SEARCH

SEARCH BY CITATION

Keywords:

  • hemin;
  • hypoxia-inducible factor-1α;
  • heat shock protein 90;
  • protein stability;
  • antiangiogenesis

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Hypoxia and growth factor stimulation induce hypoxia-inducible factor-1α (HIF-1α), conferring upon cancer cells the ability to adapt to microenvironments and enhance proliferation, angiogenesis and metastasis. Hemin, an iron-binding porphyrin, has been used to treat porphyria attacks, particularly in acute intermittent porphyria. Although the anti-inflammatory and antitumor effects of hemin were reported, no information is available regarding its effect on HIF-1α. Our study investigated whether hemin and other protoporphyrin compounds have the ability to inhibit HIF-1α activity, and if so, what is the molecular basis of inhibition. Hemin treatment prevented CoCl2-induced HIF-1α expression. HIF-1α inhibition by hemin resulted from an increase in its facilitated ubiquitination and degradation, as shown by the experimental results using cychloheximide treatment and ubiquitination assays. Consistently, hemin repressed HIF-1α-dependent gene transactivation. Intriguingly, hemin directly impeded the binding between heat shock protein 90 (HSP90) and HIF-1α, which was reversed by the addition of an excess amount of ATP required for HSP90 activity. In addition, hemin decreased the expression of client proteins of HSP90. Thus, the inhibition of HIF-1α activity by hemin might result from its interaction with HSP90. Moreover, treatment of protoporphyrin IX, ZnPP or Co(III)PP, but not Mn(III)PP, inhibited HIF-1α induction, indicating that protoporphyrin ring in association with the nature of binding metal leads to HSP90 inhibition. In an in vivo model, hemin treatment inhibited not only the formation of new vessels but also cancer cell proliferation and migration/invasion, supporting the notion that hemin may be applied to the prevention and/or treatment of angiogenesis and/or cancer metastasis.

Hemin (ferriprotoporphrin IX), which belongs to the porphyrin family containing an iron, is the oxidative form of the heme molecules. It has been used to treat porphyria attacks, especially in acute intermittent porphyria.1 Hemin administration had a therapeutic effect in β°-thalassemic patients by enhancing HbF (α2γ2) synthesis, eventually resulting in a threefold increase in hemoglobin content.1 In recent studies, it has been shown that the chemicals containing protoporphyrin ring such as hemin possess substantial antigenotoxic and anti-inflammatory effects against diverse mutagens.2–4 When heme is oxidized, its binding with hemoglobin is less tight, and consequently, it interacts more easily with alternative acceptor proteins and the lipid membranes of cells.5 Hence, hemin may promote oxidative stress, which would facilitate the antioxidant adaptive response in cells. Hemin induced heme oxygenase-1 (HO-1), which enhanced the oxidative degradation of heme molecules to release free iron, biliverdin and carbon monoxide,1 and this, in turn, raised the possibility that hemin has an anticancer effect.1

Hypoxia-inducible factor-1 (HIF-1) is a heterodimeric complex composed of the inducible subunit of HIF-1α and the constitutively expressed HIF-1β (also known as “aryl hydrocarbon nuclear translocator”).6, 7 HIF-1α can be controlled by either an oxygen-dependent or -independent mechanism.8, 9 Hypoxic conditions and hypoxia-mimetic agents (e.g., CoCl2) increase the accumulation of HIF-1α via the enhanced protein stability of HIF-1α. Consequently, activation of HIF-1α promotes cancer progression by inducing its target genes, whose protein products stimulate angiogenesis, glycolysis and tumor proliferation.9, 10 Thus, activated HIF-1α controls the expression of genes involved in the adaptation of cancer cells to unfavorable tumor microenvironments where the supply of oxygen and nutrients are limited.7, 10 Consequently, cancer cells with augmented HIF-1α expression levels become aggressive and acquire chemotherapy resistance.11

Heat shock protein 90 (HSP90) is a ubiquitously expressed molecular chaperone that plays a role in cell signaling, proliferation and survival.12, 13 Moreover, HSP90 is upregulated in cancer cells to adapt to environmental stresses, including nutrient deprivation and hypoxia.12, 13 Under these conditions, HSP90 prevents its client proteins [i.e., HIF-1α, epidermal growth factor receptor (EGFR), endothelial nitric oxide synthase (eNOS), glucocorticoid receptor (GR), p53, Bcr-Abl, cyclin-dependent kinase 4 (CDK4), Raf1, Akt and human epidermal growth factor receptor 2 (HER2)] from degradation and malfunction against cellular stress through its ATPase activity.12–14 The ability of HSP90 to regulate HIF-1α is independent of canonical O2/proline hydroxylase-domain proteins (PHD)/von Hippel-Lindau (VHL) pathway.14–16 Certain chemicals bound to the ATP-binding pocket of HSP90 promote the disruption of HSP90 function, causing ubiquitin-mediated degradation of HIF-1α.15, 17, 18 Therefore, an increasing number of studies have described a series of compounds, such as geldanamycin, 17-AAG, radicicol and deguelin, that inhibit HIF-1α in cell and animal models.7, 17, 18

