HIV-1 transcription is activated by HIV-1 Tat protein, which recruits cyclin-dependent kinase 9 (CDK9)/cyclin T1 and other host transcriptional coactivators to the HIV-1 promoter. Tat itself is phosphorylated by CDK2, and inhibition of CDK2 by small interfering RNA, the iron chelator 2-hydroxy-1-naphthylaldehyde isonicotinoyl hydrazone (311), and the iron chelator deferasirox (ICL670) inhibits HIV-1 transcription. Here we have analyzed a group of novel di-2-pyridylketone thiosemicarbazone- and 2-benzoylpyridine thiosemicarbazone-based iron chelators that exhibit marked anticancer activity in vitro and in vivo (Proc Natl Acad Sci USA 103:7670–7675, 2006; J Med Chem 50:3716–3729, 2007). Several of these iron chelators, in particular 2-benzoylpyridine 4-allyl-3-thiosemicarbazone (Bp4aT) and 2-benzoylpyridine 4-ethyl-3-thiosemicarbazone (Bp4eT), inhibited HIV-1 transcription and replication at much lower concentrations than did 311 and ICL670. Neither Bp4aT nor Bp4eT were toxic after a 24-h incubation. However, longer incubations for 48 h or 72 h resulted in cytotoxicity. Analysis of the molecular mechanism of HIV-1 inhibition showed that the novel iron chelators inhibited basal HIV-1 transcription, but not the nuclear factor-κB-dependent transcription or transcription from an HIV-1 promoter with inactivated SP1 sites. The chelators inhibited the activities of CDK2 and CDK9/cyclin T1, suggesting that inhibition of CDK9 may contribute to the inhibition of HIV-1 transcription. Our study suggests the potential usefulness of Bp4aT or Bp4eT in antiretroviral regimens, particularly where resistance to standard treatment occurs.
Increased iron stores are correlated with faster HIV-1 progression in iron-loaded patients with thalassemia major, in patients infected with HIV administered with oral iron, and in those with the haptoglobin 2–2 polymorphism (Gordeuk et al., 2001). Moreover, a retrospective study of bone marrow macrophage iron in patients infected with HIV suggested that survival was inversely correlated with higher iron stores (Gordeuk et al., 2001).
Iron chelation therapy has been considered for the control of various infections such as those mediated by protozoa or intracellular pathogens (Hershko, 1994). Several studies have demonstrated the potential of iron chelators in inhibiting HIV replication (Georgiou et al., 2000, 2002; Traoré and Meyer, 2004). In cultured T cells, excess iron stimulates HIV-1 viral replication, whereas iron chelation with desferrioxamine (DFO) lowers viral replication, as measured by decreased p24 levels and reverse transcriptase (RT) activity (Traoré and Meyer, 2004). Treatment of monocyte-derived macrophages and peripheral blood lymphocytes with one of two chelators (DFO or deferiprone) reduced p24 expression and inhibited cellular proliferation (Georgiou et al., 2000). The orally active bidentate chelators 1,6-dimethyl-3-hydroxy-4-(1H)-pyridinone-2-carboxy-(N-methyl)-amide hydrochloride (CP502) and 6-dimethyl-3-hydroxy-4-(1H)-pyridinone-2-carboxy-(N-methyl)-amide hydrochloride (CP511) decreased HIV-1 replication and cellular proliferation in a manner similar to that of DFO and deferiprone (Georgiou et al., 2002). Thus, the observed reduction of HIV-1 replication by these chelators may reflect inhibition of cellular proliferation rather than inhibition of a specific host cell factor involved in HIV-1 replication.
In our previous study, HIV-1 transcription was inhibited in CEM T cells by the tridentate iron chelators 2-hydroxy-1-naphthylaldehyde isonicotinoyl hydrazone (311; IC50 = 2 μM) and ICL670 (also known as deferasirox; IC50 = 23 μM) (Debebe et al., 2007). These iron chelators inhibited the cellular activity of cyclin-dependent kinase (CDK) 2 (Debebe et al., 2007). Although ICL670 has been recently approved for clinical use in the United States and has been successfully implemented for the treatment of iron overload (Porter, 2006), the relatively high concentrations of ICL670 needed for the inhibition of HIV-1 were of concern. The iron chelator 311 was shown to be cytotoxic at low micromolar concentrations (IC50 = 3 μM) (Becker et al., 2003). Hence, the inhibition of HIV-1 that we observed with 311 may be related in part to the inhibition of DNA synthesis and cellular proliferation.
