that in WT cells, indicating that JNK1-mediated regulation of Hsp90/Hsp70 expression occurs at the posttranscriptional levels. CHX was then used to inhibit de novo protein synthesis and to compare the degradation rate of Hsp90/Hsp70 in both cell types. As shown in Fig. 5D, in WT cells CHX treatment for up to 24 hours did not induce a noticeable reduction of Hsp90 and Hsp70 proteins, consistent with the notion that chaperon molecules are always long-life proteins. In contrast, in JNK1−/− cells CHX treatment for 20 and 24 hours led to the obvious reduction of Hsp90 and Hsp70 proteins, clearly indicating that the degradation rate of Hsp90/Hsp70 proteins in JNK1−/− cells was more rapid than that in WT cells. Moreover, this rapid degradation of Hsp90/Hsp70 in JNK1−/− was dependent on the 26S proteasome pathway, because ubiquitin E1 inhibitor (UBEI-41) treatment caused the accumulation of Hsp90/Hsp70 proteins in JNK1−/− cells (Fig. 5D). Collectively, our results show that JNK1 protects Hsp90/Hsp70 from proteasome-dependent degradation, which subsequently contributes to HIF-1α stabilization in cellular response to nickel exposure.

Posttranslational modifications, such as reversible acetylation, can affect function of Hsp90 on ligand-dependent activation of glucocorticoid receptor, a nuclear receptor client of Hsp90 (37, 38), and HDAC6 was identified as a regulator of Hsp90 acetylation (37, 38). Therefore, we tested whether JNK1 could affect Hsp90 activity by regulating HDAC6-mediated Hsp90 acetylation. To this end, we used trichostatin A, a pan HDAC inhibitor, and found that trichostatin A pretreatment decreased nickel-induced HIF-1α accumulation (Fig. 6A), suggesting that HDACs were associated with HIF-1α protein accumulation. To further verify the role of HDAC6 in this process, functional deficient HDAC6 (HDAC6-DC, H216/611A catalytically inactive mutant; ref. 37) was used. As shown in Fig. 6A, the HIF-1α protein induction by nickel exposure was markedly attenuated after an overexpression of HDAC6-DC, indicating the requirement of HDAC6 activity for HIF-1α stabilization upon nickel exposure. We further compared HDAC6 expression levels in WT and JNK1−/− cells. We found that HDAC6 protein was almost absent in JNK1−/− cells, and consequently, the acetylation of Hsp90 was much higher in JNK1−/− cells (Fig. 6B). The RT-PCR results indicated that mRNA expression of hdac6 did not parallel with protein level in JNK1−/− cells. Although nickel treatment reduced hdac6 mRNA in JNK1−/− cells, the basal level of hdac6 mRNA was comparable between these two cells (Fig. 6C,, top left). Therefore, the investigation of posttranscriptional regulation of HDAC6 was performed. Our results showed that HDAC6 degradation was not through MG132-sensitive proteasome pathway, because MG132 could not accumulate HDAC6 protein in WT MEFs (data not shown). Rather, we found that pretreatment of cells with leupeptin, a lysosome inhibitor (39), led to a strong accumulation of HDAC6 in WT MEFs (Fig. 6C,, middle left), suggesting that HDAC6 was decayed in a lysosome-dependent manner. However, leupeptin failed to accumulate HDAC6 in JNK1−/− cells (Fig. 6C,, bottom left). Therefore, we anticipate that the translation machinery of HDAC6 was defective in JNK1−/− cells. Previous studies show that nearly 90% of cellular proteins are translated in eukaryotic cells in a manner that depends on the m⁷GpppN 5′-cap structure of mRNA, which recruits cap-binding protein eIF4E along with the ribosomal preinitiation complex (40). Functional disruption of mTOR by rapamycin could inhibit those cap-dependent protein translation (41). Therefore, we used rapamycin (10 and 20 nmol/L) to treat WT MEFs for 12 and 24 hours, in which conditions the activation of p70S6K was markedly inhibited as evidenced by the disappearance of the phosphorylated form of the p70S6K on T^389/421 (Fig. 6C,, right). However, rapamycin did not show any inhibition on HDAC6 protein expression, revealing that HDAC6 is translated in a cap-independent mechanism. It has been reported that some key proteins that were indispensable for cell survival, differentiation, or apoptosis possesses a cap-independent, internal ribosomal entry site–dependent translation capability (40). Thus, we presumably anticipate that HDAC6 might be translated in an internal ribosomal entry site–dependent mechanism, which is likely mediated by JNK1. The details about JNK1-mediated HDAC6 translation are under investigation in our laboratory. Taken together, our results suggest that the deficiency of HDAC6 in JNK1−/− cells leads to hyperacetylation of Hsp90 protein, which alters Hsp90 activity, and in turn results in HIF-1α protein degradation. Based on these results, we conclude that JNK1 not only stabilizes Hsp90/Hsp70 from proteasome-dependent degradation but also maintains Hsp90 activity by removing acetylation of Hsp90 through regulating HDAC6 expression as summarized in Fig. 6D.