In view of the possibility that hemin affects the pathways responsible for oxidative degradation of heme molecules, our study explored the effects of hemin and other protoporphyrin compounds on HIF-1α expression and its activity and target gene induction. We determined whether hemin prevents the ability of CoCl2 or hypoxia to increase HIF-1α activity and studied the underlying molecular basis. Here, we report that hemin treatment accelerates HIF-1α ubiquitination and its degradation. As hemin activated various signaling pathways known to affect HIF-1α, our study examined the possible causal relationship between the activated signaling pathways and a decrease in HIF-1α stability by hemin. Finally, the results of the chick chorioallantoic membrane (CAM) model, and cancer cell proliferation and migration/invasion assays confirmed the efficacy of hemin in the repression of HIF-1α and angiogenesis. Our findings indicate that hemin directly interferes with the interaction between HSP90 and HIF-1α. This approach enabled us to identify hemin and other protoporphyrins as novel compounds that inhibit HIF-1α-mediated angiogenesis.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Cells and cell culture conditions

HCT116, a human colon cancer cell line, was obtained from ATCC (Rockville, MD). The cells were incubated in growth medium containing Dulbecco's modified Eagle's medium (DMEM), 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37°C in a humidified atmosphere containing 5% CO2. For all experiments, cells were grown to 80–90% confluency and were starved of serum for 16 hr before treatment. To generate hypoxic conditions, the cells were transferred to a hypoxic chamber (Forma Scientific, Marietta, OH), where they were maintained at 37°C in an atmosphere containing 5% CO2, 1% O2 and 94% N2.

Materials

Hemin, zinc protoporphyrin (ZnPP), insulin, H2O2, MG132, CoCl2, adenosine triphosphate (ATP) and dichlorofluorescein diacetate (DCFH-DA) were purchased from Sigma-Aldrich (St. Louis, MO). Protoporphyrin IX, manganese(III) protoporphyrin (Mn(III)PP) and cobalt(III) protoporphyrin (Co(III)PP) were purchased from Frontier Scientific (Logan, UT). Anti-HIF-1α and anti-HIF-1β antibodies were obtained from BD Biosciences Pharmingen (San Jose, CA). Anti-HSP90 antibody was purchased from StressGen (Victoria, BC, Canada). An antibody specifically directed against ubiquitin was supplied from Sigma-Aldrich (St. Louis, MO). Antibodies recognizing phosphorylated (phospho-) mammalian target of rapamycin (mTOR), phospho-p70 ribosomal S6 kinase 1 (S6K1), phospho-p70S6, phospho-Akt, p70S6K1, p70S6, Akt, lamin A/C and HSP70 were provided from Cell Signaling Technology (Beverly, MA). [Methyl-3H]-thymidine was purchased from Amersham Biosciences (Buckinghamshire, UK).

Immunoblot analysis

Cell lysates were prepared according to previously published methods.19 Briefly, cells were lysed in buffer containing 10 mM Tris-HCl (pH 7.1), 100 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% Triton X-100, 0.5% Nonidet P-40, 1 mM dithiothreitol and 0.5 mM phenylmethylsulfonyl fluoride, which was supplemented with a protease inhibitor cocktail (Calbiochem, La Jolla, CA). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblot analyses were performed as described previously.19 Immunoreactive protein was visualized by the ECL chemiluminescence detection kit (Amersham Biosciences, Buckinghamshire, UK). Equal loading of proteins was verified by immunoblotting for β-actin or HIF-1β. Scanning densitometry of the immunoblots was performed with the Image Scan & Analysis System (Alpha Innotech Corp., San Leandro, CA). The total area of each lane was integrated using AlphaEaseTM version 5.5 software, followed by background subtraction.

HRE reporter gene assay

HRE (hypoxia response element)-A549 cell line was established by transfection of HRE-luciferase reporter plasmid20 into the human lung carcinoma cell line, A549, using LipofectaminPlus (Invitrogen, Carlsbad, CA) and subsequent selection by treatment with G418 (600 μg/ml)(GIBCO, Carlsbad, CA). The cells were maintained in Dulbecco's modified Eagle's medium. Following overnight serum starvation, they were exposed to 100 μM CoCl2 for 24 hr at 37°C with or without hemin (10–30 μM). Luciferase activity was measured by adding luciferase assay reagent (Promega, Madison, WI).