The goal of the present study was to screen and identify novel iron chelators that would be more potent inhibitors of HIV-1 transcription than ICL670 or equally inhibitory but less cytotoxic than 311. We also were interested to determine whether inhibition of HIV-1 transcription was due to the inhibition of CDK9/cyclin T1, a major RNA polymerase II kinase that interacts with HIV-1 Tat and induces HIV-1 transcription (Nekhai and Jeang, 2006). For this purpose, we examined a library of di-2-pyridylketone thiosemicarbazone (DpT) and 2-benzoylpyridine thiosemicarbazone (BpT)-based tridentate iron chelators (Fig. 1A) (Yuan et al., 2004; Whitnall et al., 2006; Kalinowski et al., 2007). The DpT chelators, especially di-2-pyridylketone 4,4-dimethyl-3-thiosemicarbazone (Dp44mT), have been shown to have potent antitumor activity in vitro and in vivo (Yuan et al., 2004; Whitnall et al., 2006). It is noteworthy that mice treated with optimal Dp44mT showed little alteration in hematological and biochemical indices at the chelator doses (0.4 mg/kg/day) required to induce antitumor activity (Whitnall et al., 2006). The BpT-based iron chelators exhibited comparable antineoplastic activity than their DpT homologs in vitro and are among the most active anticancer agents developed by Richardson and colleagues (Kalinowski et al., 2007).
In the current investigation, we identified the two most potent inhibitors of HIV-1 from the BpT and DpT series of chelators and discovered novel molecular mechanisms involved in their activity. Utilization of these new iron-chelating agents in treatment schedules could be of clinical significance in patients with HIV who are resistant to current antiviral treatment regimens.
Materials and Methods
293T and CEM T cells were purchased from the American Type Culture Collection (Manassas, VA). The CEM-HIV-1 [long terminal repeat (LTR)] GFP cells (courtesy of Dr. Jacques Corbeil, Centre de Recherche en Infectiologie du CHUL, Québec, QC, Canada) and OM10.1 cells (courtesy of Dr. Salvatore Butera, Retrovirus Diseases Branch, DVRD/NCID, Centers for Disease Control and Prevention, Atlanta, GA) were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program. Histone H1 was purchased from Millipore (Billerica, MA). Anti-FLAG monoclonal antibody, protein G, and protein A agarose were purchased from Sigma (St. Louis, MO). All radioactive reagents were purchased from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK). Antibodies against CDK9, CDK2, and cyclin T1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Horseradish peroxidase-conjugated F(ab)2 fragment was purchased from GE Healthcare. All other inorganic reagents were purchased from Thermo Fisher Scientific (Waltham, MA) or Sigma. Roscovitin was purchased from Calbiochem (La Jolla, CA), and 4-amino-6-hydrazino-7-β-d-ribofuranosyl-7H-pyrrolo[2,3-d]-pyrimidine-5-carboxamide (ARC) was a gift from Dr. Andrei Gartel (University of Illinois, Chicago, IL).
All chelators were synthesized by Schiff base condensation between the relevant aromatic ketone or aldehyde with a thiosemicarbazide or acid hydrazide using standard methods (Kalinowski et al., 2007). The chelators were characterized using a combination of elemental analysis, 1H NMR spectroscopy, infrared spectroscopy, and mass spectroscopy.
All cells were cultured at 37°C in a 5% CO2 atmosphere. CEM T cells and peripheral-blood mononuclear cells (PBMCs) were cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum, and 1% antibiotic solution (penicillin and streptomycin; Invitrogen). CEM GFP cells were cultured in RPMI 1640 medium with all the additions above plus 500 μg/ml G418 (Invitrogen). The 293T cells and OM10.1 cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum and 1% antibiotic solution (penicillin and streptomycin).
Induction of HIV-1 Transcription with Adenovirus Carrying Tat.
The E1-deleted recombinant adenovirus carrying Tat was generated as described previously (Debebe et al., 2007). In brief, CEM-GFP cells were infected in 96-well plates containing 4 × 105 cells/well. After 24 h at 37°C, 10-μl aliquots were removed, supplemented with trypan blue, and counted to determine cellular viability. The remaining cells were transferred to a white plate (PerkinElmer Life and Analytical Sciences, Waltham, MA), and fluorescence was measured at an excitation wavelength of 488 nm and an emission wavelength of 510 nm by using a luminescence spectrometer (LS50B; PerkinElmer Life and Analytical Sciences).
Cytotoxicity data are presented as a percentage of the control versus logarithm of the chelator concentration and analyzed using Prism software (ver. 3; GraphPad Software, La Jolla, CA). The IC50 values were determined from the dose-response (variable slope) curve using a four parameter logistic equation: Y = Bottom + (Top − Bottom)/(1 + 10((logEC50−X) × nH)), where Y is a percentage of the control, X is logarithm of the chelator concentration, and nH is Hill slope.
The HIV-1 reporter containing HIV-1 LTR (−138 to +82) followed by a nuclear localization signal and the LacZ reporter gene (courtesy of Dr. Michael Emmerman, Fred Hutchinson Cancer Institute, Seattle, WA) and the pNL4-3.Luc.R−E− (courtesy of Prof. Nathaniel Landau, New York University School of Medicine, New York, NY) were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program. The NF-κB driven e-selectin-luciferase reporter was a gift from Dr. Chou-Zen Giam (Uniformed Services University of the Health Sciences, Bethesda, Maryland) (Fu et al., 2003). The following plasmids were kindly provided by Dr. Manuel López-Cabrera (Unidad de Biología Molecular, Madrid, Spain): WT HIV-1 LTR (−105– +77) followed by the luciferase reporter gene (WT SP KB); HIV-1 (−105– +77) with Sp1-inactivated sites followed by the luciferase reporter gene (WT SP 123); and HIV-1 (−81– +77) with NF-κB-deleted sites followed by the luciferase reporter gene (WT SP).