Epidemiologic studies have associated occupational exposure to nickel compounds to elevated incidences of human cancers, such as lung and nasal cancers (42). Nickel has been proposed to contribute to human carcinogenesis by multiple mechanisms, including activation or silencing of certain genes and transcription factors like HIF-1α (19). The molecular mechanism investigation in the current studies further revealed that HIF-1–dependent transcription activity, as well as expression of the downstream target VEGF, can be regulated by JNK1 through modulating HIF-1α stability in a VHL-independent, Hsp90/Hsp70-dependent manner.

It has been reported that a group of positive regulators, such as coactivator proteins and molecular chaperones, contributes to HIF-1α transcriptional activity and stability under hypoxia and normoxia conditions (6). Hsp90 can interact with HIF-1α in vivo and in vitro by binding to HIF-1α Per-ARNT-Sim domain, and such binding is crucial for rapid HIF-1α protein accumulation (17, 32). Hsp90 inhibitors, such as geldanamycin, promote HIF-1α ubiqitination and proteasomal-dependent degradation (34). Our present studies clearly show that, although nickel exposure reduces VHL-dependent HIF-1α degradation due to the inactivation of PHDs, HIF-1α still undergoes a VHL-independent degradation, the efficiency of which is compromised by the presence of JNK1. JNK1 protects HIF-1α from this inefficient turnover after nickel exposure through facilitating the expression of Hsp90/Hsp70, the molecular chaperons known for regulating proper folding, maturation, and stabilization of ∼100 client proteins.

Although JNKs are known to phosphorylate heat shock transcription factor-1, which controls hsp90/hsp70 transcription (43), our RT-PCR results strongly indicate that the hsp90/hsp70 mRNA is not reduced in JNK1−/− cells, implying that JNK1 does not affect their transcription. Actually we have shown that JNK1 prevents Hsp90/Hsp70 from proteasome-dependent degradation. This conclusion is based on these facts: (a) after cycloheximide treatment, the preexisting Hsp90/Hsp70 degraded faster upon JNK1 deficiency and (b) the ubiquitin E1 inhibitor, UBEI-41, could restore Hsp90/Hsp70 expression in JNK1−/− cells. Based on these results, we anticipate that certain unspecified E3 ligase functions as a mediator for Hsp90/Hsp70 protein degradation.

Our present studies further show that JNK1 is able to sustain HDAC6 expression, which is responsible for removing Hsp90 acetylation and consequently facilitating its chaperon activity. The chaperon activity of Hsp90 protein is regulated by its expression level (under heat shock conditions; ref. 44), the interactions with the cochaperones (45), and posttranslational modifications, such as acetylation (37, 46). Although a specific acetyltransferase for Hsp90 still needs to be identified, the deacetylase of Hsp90 protein investigated by Yao and colleagues is attributable to HDAC6 in the case of the ligand-dependent activation of the glucocorticoid receptor (37). HDAC6 belongs to the class II histone deacetylases and possesses unique structural modules so that it may have special biological functions, such as regulation of acetylation of nonhistone proteins, microtubulin (21), and Hsp90 (47). In our present study, we also show that Hsp90 acetylation is elevated in JNK1−/− cells resulting from the total absence of HDAC6 expression.

In summary, our present studies for the first time disclose a novel function of JNK1 in the modulation of HIF-1α stabilization under nickel-mimicked hypoxia conditions through regulation of Hsp90/Hsp70 expression as well as HDAC6-mediated Hsp90 acetylation modification. This novel function may contribute to JNK1-mediated carcinogenic effects in response to nickel exposure.

No potential conflicts of interest were disclosed.

We thank Dr. Laura S. Schmidt (Science Applications International Corporation-Frederick, Inc., National Cancer Institute-Frederick) for providing VHL conditional knockout cells, Dr. Dörthe M. Katschinski1 (Martin-Luther-University Halle) for the generous gift of Hsp90 expression construct, Dr. Hector R. Wong (Children's Hospital Medical Center) for providing Hsp70 expression construct, Dr. Tso-Pang Yao (Duke University) for providing catalytically inactive mutant of HDAC6 construct, and Dr. JaWanda Grant for her critical reading of this manuscript.

Grant Support: Grants from NIH CA112557 (C. Huang), ES012451 (C. Huang), ES010344 (M. Costa), ES005512 (M. Costa), ES014454 (M. Costa), ES000260 (M. Costa), and from the Natural Science Foundation of China 30928023 (C. Huang) and 30971516 (J. Gao).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.