Real-time polymerase chain reaction assays

Total RNA was isolated from cells using Trizol (Invitrogen, Carlsbad, CA), and cDNA was synthesized by reverse transcription using an oligo(dT) primer. Then, real-time polymerase chain reaction (PCR) was performed with a Light Cycler 1.5 (Roche, Mannheim, Germany) using a Light Cycler DNA master SYBR green-I kit, according to the manufacturer's instructions. PCR was performed using primers selective for the genes encoding vascular endothelial growth factor (VEGF) (sense, 5′-AGGAGGGCAGAATCATCACG-3′; antisense, 5′-CAAGGCCCACAGGGATTTTCT-3′),21 glucose transporter 1 (GLUT1) (sense, 5′-CGGGCCAAGAGTGTGCTAAA-3′; antisense, 5′-TGACGATACCGGAGCCAATG-3′)22 and β-actin (sense, 5′-CTCTTCCAGCCTTCCTTCCTG-3′; antisense, 5′-CAGCACTGTGTTGGCGTACAG-3′).23

Measurement of cellular H2O2

DCFH-DA is a cell-permeable nonfluorescent probe, which is cleaved by intracellular esterases and oxidized primarily by H2O2, producing the fluorescent dichlorofluorescein. The level of H2O2 generation was determined by the concomitant increase in dichlorofluorescein fluorescence. HCT116 cells were incubated with 100 μM CoCl2 after 30 μM hemin treatment for 1 hr. The cells were then stained with 10 μM DCFH-DA for 30 min at 37°C, and fluorescence intensity in the cells was measured using a BD FACSCalibur flow cytometer (San Jose, CA).24 In each analysis, 10,000 events were recorded.

Transient transfection

The cells (5 × 105 cells per well) were plated in six-well plates overnight, serum starved for 3 hr and transfected with a plasmid encoding His-tagged ubiquitin (His-ubi) for 6 hr in the presence of FuGENE® HD Reagent (Roche, Indianapolis, IN). Transfected cells were then maintained in Eagle's minimum essential medium containing 1% FBS for 18 hr.

Immunoprecipitation assay

To assess HIF-1α ubiquitination, cells were transfected with the plasmid of His-tagged ubiquitin. Cell lysates were incubated with anti-HIF-1α antibody overnight at 4°C. After immunoprecipitation, the antigen–antibody complex was precipitated following incubation for 2 hr at 4°C with protein G-agarose. The immune complex was solubilized in 2× Laemmli buffer and boiled for 5 min. The samples were resolved and analyzed using 6% SDS-PAGE and then transferred to nitrocellulose membrane. They were then immunoblotted with the antibody directed against His or HIF-1α. In vitro assay was performed using the lysates of HCT116 cells that had been treated with 100 μM CoCl2 for 3 hr and incubated with 60 μM hemin in the presence or absence of 20 mM ATP for 30 min at 37°C. The samples were immunoprecipitated and immunoblotted with the same method as described above.

CAM assay

CAM assay was performed, as previously described.25 Fertilized chick eggs were incubated at 37°C. On the third day of incubation, 3 ml of egg albumin was aspirated by an 18-gauge hypodermic needle to detach the developing CAM from the shell. At 4.5 days, sample-loaded coverslips were air dried and applied to the CAM surface for the testing of angiogenesis inhibition by hemin (30 μM per egg). Two days later, 1–2 ml of 20% fat emulsion (Sigma-Aldrich, St. Louis, MO) was injected into the chorioallantois, and angiogenesis was observed.

[3H]-Thymidine incorporation assay

The rate of DNA synthesis was measured using [3H]-thymidine incorporation assay.26 Postconfluent cells in 12-well plates were incubated with 10% serum for 24 hr after 30 μM hemin treatment. The cells were pulse labeled with 1 μCi/ml [methyl-3H]-thymidine for 8 hr. Chromosomal DNA was precipitated with trichloroacetic acid and extracted with a solution containing 0.5 M NaOH and 0.1% SDS. The radioactivity was quantified using a liquid scintillation counter.

siRNA knockdown of HIF-1α

To knockdown HIF-1α, HCT116 cells were transfected with either an siRNA directed against HIF-1α (5′-CCUAUAUCCCAAUGGAUGAUGTT-3′ and 5′-TTGGAUAUAGGGUUACCUACUAC-3′)27 or a nontargeting control siRNA (100 pmol/ml) using FuGENE® HD Reagent (Roche, Indianapolis, IN) according to the manufacturer's instructions.

In vitro cell invasion/migration assay

An in vitro cell invasion assay was performed using a 24-well Transwell® as described previously.28, 29 The lower side of the filter was coated with Type I collagen, whereas its upper side was coated with matrigel (Collaborative Research, Lexington, KY). The lower compartment was filled with serum-free media containing 0.1% bovine serum albumin. HCT116 cells were placed in the upper part of the Transwell® plate, incubated with 10% serum for 24 hr after 30 μM hemin treatment, fixed with methanol and then stained with hematoxylin for 10 min, briefly followed by eosin. The invasive phenotypes were determined by counting the cells that migrated to the lower side of the filter with microscopy (magnification, 400×). Thirteen visual fields were counted for each filter, and each sample was assayed in triplicate. An in vitro cell migration assay was performed using a 24-well Transwell® unit with polycarbonate filters. Experimental procedures were the same as for the in vitro cell invasion assay except that the filter was not coated with matrigel.