293T cells were cotransfected with Tat-expressing vector and HIV-1 LTR-LacZ at 30% confluence using Lipofectamine and Plus reagents (Invitrogen). After transfection, the cells were cultured for an additional 48 h at 37°C, and β-galactosidase activity was analyzed with the use of the quantitative o-nitrophenyl-β-d-galactopyranoside (ONPG)-based assay (see next section) (Debebe et al., 2007).
Cells were washed with PBS and lysed for 20 min at room temperature in 50 μl of lysis buffer containing 20 mM HEPES at pH 7.9, 0.1% Tergitol-type NP-40 (NP-40), and 5 mM EDTA. Subsequently, 100 μl of ONPG solution (72 mM Na2PO4 at pH 7.5, 1 mg/ml ONPG, 12 mM MgCl2, and 180 mM 2-mercaptoethanol) was added and incubated at room temperature. Standards were incubated for 10 min. Based on the linear part of the standard curve, the color development was monitored and stopped when the signal was in the middle of the linear range. The reaction was terminated by the addition of 100 μl of 1 M Na2CO3. The 96-well plate was analyzed in a microplate reader at 414 nm (Multiskan MS; Thermo Fisher Scientific).
Cell Viability Assays (Propidium Iodide Uptake, Calcein-AM Uptake, and Trypan Blue Exclusion Assay).
CEM T cells were grown in 96-well plates at 37°C and incubated with iron chelators for the periods of time indicated in the figure legends. To measure toxicity by propidium iodide (PI; Sigma) uptake, PI (250 μg/ml) was incubated with cells for 30 min at 37°C. PI fluorescence was measured at an excitation and emission wavelengths of 488 and 617 nm, respectively, by using the luminescence spectrometer described above.
To assess cytotoxicity with calcein, media was removed and the cells were washed with Dulbecco's PBS to remove serum esterase activity that may cause an increase in fluorescence through the hydrolysis of calcein-AM. Cells were then supplemented with 0.2 μM calcein-AM (Invitrogen) for 10 min at 37°C. A positive control containing 100% dead cells was prepared by treating cells with Triton X-100 [1% (v/v)] that were then incubated with 0.2 μM calcein-AM. Fluorescence was measured using the luminescence spectrometer described above implementing an excitation wavelength of 495 nm and emission filters at 515 nm. To measure cellular viability, the cells were supplemented with 0.2% trypan blue, transferred to a plastic disposable counting chamber and counted on a Cellometer Automatic Cell Counter (Nexcelcom Bioscience, Lawrence, MA).
Preparation of Pseudotyped HIV-1 Expressing Luciferase.
293T cells were grown on 100-mm plates at 37°C and transfected using the calcium phosphate method. Cells were transfected with the VSVG-expressing vector (gpHEF-VSVG) and pNL4-3.Luc.R−E− molecular clone that contains two nonsense frame shifts in the Env and Vpr genes. In this vector, the luciferase gene was cloned in place of nef (He and Landau, 1995). Seventy-two hours after transfection, the media were collected and briefly centrifuged at 1000g for 10 min, and the virus was collected by centrifugation at 4°C for 6 h at 14,000g. The precipitated virus was resuspended in PBS containing 10% glycerol, separated into aliquots, and stored at −70°C. The p24 analysis of the HIV-1 Luc was performed by enzyme-linked immunosorbent assay using a kit from ZeptoMetrix (Buffalo, NY).
CEM T cells were infected with VSVG-pseudotyped pNL4-3.Luc.R−E− virus and then cultured at 0.5 × 106 cells/ml in six-well plates at 37°C and 5% CO2. The cells were collected, washed with PBS, and resuspended in 100 μl of PBS. Then, 100 μl of reconstituted luciferase buffer (Luclite Kit; PerkinElmer Life and Analytical Sciences) was added to each well. After a 10-min incubation, the lysates were transferred into white plates (PerkinElmer Life and Analytical Sciences), and luminescence was measured using LuminoSkan RT equipment (Thermo Fisher Scientific).
Source of PBMCs.
PBMCs were purchased from Astarte Biologics (Redmond, WA). Donors were negative for HIV-1 and -2, hepatitis B, hepatitis C, and human T-lymphotropic virus type 1. PBMCs were isolated from peripheral blood by apheresis with additional purification by density gradient centrifugation and then cryopreserved until used.
Infection of PBMCs with VSVG HIV-Luc and Inhibition of HIV-1 with Iron Chelators.
PBMCs were infected with HIV-1 Luc virus at approximately 1 ng of p24 per 5 × 106 cells. After 48 h, the cells were seeded on 96-well white plates and incubated with the iron chelators for 24 h at 37°C. Luciferase buffer was then added to each well, and luminescence was measured using Luminoskan RT equipment as described under Luciferase Assay.
Viability of PBMCs after Iron Chelator Treatment.
PBMCs were grown on a 96-well white plate and incubated with the iron chelators for different time periods at 37°C. Later, cells were washed with PBS and incubated with 0.2 μM calcein-AM for 15 min, and fluorescence was measured at an excitation wavelength of 495 nm and an emission wavelength of 515 nm using Luminescence Spectrometer LS 50B (PerkinElmer Life and Analytical Sciences). Values presented are mean ± S.D. (n = 4) and are typical of three independent experiments.
Inhibition of HIV-1 Replication Using Iron Chelators.
In these studies, OM10.1 cells were used where the integrated HIV-1 proviral DNA was originally derived from the subtype B, LAI strain (Butera et al., 1994). To assess the activity of chelators on inhibiting HIV-1 replication, 106 OM10.1 cells were grown to 70% confluence and treated with 10 ng/ml TNF-α for 2 h at 37°C. Subsequently, TNF-α was removed, and the cells were washed with PBS and then treated with the concentrations of iron chelators indicated in the figure legends. Media was collected 48 h after the drug treatment, and the activity of RT was determined. Viral supernatants (10 μl) were incubated in a 96-well plate with RT reaction mixture containing 1× RT buffer (50 mM Tris-HCl, 1 mM DTT, 5 mM MgCl2, and 20 mM KCl), 0.1% Triton X-100, 0.01 units of poly(A), 0.01 units of poly(dT), and [3H]TTP. The mixture was incubated overnight at 37°C, and 5 μl of the reaction mix was spotted on a DEAE Filtermat (PerkinElmer Life and Analytical Sciences), washed four times with 5% Na2HPO4 and three times with water, and then dried completely. RT activity was measured using a Betaplate counter (PerkinElmer Life and Analytical Sciences).
Analysis of the NF-κB-Driven E-Selectin-Luciferase Reporter Activity.
293T cells were seeded at 30% confluence in 96-well plates and allowed to grow overnight. The following day, the cells were transfected with CMV-EGFP reporter in combination with the NF-κB driven e-selectin-luciferase reporter (Fu et al., 2003), or pNL4-3.Luc.R−E− (HIV-1 promoter) for 3 h at 37°C using Lipofectamine and then incubated with Bp4aT or Bp4eT for 48 h at 37°C. The media were then removed from the cells, and equal amounts of PBS and luciferase buffer were added to measure luciferase activity. After luciferase activity, GFP was measured at an excitation wavelength of 495 nm and an emission wavelength of 515 nm.
Analysis of HIV-1 Promoters with the Deletion of NF-κB and Inactivation of Sp1 sites.
293T cells were transfected at 30% confluence in 96-well plates with vectors in which the luciferase reporter was driven by WT HIV-1 LTR (−105 to +77), HIV-1 LTR (−105 to +77), with Sp1-inactivated sites or HIV-1 LTR (−81 to +77) with NF-κB-deleted sites (courtesy of Dr. Manuel López-Cabrera, Unidad de Biología Molecular, Madrid, Spain). The cells were transfected for 3 h at 37°C using Lipofectamine Plus and then incubated with iron chelators or control medium for 18 h at 37°C. The media was then removed from the cells and equal amounts of 1× PBS and luciferase buffer were added to measure luciferase activity.
Immunoblots and Kinase Assay.
293T cells were incubated with control medium or control medium containing chelators and incubated for 24 h at 37°C. Cells were then washed with PBS, and whole-cell lysates were prepared from the cells using whole-cell lysis buffer (50 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 1% NP-40, and 0.1% SDS) supplemented with antiprotease cocktail (Sigma). Protein concentration of the lysates was measured by the Bradford assay (Bio-Rad Laboratories, Hercules, CA) and 30 to 50 μg of total protein was subjected to electrophoresis using 10% SDS-PAGE. The gels were transferred onto a polyvinylidene fluoride (PVDF) membrane (Millipore) and analyzed for CDK2, CDK9, and cyclin T1. The blots were developed and quantified using a ChemiDoc XRS Station (Bio-Rad Laboratories). Immunoprecipitation assays were carried out with 400 μg of the cell lysate and 800 ng of antibody combined with 50 μl of a 50% slurry of protein A agarose for 2 h at 4°C in TNN buffer (50 mM Tris-HCl, pH 7.5, 0.15 M NaCl, and 1% NP-40). The agarose beads were precipitated, washed with TNN buffer, and divided into two parts and used for the kinase assay and Western blotting. The kinase assay was performed at 30°C for 30 min in a kinase assay buffer (50 mM HEPES-KOH, pH 7.9, 10 mM MgCl2, 6 mM EGTA, and 2.5 mM DTT) containing 1 μg of histone H1 as a substrate, 200 μM ATP, and 5 μCi of [γ-32P]ATP.
C-Terminal Domain of RNA Polymerase II Phosphorylation Assay.