Data analysis

One-way analysis of variance procedures were used to assess significant differences among treatment groups. For each significant treatment effect, the Newman–Keuls test was used to compare multiple group means.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Inhibition of HIF-1α induction

Hypoxia and its mimetic agent, CoCl2, induce HIF-1α accumulation through protein stabilization.6–8 First, we investigated whether hemin and other protoporphyrin compounds (chemical structures shown in Fig. 1a) prevent the induction of HIF-1α by CoCl2 (100 μM, for 3 hr) in HCT116 cells, a human colon cancer cell line. Hemin had an inhibitory effect on HIF-1α accumulation by CoCl2 in a dose-dependent manner, completely abrogating HIF-1α induction at 10 μM or greater concentrations (Fig. 1b, upper). Hemin was nontoxic to the cells at these concentrations (data not shown). Also, it inhibited CoCl2-induced HIF-1α protein level in a time-dependent manner (Fig. 1b, lower). In addition, we examined the effect of ZnPP, another member of protoporphyrin IX, on HIF-1α induction by CoCl2 (Fig. 1c). Treatment of HCT116 cells with 30 μM ZnPP for 1 hr totally abolished HIF-1α protein induction by CoCl2. As an effort to identify the role of protoporphyrin ring in the inhibition of HIF-1α, the effects of protoporphyrin IX and other metal-binding protoporphyrins [i.e., Mn(III)PP and Co(III)PP] on the induction of HIF-1α by CoCl2 were examined. Either protoporphyrin IX or Co(III)PP treatment inhibited the induction of HIF-1α, whereas Mn(III)PP treatment failed to do so (Fig. 1d). Our results indicate that protoporphyrin ring plays a role in the inhibition of HIF-1α.

thumbnail image

Figure 1. Inhibition of HIF-1α induction by hemin. (a) The chemical structures of hemin, ZnPP, protoporphyrin IX, Mn(III)PP and Co(III)PP. (b) Inhibition of CoCl2 induction of HIF-1α by hemin. After serum starvation for 16 hr, HCT116 cells were incubated in a medium containing CoCl2 (100 μM, 3 hr) after treatment with vehicle or 1–30 μM hemin for 1 hr (upper). The expression of HIF-1α was determined in the lysates of HCT116 cells exposed to CoCl2 for the indicated time periods after vehicle or 30 μM hemin treatment for 1 hr (lower). (c) Inhibition of CoCl2 induction of HIF-1α by ZnPP. HCT116 cells were maintained in a medium containing CoCl2 (100 μM, 3 hr) following vehicle, hemin or ZnPP (30 μM each) treatment for 1 hr. (d) The effects of protoporphyrins on HIF-1α induction. Cells were treated with vehicle, protoporphyrin IX, Mn(III) PP or Co(III) PP (30 μM each) for 1 hr and exposed to CoCl2. PP, protoporphyrin IX. (e) Hemin's inhibition of HIF-1α induction by hypoxia, insulin or H2O2. HCT116 cells were incubated under the condition of normoxia or hypoxia (1% oxygen) for indicated time periods (upper). Cells were incubated in a medium containing 100 nM insulin (6 hr) or 300 μM H2O2 (1 hr) following vehicle or hemin (30 μM) treatment for 1 hr (lower).

Download figure to PowerPoint

Next, we determined whether hemin inhibits HIF-1α induction by other stimuli. Hemin treatment markedly reduced the ability of hypoxia to increase HIF-1α protein level (Fig. 1e, upper). Insulin induces HIF-1α accumulation through de novo synthesis or protein stabilization, whereas H2O2 promotes HIF-1α stabilization.9, 30 Hemin treatment repressed HIF-1α induction by either insulin or H2O2 treatment (Fig. 1e, lower). HIF-1α is expressed as a full length (826 amino acids) or smaller form (735 amino acids).31 The small one is produced from its mRNA by alternative splicing and is less active because of the absence of C-terminal transactivation domain: the arrow head designates the full-length form.

CoCl2 treatment increased cellular H2O2 generation (Supporting Information Fig. S1a) and, thus, promoted the stabilization of HIF-1α. As hemin completely inhibited HIF-1α stabilization by either insulin or H2O2, we next explored whether the inhibitory effect of hemin on HIF-1α expression is associated with its antioxidant capacity. Indeed, pretreatment of cells with N-acetyl-L-cysteine (2 mM) or trolox (100 μM) as antioxidant for 1 hr inhibited CoCl2-induced HIF-1α expression (Supporting Information Fig. S1b). However, treatment of these antioxidants failed to prevent HIF-1α repression by hemin (Supporting Information Fig. S1c). Unexpectedly, hemin treatment alone substantially increased H2O2 production and further enhanced CoCl2-induced H2O2 production (Supporting Information Fig. S1d). Therefore, the inhibition of HIF-1α by hemin may not result from an antioxidant effect.