The CTD phosphorylation assay was performed at 30°C for 30 min in 20 μl of kinase assay buffer (50 mM HEPES-KOH, pH 7.9, 10 mM MgCl2, 6 mM EGTA, and 2.5 mM DTT) containing 4 μg of CTD substrate, 250 μM ATP, and 5 μCi of [γ-32P]ATP. Recombinant CDK2/cyclin E or CDK9/cyclin T1 were used as controls. The reaction was stopped with SDS-loading buffer and resolved using 10% PAGE. The dried gel was subjected to filmless autoradiographic analysis. In parallel, the products of immunoprecipitation were resolved using 10% Tris-glycine SDS-PAGE, transferred to PVDF membranes (Millipore), and immunoblotted with appropriate antibodies.
Separation of Large and Small CDK9/Cyclin T1-Containing Complexes by Differential Salt Extraction.
We used a method developed by Biglione et al. (2007) to quantify the relative amounts of large and small CDK9/cyclin T1-containing complexes based on differential nuclear extraction. 293T cells were grown on six-well plates at 37°C and 5% CO2 and then incubated with the iron chelators (1 μM) for 24 h at 37°C. Cells were then suspended in buffer A [10 mM HEPES, pH 7.9, 10 mM KCl, 10 mM MgCl2, 1 mM EDTA, 250 μM sucrose, 1 mM DTT, 0.5% NP-40, and protease inhibitor cocktail (Sigma)] added at 500 μl/107 cells. The mixture was incubated on ice for 10 min and centrifuged at 1000g for 5 min to pellet the nucleus. The supernatant was removed and saved as the low-salt (LS) extract. The remaining pellet was resuspended in buffer B (20 mM HEPES-KOH, pH 7.9, 450 mM NaCl, 1.5 mM MgCl2, 0.5 mM EDTA, 1 mM DTT, and protease inhibitor cocktail) added at 500 μl/107 cells. The mixture was incubated on ice for 10 min and centrifuged at 10,000g for 1 h at 4°C. The supernatant was collected as the high-salt (HS) extract and the pellet was discarded. The lysates were loaded on 10% polyacrylamide gel and transferred to PVDF membrane and western blotting analysis was performed for the expression of CDK9 and cyclin T1.
Results were expressed as mean ± S.D. Data were compared using Student's t test. Results were considered significant when p < 0.05.
Identification of Bp4aT and Bp4eT as the Most Efficient and Least Cytotoxic Inhibitors of HIV-1 Transcription.
We analyzed 15 selected DpT and BpT-derived iron chelators, including a control nonchelating analog, Dp2mT, that was designed and characterized in previous studies (Yuan et al., 2004; Kalinowski et al., 2007) (Fig. 1A and Table 1). These agents were assessed using CEM-T cells containing an integrated HIV-1 LTR-GFP (CEM-GFP). HIV-1 transcription in CEM-GFP cells was induced by infection with adenovirus expressing Tat, and GFP fluorescence was measured to quantify viral transcription, as described previously (Debebe et al., 2007; Nekhai et al., 2007). Data were analyzed using Prism 3 software and showed that after a 24-h incubation of CEM T cells with the 15 chelators tested, 11 ligands inhibited HIV-1 transcription with submicromolar IC50 values (Table 1 and Fig. 1, B and C). The six most active chelators (IC50<0.2 μM) were Dp44mT, Dp4pT, Bp44mT, Bp4eT, Bp4aT, and Bp4pT (Table 1). Bp44mT had the lowest IC50 for inhibition of HIV-1 transcription, namely 15 nM (Table 1 and Fig. 1C).
The short-term toxicity of the iron chelators was analyzed by propidium iodide uptake in the treated CEM T cells in comparison with the untreated control. As a positive control for cell death, untreated cells were lysed by the addition of the nonionic detergent, Triton X-100 [1% (v/v)]. Toxicity of the chelators and that of the noncytotoxic, nonchelating control compound Dp2mT (Yuan et al., 2004) were then analyzed. These results showed that Dp4pT, Bp44mT, Bp4pT, and di-2-pyridylketone thiobenzoyl hydrazone (PKTBH) were clearly cytotoxic over 24 h (see IC50 values in Table 1 and Supplemental Fig. 1). As shown previously, Dp2mT was not cytotoxic at concentrations below 10 μM, because it cannot bind cellular iron pools (Yuan et al., 2004).
Inhibition of One-Round HIV-1 Replication in CEM T Cells.
Based on the results above, we selected Bp4aT and Bp4eT for further analysis as the most potent and least cytotoxic compounds that inhibited HIV-1 transcription in 293T and CEM T cells. We analyzed the effect of Bp4aT and Bp4eT on one round of HIV-1 replication in comparison with the positive and negative controls, Bp4pT and BpT, which showed marked and little activity in inhibiting HIV-1 transcription, respectively (Table 1). In these experiments, CEM T cells were infected with pseudotyped HIV-1 virus containing luciferase in place of the nef gene (HIV-1 Luc) (He and Landau, 1995). CEM T cells were infected overnight with the virus and then cultured for 24 h at 37°C with or without the indicated concentration of chelators (Fig. 2A). In these experiments, Bp4eT was the most efficient inhibitor (IC50 = 75.0 ± 5.5 nM), followed by Bp4aT (IC50 = 160 ± 16 nM), Bp4pT (IC50 = 340 ± 53 nM), and BpT (IC50 = 1.60 ± 0.26 μM) (Fig. 2A).
Inhibition of One-Round HIV-1 Replication in PBMCs.
We also analyzed the effect of Bp4aT and Bp4eT on one round of HIV-1 replication in PBMCs in comparison with BpT. PBMCs were infected with pseudotyped HIV-1 Luc virus for 48 h at 37°C and then cultured for 24 h at 37°C with or without the indicated concentration of chelators (Fig. 2B). In these cells, only Bp4aT markedly inhibited luciferase activity with an IC50 = 25 nM, whereas Bp4eT inhibited HIV-1 replication at a much higher concentration (IC50 = 1.3 μM; Fig. 2B).
Inhibition of HIV-1 Replication in OM10.1 Cells.
Considering the effective antiviral activity observed above on HIV-1 transcription or one-round HIV-1 replication, it was important to compare the efficacy of these chelators on productive HIV-1 replication. To perform these studies, we used HIV-1-infected OM10.1 cells that were treated with TNF-α to analyze the inhibitory effects of iron chelators. Both Bp4eT and Bp4aT inhibited HIV-1 replication, the IC50 being 2.4 ± 0.4 and 6.4 ± 2.7 μM, respectively (Fig. 2C). Thus, the chelators were also able to inhibit productive HIV-1 replication.
Comparison of the Inhibition of HIV-1 Transcription by Bp4aT and Bp4eT with That of ARC and Roscovitin.
Considering the effective antiviral activity of Bp4eT and Bp4aT, it was important to compare the efficacy of these chelators with known HIV-1 transcription inhibitors. We compared inhibition of HIV-1 transcription in 293T cells transfected with HIV-1 LTR-LacZ and Tat expression vectors. Bp4aT and Bp4eT inhibited Tat-induced HIV-1 transcription in 293T cells at low micromolar concentrations (Fig. 3A). To compare the effect of the iron chelators with the effect of known HIV-1 transcription inhibitors, we incubated these transfected cells with ARC or roscovitin, which inhibit HIV-1 transcription through the deregulation of CDK9 and CDK2, respectively (for review, see Nekhai and Jeang, 2006). After a 24-h incubation at 37°C, HIV-1 transcription was inhibited by ARC (IC50 = 180 nM) and by roscovitin (IC50 = 1.5 μM) (Fig. 3B). Thus, the iron chelators, Bp4aT and Bp4eT, were approximately 4- to 16-fold less efficient inhibitors of HIV-1 than ARC but were comparable to roscovitin. However, it is relevant to note that these chelators have not been specifically designed as antiviral agents, and their potential may lie in their novel mechanism of action (examined under Inhibition of CDK9).
Toxicity of Iron Chelators.
Analysis of short-term cytotoxicity of Bp4aT and Bp4eT was performed in comparison with the negative control chelator, BpT, which showed little toxicity after 24 h (Table 1). In these studies using calcein-AM uptake (Papadopoulos et al., 1994), the chelators were demonstrated to be equally nontoxic after a 24-h incubation period (IC50 >100 μM; Fig. 4A). We further analyzed long-term cytotoxicity of Bp4aT and Bp4eT in comparison with the negative control BpT in CEM T cells after 1 to 5 days of incubation implementing the calcein-AM assay. The number of viable cells decreased after 2-day treatment and remained low for up to 5 days of incubation (Fig. 4B). Analysis of cell growth using the trypan blue viability assay showed that although control cells continued to grow, cells in the presence of chelators did not proliferate (Supplemental Fig. 2A). Analysis of the long-term cytotoxicity of Bp4aT, Bp4eT, and BpT in PBMCs at 0.3 μM showed that Bp4aT was more toxic compared with Bp4eT or BpT (Fig. 4C). These results showed that although iron chelators were not toxic during the 24-h incubation when the antiviral effect was observed, longer incubation times demonstrated cytotoxicity. Taken together, the results above demonstrated that among the analyzed iron chelators, Bp4eT was the most promising antiviral agent.
Examination of the Effect of Chelators on Sp1- and NF-κB-Driven Transcription.
Sp1 activity is important for HIV-1 basal transcription (Jochmann et al., 2009) and Tat-induced HIV-1 transcription (Chun et al., 1998). NF-κB is activated in various cell types in response to hypoxia and is important for Tat-activated HIV-1 transcription, because it acts in concert with Tat and CDK9/cyclin T1 (West et al., 2001). Considering this, to determine whether iron chelators have an effect on NF-κB- and Sp1-driven activity, we analyzed their effect on the NF-κB-dependent e-selectin promoter as a relevant positive control (Fu et al., 2003). This was then compared with the effects of the chelators on the activity of HIV-1 promoters with the deletion of NF-κB sites or inactivation of Sp1 sites in 293T cells (Gómez-Gonzalo et al., 2001). Bp4eT at 10 μM inhibited the activity of WT HIV-1 LTR in the HIV-1 proviral pNL4-3.Luc.R−E− (pNL-Luc) promoter in 293T cells (Fig. 5A). In contrast, the activity of the NF-κB-dependent e-selectin promoter (Fu et al., 2003) was not affected in 293T cells incubated with this chelator (Fig. 5A). This suggested that NF-κB activity was not changed after incubation of cells with iron chelators. It should be noted that analysis of cytotoxicity in 293T cells by the calcein-AM uptake assay showed that Bp4eT (10 μM) did not exhibit significant toxicity compared with the control after 24 h (Supplemental Fig. 2B).