Inhibition of HIF-1α-dependent gene transcription

HIF-1α makes a complex with HIF-1β to form a heterodimer.7 This complex is translocated into the nucleus and mediates target gene induction. The levels of nuclear HIF-1α were determined as an active form in HCT116 cells treated with CoCl2 or CoCl2 + hemin; hemin almost completely prevented the ability of CoCl2 to increase nuclear HIF-1α content for the time periods examined (Fig. 2a). To assess whether hemin suppresses HIF-1α-driven gene transactivation, reporter gene assays were conducted in A549 cells that had been stably transfected with a HRE reporter construct containing the luciferase gene downstream of an HRE element. Although treatment of the cells with CoCl2 caused a 2.5-fold increase in luciferase activity compared to control, hemin inhibited the ability of CoCl2 to induce luciferase compared to control (Fig. 2b). Consistently, hemin treatment abrogated the CoCl2-induced transcription of HIF-1α target genes, specifically VEGF and GLUT1 (Fig. 2c).

thumbnail image

Figure 2. Repression of HIF-1α-dependent gene transcription by hemin. (a) Inhibition of HIF-1α nuclear accumulation by hemin (30 μM). Immunoblottings of lamin A and HSP70 confirmed the purities of nuclear (NF) and cytoplasmic fractions (CF), respectively. (b) HRE reporter gene assay. HRE-A549 cells that had been stably transfected with the HRE-luciferase construct were incubated with CoCl2 (100 μM) for 24 hr following hemin (30 μM) treatment for 1 hr. Luciferase activity was measured in the cell lysates. (c) Real-time PCR assays of HIF-1α target gene transcripts. After serum starvation for 16 hr, HCT116 cells were treated with hemin for 1 hr and continuously incubated with CoCl2 for 6 hr. The mRNA levels of VEGF and GLUT1 were assessed by real-time PCR assays with the level of β-actin mRNA used as a normalizing reference. Values represent mean ± SEM from five independent experiments (treatment mean significantly different from vehicle-treated control, **p < 0.01, or CoCl2, ##p < 0.01).

Download figure to PowerPoint

Increase in ubiquitin-dependent HIF-1α degradation

In a continuing effort to find the molecular mechanism of HIF-1α inhibition, we determined whether hemin affects the stability of HIF-1α protein. Cycloheximide, a compound that prevents protein synthesis, was used to inhibit de novo synthesis of HIF-1α. HCT116 cells treated with cycloheximide in combination with CoCl2 displayed a gradual decrease in HIF-1α level as a function of time. Concomitant hemin treatment facilitated HIF-1α degradation (Fig. 3a). CoCl2 treatment augments HIF-1α expression through impediment of ubiquitination.7 To fully understand the molecular mechanism underlying the hemin-mediated acceleration of HIF-1α degradation, we examined the possibility that hemin facilitates ubiquitination before proteasomal degradation. Immunoprecipitation and immunoblot detection assays indicated that hemin enhances HIF-1α ubiquitination in HCT116 cells (Fig. 3b, upper). These results match the ability of hemin to prevent the induction of HIF-1α by CoCl2 (Fig. 3b, lower). In MG132-treated cells, hemin had no effect on ubiquitination of HIF-1α; however, it further enhanced the ability of CoCl2 to increase the level of total HIF-1α protein (Fig. 3c), suggesting that hemin might affect the equilibrium between HIF-1α stabilization and degradation.

thumbnail image

Figure 3. Increase in proteasomal degradation of HIF-1α by hemin. (a) The effect of hemin on HIF-1α protein stability. HCT116 cells that had been incubated with CoCl2 or CoCl2 + hemin were treated with cycloheximide (CHX, 20 μg/ml) for the indicated time periods and subjected to immunoblottings for HIF-1α and HIF-1β. Values represent mean ± SEM from five independent experiments (treatment mean significantly different from vehicle-treated control, **p < 0.01, or CoCl2, ##p < 0.01). (b) Ubiquitination of HIF-1α. Cells that had been transfected with the plasmid encoding His-tagged ubiquitin (His-Ubi) were treated with vehicle or hemin for 1 hr and treated with CoCl2 for 3 hr. HIF-1α immunoprecipitates were immunoblotted with anti-ubiquitin antibody. HIF-1α and HIF-1β were immunoblotted in the cell lysates. (c) The effect of MG132 on ubiquitinated or total HIF-1α levels. HCT116 cells that had been transfected with the plasmid encoding His-tagged ubiquitin (His-Ubi) were treated with vehicle or hemin for 1 hr and continuously treated with CoCl2 in the presence or absence of 10 μM MG132 for 3 hr. HIF-1α and HIF-1β were immunoblotted in the cell lysates.

Download figure to PowerPoint

The role of HSP90 in the inhibition of HIF-1α by hemin

The molecular chaperon function of HSP90 protects HIF-1α from ubiquitin-dependent proteasomal degradation and, consequently, increases the accumulation of HIF-1α (Fig. 4a). To investigate a precise mechanism responsible for the inhibitory effect of hemin on HIF-1α, the effect of hemin on the interaction between HSP90 and HIF-1α was determined in cells or in a cell-free in vitro condition. In the immunoprecipitation assay using HCT116 cells, hemin treatment hampered the binding of HSP90 with HIF-1α elicited by CoCl2 (Fig. 4b, upper). This effect was verified by reverse immunoprecipitation assay (Fig. 4b, lower). A cell-free in vitro immunoprecipitation assay confirmed the inhibitory effect of hemin on the binding of HSP90 and HIF-1α (Fig. 4c, upper). Moreover, this inhibition was reversed by the addition of excess ATP (20 mM), suggesting that hemin may directly interact with the ATP-binding site of HSP90 (Fig. 4c, lower).