In the absence of Tat, basal activity of the WT HIV-1 LTR promoter (denoted WT SP KB) was reduced by 2.7-fold in 293T cells incubated with 10 μM Bp4eT (Fig. 5B). In contrast, the activities of the HIV-1 LTR without NF-κB sites (WT SP) or with inactivated Sp1 sites (WT SP 123) were reduced by only 1.3- and 1.4-fold, respectively (Fig. 5B). These results indicated that inactivation of Sp1 sites or deletion of NF-κB sites makes the remaining promoter activity less sensitive to Bp4eT. Thus, the effect of the iron chelators on basal HIV-1 transcription might involve deregulation of a factor other than Sp1 or NF-κB. Other possible factors that could be affected by the chelators include CDK9/cyclin T1, and this is assessed under Inhibition of CDK9.
Inhibition of CDK2.
We have demonstrated previously that 311 and ICL670 inhibited CDK2 activity (Debebe et al., 2007). To determine whether the most effective BpT and DpT chelators had a similar effect, CDK2 was immunoprecipitated from 293T cells incubated for 24 h at 37°C with Dp44mT, Bp4aT, or Bp4eT and also the positive control chelator 311 or ICL670 (all at 10 μM except ICL670, which was used at 100 μM because of its low efficacy; Fig. 6A). These studies showed that Bp4eT, Dp44mT, and particularly Bp4aT were significantly (p < 0.001) more active than 311 or ICL670 at inhibiting CDK2 kinase activity (Fig. 6A).
Next, we analyzed whether the inhibition of CDK2 activity parallels the inhibition of HIV-1 transcription using 293T cells. CDK2 was immunoprecipitated from 293T cells incubated with three different concentrations of Bp4eT (0.1, 1, and 10 μM), and its activity was analyzed with histone H1 as a substrate. Histone H1 phosphorylation was reduced in a dose-dependent manner, with approximately 50% inhibition at 1 μM and 85% inhibition at 10 μM (Fig. 6, B and C), being consistent with the inhibition of HIV-1 transcription (IC50 = 2 μM; Fig. 3A).
Inhibition of CDK9.
We then analyzed whether CDK9 activity was affected in 293T cells incubated with iron chelators (all at 10 μM except ICL670A at 100 μM) for 24 h at 37°C. This was important to assess because CDK9/cyclin T1 is recruited by HIV-1 Tat to the HIV-1 promoter and is crucial for viral replication (Nekhai and Jeang, 2006). Cyclin T1 was immunoprecipitated (Fig. 7A) and the activity of coprecipitated CDK9 was assayed with GST-CTD as a substrate. The hyperphosphorylated C-terminal domain of RNA polymerase II (CTDo) that contains more than three phosphorylated serine residues migrates as a higher molecular weight band on SDS-PAGE (Fig. 7A). In contrast, the hypophosphorylated C-terminal domain of RNA polymerase II (CTDa), which appears as a lower molecular weight band on SDS-PAGE (Fig. 7A), comigrates with the nonphosphorylated CTD of RNA Polymerase II. CTD phosphorylation by CDK9/cyclin T1 resulted in formation of CTDo (Fig. 7A, lane 2). There was a significant (p < 0.005) decrease in the CTD kinase activity of CDK9/cyclin T1 in cells incubated with all the chelators (Fig. 7A).
It is noteworthy that we could not detect CDK9 that coprecipitated with cyclin T1 in cells incubated with most iron chelators, except for Bp4aT and Dp44mT, where only very faint bands were observed (Fig. 7A). Analysis of the kinase activity of CDK9 precipitated with anti-CDK9 antibody showed a marked inhibition of CDK9 activity (as demonstrated by a decrease in the intensity of the CTDo and CTDa bands) in the cells incubated with Bp4eT (10 μM) compared with untreated control cells or cells incubated with the negative control, Dp2mT (10 μM; Fig. 7B).
We then determined whether the effect of CDK9 inhibition could be due to changes in its interaction with cyclin T1 and distribution between small CDK9/cyclin T1 complexes and large CDK9/cyclin T1/HEXIM1/7SK RNA complexes. We analyzed the distribution of CDK9 and cyclin T1 between the small and large complexes using the procedure described by Biglione et al. (2007). Using this method, large and small complexes are sequentially extracted with LS and HS concentrations. Levels of CDK9 after incubating 293T cells with Bp4eT (1 μM) for 24 h at 37°C were lower than that of the control in both LS and HS extracts, whereas cyclin T1 levels were reduced only in the HS fraction (Fig. 8, A and B). These results indicate that the overall level of CDK9 expression is reduced. Furthermore, the amount of high molecular weight CDK9/cyclin T1 complex, which supplies CDK9/cyclin T1 for HIV-1 transcription (Biglione et al., 2007), was reduced after incubation with Bp4eT as demonstrated by the reduction of CDK9 in this complex (Fig. 8B).