thumbnail image

Figure 4. The role of HSP90 in HIF-1α inhibition by hemin. (a) A proposed model by which hemin inhibits inducible HIF-1α activity. (b) Hemin's inhibition of HSP90 interaction with HIF-1α. HSP90 immunoprecipitates were immunoblotted with anti-HIF-1α antibody (upper), and vice versa (lower). (c) In vitro inhibition of HSP90 and HIF-1α binding by hemin. The lysates prepared from HCT116 cells that had been treated with 100 μM CoCl2 for 3 hr were incubated with 60 μM hemin for 30 min at 37°C. HSP90 immunoprecipitates were immunoblotted with anti-HIF-1α antibody (upper). The lysates from HCT116 cells that had been treated as described above were incubated with 60 μM hemin in the presence or absence of 20 mM ATP for 30 min. HIF-1α immunoprecipitates were immunoblotted with anti-HSP90 (lower). (d) Inhibition of the chaperone functions of HSP90 by hemin. HSP90 client proteins (GR, EGFR, eNOS and HIF-1α) were immunoblotted on the lysates of cells treated with 30 μM hemin or 10 μM geldanamycin (an HSP90 inhibitor, Geld) for 48 hr.

Download figure to PowerPoint

In view of the known chaperone function of HSP90 on HIF-1α protein stability, we examined whether hemin affects the fates of HSP90 client proteins. Apparently, hemin treatment significantly decreased the expression of conventional HSP90 client proteins including GR, EGFR, eNOS and HIF-1α (Fig. 4d), showing that hemin may negatively regulate HSP90 activity. Collectively, these results provide evidence that hemin has the capability to inhibit HIF-1α activity through ubiquitin-dependent proteasomal degradation, and which may result from the inhibition of HSP90 responsible for HIF-1α stabilization.

Inhibition of in vivo angiogenesis and of cancer cell proliferation and invasion/migration

HIF-1α promotes new blood vessel formation, which represents a crucial process in the blood supply. It is expected that the inhibition of HIF-1α by hemin results in an antiangiogenic effect. Having identified the ability of hemin to inhibit HIF-1α, we determined the in vivo inhibitory effect of hemin on angiogenesis using CAM assay. As expected, hemin treatment (30 μM) markedly reduced the number of branched vessels (Fig. 5a).

thumbnail image

Figure 5. The effects of hemin on angiogenesis and on cell proliferation and invasion/migration. (a) Chick chorioallantoic membranes assay. Saline (control) or hemin (30 μM) was loaded at glass coverslip on the chick chorioallantoic membranes (CAM) of chick embryos. After 2 days of incubation, 10% v/v formaldehyde was added onto the surface of CAMs to fix the blood. The disc and surrounding CAMs were incised and photographed; each photograph is representative of CAMs in two independent experimental groups. The result was confirmed by two additional sets of experiments. (b) Inhibition of DNA synthesis by hemin (30 μM, 24 hr). Cell proliferation was assessed by measuring [methyl-3H]-thymidine incorporation (left). DNA synthesis rate was measured in HCT116 cells transfected with siRNA directed against HIF-1α and subsequently treated with serum (10%) for 24 hr (right lower). Immunoblots confirmed protein silencing (right upper). Values represent mean ± SEM from four independent experiments (treatment mean significantly different from vehicle-treated control, *p < 0.05, **p < 0.01, or serum, ##p < 0.01). (c) Inhibition of cell invasion/migration by hemin. Invaded/migrated cells were examined with light microscopy (magnification, ×100, upper). Numbers of invaded cells per field were counted under ×400 light microscopy and quantified (lower). Values represent mean ± SEM from four independent experiments (treatment mean significantly different from vehicle-treated control, **p < 0.01, or serum, ##p < 0.01). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

As HIF-1α activation promotes the proliferation of cancer cells,7–9 we finally assessed the effect of hemin on the DNA synthesis of HCT116 cells using [3H]-thymidine incorporation assay. Hemin treatment resulted in the inhibition of DNA synthesis (Fig. 5b, left). The result of HIF-1α knockdown experiment confirmed that HIF-1α attributes at least in part to cancer cell proliferation (Fig. 5b, right). Thus, the inhibition of cancer cell proliferation may result from hemin's ability to suppress HIF-1α accumulation. Also, hemin treatment inhibited serum-induced invasion/migration of HCT116 cells (Fig. 5c), supporting the notion that the inhibition of HIF-1α by hemin may contribute to the repression of cancer cell invasion/migration. These results demonstrate that hemin may have the ability to restrain not only angiogenesis but also cancer cell proliferation and invasion/migration.

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

HIF-1α belongs to the bHLH and PAS families and forms a heterodimer complex with HIF-1β.6, 32 This complex binds to HIF-1-specific DNA sequences in the mammalian HRE (5′-B(A/G)CGTGVBBB-3′; B represents to all bases except A, and V represents to all bases except T) of target genes.6, 32 HIF-1α contains two transactivation domains, N-terminal and C-terminal. The C-terminal domain especially attributes to the binding with transcription coactivators, such as CBP/p300.9 Activation of HIF-1α by hypoxia or growth factors causes cancer progression by initiating angiogenic gene transcription. Therefore, HIF-1α is considered as a promising pharmacological target against cancer cell development and growth independent of normal cells. A series of anticancer agents have been reported to inhibit HIF-1α.33 Moreover, additional new chemicals are being investigated for HIF-1α inhibition.

The oxygen-dependent degradation domain (ODDD) of HIF-1α is essential for its regulation by oxygen.7 This domain has two proline residues hydroxylated by PHD, which recruits the E3 ubiquitin ligase complex.7 Thus, HIF-1α constantly undergoes the ubiquitin-proteasome degradation under normoxic conditions.7 In our study, we found that hemin increased HIF-1α ubiquitination in cancer cells exposed to CoCl2. Our data showing a substantial increase in HIF-1α ubiquitination by hemin correlate well with a decrease in total HIF-1α level. The hypothesis that hemin enhances proteosomal degradation of HIF-1α is supported by our data showing that hemin accelerates HIF-1α ubiquitination, decreases nuclear HIF-1α content and inhibits HRE target gene transactivation. Hypoxic conditions impede PHD from hydroxylating proline residues due to oxygen deficiency.34, 35 We found that hemin also inhibited HIF-1α activation stimulated by hypoxia. Therefore, it was predicted that the target site of hemin, which led to accelerated HIF-1α degradation, might not depend on the ubiquitination step.

Our observation that hemin, ZnPP, protoporphyrin IX and Co(III)PP, but not Mn(III)PP, had the ability to inhibit HIF-1α induction indicates that metal-binding protoporphyrins have differential inhibitory effects. In addition, we found that cyanocobalamin and hydroxocobalamin, which belong to protoporphyrin IX families, failed to inhibit HIF-1α activation (data not shown), suggesting that the structure of aliphatic chains of protoporphyrin is important. Hence, it is highly likely that the inhibitory effect of hemin on HIF-1α results from protoporphyrin ring in its chemical structure in association with the chemical nature of binding metal. The exact role and chemical properties of metals comprised in protoporphyrin ring remain to be established.

An intriguing finding of our study is that hemin has an inhibitory effect on HSP90 for the repression of HIF-1α activity. This hypothesis is supported by the result of the immunoprecipitation assays showing that hemin attenuated the interaction between HSP90 and HIF-1α. This finding, in turn, led to 26S proteasome-mediated degradation of HIF-1α. In addition, the ability of hemin to suppress the mRNA levels of HSP90′s client proteins strongly supports our hypothesis. Moreover, the result of cell-free in vitro assay strengthens the concept that the inhibitory effect of hemin on the interaction of HSP90 (ATP-dependent chaperone) and HIF-1α may result from the competition of hemin with ATP for the ATP-binding site of HSP90. This contention was strengthened by the reversal of hemin's inhibitory effect on the binding of HSP90 and HIF-1α by the addition of excess amount of ATP. Studies have shown that geldanamycin, radicicol and deguelin impede the interaction of HSP90 and oncogenic proteins, thereby facilitating their protein degradation.17, 36 These chemicals possess a high affinity to the ATP-binding site of HSP90, which contributes to HSP90 dimerization via its conformational change.36 Oh et al. displayed the docked conformation of geldanamycin (a model structure) or deguelin attached to the ATP-binding site of HSP90.17 These structural analyses suggest the possibility that hemin also binds to the same region of HSP90 and competes with ATP.

Reactive oxygen species, in particular H2O2, inactivates PHD enzymes, thereby stabilizing HIF-1α.34, 35 Thus, antioxidants exert inhibitory effects against tumorigenesis and tumor proliferation that rely on HIF-1α activation.37 The roles of hemin on oxidative stress in the cell remain controversial. It has been reported that hemin transcriptionally induced thioredoxin, which acts as an antioxidant and redox regulator.38 In addition, phase II enzymes induced by hemin may help protect cells from oxidative stress-related condition through its radical scavenging activity.38, 39 Hemin induces activation of NF-E2-related factor 2 (Nrf2) and its target antioxidative genes.38 These results raised the possibility that hemin may inhibit HIF-1α through its increased expression of phase II enzymes including hemeoxygenase-1 (HO-1). However, we found that ZnPP, a well-known HO-1 inhibitor, also suppressed CoCl2- or hypoxia-inducible HIF-1α activation in HCT116 cells, which rules out the possibility that Nrf2-mediated HO-1 induction by hemin contributes to its inhibitory action on HIF-1α. Moreover, hemin treatment alone substantially increased H2O2 production and further enhanced CoCl2-induced H2O2 production (Supporting Information Fig. S1b). This result agreed to another report, which stated that high concentration of hemin would cause oxidative stress, leading to metabolic stress, tissue damage and cell death.40 Therefore, potential antioxidant effect of hemin might not be associated with its inhibition of HIF-1α by hemin.

In view of the known signaling pathways controlling HIF-1α, our study further examined changes in the phosphorylation status of major kinases that regulate HIF-1α activity in cells treated with hemin. S6K1 is involved in the regulation of HIF-1α: (i) insulin promotes HIF-1A mRNA translation through S6K1 and thus increases de novo synthesis of the protein, and (ii) S6K1 enhances translation of HIF-1α mRNA via 5′ terminal oligopolypyrimidine (TOP) sequences.41, 42 In HCT116 cells, hemin treatment alone enhanced phosphorylations of Akt and S6K1 rapidly, which did not match the inhibitory action of hemin on HIF-1α (Supporting Information Fig. S2a). Consistently, chemical inhibitors such as Ly294002 and rapamycin (respectively, a phosphatidylinositol 3-kinase/Akt inhibitor and an S6K1 inhibitor) failed to restore the ability of hemin to inhibit HIF-1α induction by CoCl2 (Supporting Information Fig. S2b, top).

AMP-activated protein kinase (AMPK) is also implicated with HIF-1α activity.43, 44 AMPK activation by either 5-aminoimidazole-4-carboxamide ribonucleotide or metformin antagonized HIF-1α induction by insulin or insulin-like growth factor, indicating that chemical activation of AMPK might be associated with HIF-1α inhibition.45 In the subsequent experiment, we examined the possible role of AMPK activation in hemin's inhibition of HIF-1α in cells. Although hemin enhanced the phosphorylation of AMPK at 3–6 hr, compound C (an AMPK inhibitor) treatment did not show reversal of hemin's inhibitory effect on HIF-1α induction (Supporting Information Fig. S2b, upper middle). Recently, it has been known that glycogen synthase kinase-3β (GSK3β) phosphorylates three residues (i.e., S551, T555 and S589) of HIF-1α and provokes its proteasomal degradation.46 Hemin increased the inhibitory phosphorylation of GSK3β at 10–60 min. Nonetheless, the inhibitory effect of hemin on CoCl2 induction of HIF-1α was not antagonized by SB21676, a well-known GSK3β inhibitor (Supporting Information Fig. S2b, upper middle). Therefore, neither AMPK nor GSK3β might be engaged in the inhibition of HIF-1α by hemin.

Protein kinase C δ (PKCδ) is phosphorylated and activated under hypoxia.47 The phosphorylation of PKCδ contributes to stabilization of HIF-1α protein.47 Constitutively active c-Jun N-terminal kinase 1 (JNK1) is also implicated in HIF-1α stabilization in hypoxia-mimicked conditions.48 Therefore, PKCδ and JNK are involved in the regulation of HIF-1α. Unexpectedly, hemin activates both PKCδ and JNK (Supporting Information Fig. S2a), and the inhibitory action of hemin on HIF-1α was not reversed by treatment with either a PKCδ inhibitor (rottlerin) or a JNK inhibitor (SP600125) (Supporting Information Fig. S2b, lower middle and bottom), suggesting that the inhibitory effect of hemin on HIF-1α activity might not result from the activation of PKCδ or JNK. Collectively, our results indicate that the inhibitory effect of hemin on HSP90 interaction with HIF-1α, which leads to 26S proteasome-mediated HIF-1α degradation, may not depend on the signaling pathways examined.

After heme synthesis in mitochondria, it would be subsequently incorporated as a prosthetic group into cellular proteins (i.e., hemoglobin and myoglobin).49 Heme incorporated into apoprotein is unlikely to affect HIF-1α activity because the side chains and metal of protoporphyrin would interact with amino acid residues of apoproteins. Therefore, free heme might have rare chance to encounter and inhibit HSP90. Pharmacokinetic studies showed that the plasma concentration of hemin was calculated to be ∼46.1 μM with the half-life time of 10.8 hr after intravenous administration of hemin arginate (3 mg/kg) in healthy men and porphyric patients.50 During pathological conditions such as ischemia reperfusion or malaria, the level of free heme may be raised up to 20 μM because of severe hemolysis.40 Considering the fact that heme in neutral solutions and in the presence of oxygen is quickly converted to hemin, the estimated half maximal inhibitory concentration (IC50) (∼10 μM) of hemin required for HIF-1α inhibition in our study seems to be physiologically relevant.

Our in vivo study confirmed a decrease in new vessel formation in a CAM assay, strengthening the concept that hemin inhibits HIF-1α activity and thereby diminishes angiogenesis. Moreover, the findings that hemin repressed DNA synthesis and invasion/migration of HCT116 cells support the concept that hemin may have inhibitory effects on angiogenesis, and tumor proliferation and invasion/migration presumably at least in part through HIF-1α inhibition. In summary, the results of our study demonstrate that hemin inhibits HIF-1α activity and HIF-1α-dependent functions (i.e., angiogenesis, cancer cell growth, invasion and migration), and which may result from a decrease in HIF-1α stability via the inhibition of HSP90 responsible for HIF-1α protection. Our finding showing hemin and other protoporphyrins as new HIF-1α inhibitors implies their potential applications not only for the prevention and/or treatment of diseases associated with angiogenesis but also for the inhibition of tumor growth and metastasis.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_26075_sm_suppinfofig1.eps1698KSupporting Information Figure 1.
IJC_26075_sm_suppinfofig2.eps6001KSupporting Information Figure 2.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.