Taken together, our data show that HIV-1 transcription is efficiently inhibited by Bp4eT- and Bp4aT-based iron chelators without apparent cytotoxicity over 24 h. This mechanism of inhibition may include the inhibition of CDK9/cyclin T1 activity, in addition to the previously shown inhibition of CDK2 activity (Debebe et al., 2007).
Iron is required for several steps in the HIV-1 life cycle, including reverse transcription, HIV-1 gene expression, and capsid assembly (Drakesmith and Prentice, 2008). Therefore, the use of iron chelators as a therapeutic strategy has potential and needs to be carefully assessed. Here, we focused on the effect of iron chelators on HIV-1 transcription and identified Bp4aT and Bp4eT as the two most promising agents that inhibited HIV-1 transcription in cultured cell lines and primary cells. Bp4eT was relatively nontoxic after a 24-h incubation and could be the most suitable candidate as a potential HIV-1 inhibitor. However, only Bp4aT effectively inhibited HIV-1 transcription in primary PMBCs and thus should also be further considered.
CDK2, a previously identified regulator of HIV-1 transcription (for review, see Nekhai and Jeang, 2006), was found to be a molecular target of the BpT and DpT chelators. Thus, CDK2 inhibition could be partly responsible for the antiviral activity of the chelators. Inhibition of CDK2 by either pharmacological inhibitors or CDK2-directed small interfering RNA has been shown to suppress HIV-1 replication (Agbottah et al., 2005; Ammosova et al., 2005). Iron chelators affect the activity or expression of CDK2 (Gao and Richardson, 2001; Debebe et al., 2007; Pahl et al., 2007), and deregulation of CDK2/cyclin E activity by these agents arrests cell cycle progression (Lucas et al., 1995; Pahl et al., 2000). The iron chelator desferriexochelin inhibits the binding of cyclin A and cyclin E to CDK2 in human mammary epithelial cells, whereas in human breast cancer cells, the binding is increased (Pahl et al., 2007). On the other hand, the ligand 311 inhibits the expression of cyclins D1, D2, D3, A, and B1 and also CDK2, but not cyclin E (Gao and Richardson, 2001). The chelator-mediated mechanism of inhibition of CDK2 activity may include deregulation of CDK2 binding to cyclin E or cyclin A and increased binding of the CDK inhibitor p21 to the CDK2/cyclin A complex. Considering this, iron chelators have been shown to down-regulate p21 protein expression in some cell types (Fu and Richardson, 2007) but up-regulate it in others, the effect being dependent on the cell-type assessed (R. Moussa and D. R. Richardson, unpublished data). Hence, further studies are required to assess the role of p21 in the mechanism of CDK2 inhibition by iron chelators.
In the current study, we demonstrated that the BpT- and DpT-based iron chelators can inhibit CDK2 activity. It is noteworthy that this investigation showed that in addition to the inhibition of CDK2, the chelators also inhibit CDK9/cyclin T1 activity that is recruited by HIV-1 Tat-induced HIV-1 transcription (Nekhai and Jeang, 2006). CDK9/cyclin T1 is present in cells within a large complex that contains 7SK RNA and the hexamethylene bisacetamide-induced protein and also a small complex (for review, see Michels and Bensaude, 2008). The activity of CDK9 in the large complex is inhibited by the interaction with 7SK RNA and hexamethylene bisacetamide-induced protein. The CDK9 inhibitors flavopiridol and roscovitin reduce the formation of the large complex at a concentration that inhibits HIV-1 (Biglione et al., 2007). This suggests that the large complex is important for the activation of HIV-1 transcription and may supply CDK9/cyclin T1 for Tat recruitment. We demonstrate here that CDK9/cyclin T1-kinase activity is markedly inhibited in cells incubated with Bp4aT, Bp4eT, or Dp44mT, thus providing a mechanistic explanation for the inhibitory effects of these agents on HIV-1 replication. Our analysis of the large and small complexes in the cells incubated with iron chelators showed a reduction of the CDK9, but not cyclin T1, in the high molecular weight CDK9/cyclin T1 complex. Hence, the molecular mechanism involved in preventing HIV-1 transcription by chelators could involve their targeting and inhibition of CDK9/cyclin T1 activity in addition to CDK2 activity.
HIV-1 basal transcription is largely regulated by the Sp1 transcription factor (Jochmann et al., 2009), whereas in Tat-activated transcription, NF-κB plays an important regulatory role by acting in concert with Tat and CDK9/cycin T1 (West et al., 2001). Our analysis showed that transcription from the NF-κB-dependent e-selectin promoter was not changed in cells incubated with Bp4eT, thus excluding NF-κB as a target for iron chelators. However, because inactivation of Sp1 as well as NF-κB sites equally decreased sensitivity of the HIV-1 promoter to iron chelator treatment, it is possible that these agents cooperatively affect both the NF-κB and Sp1 sites. Because NF-κB acts in concert with CDK9/cyclin T1 (West et al., 2001) and Sp1 targets CDK9/cyclin T1 to the HIV-1 promoter in the absence of Tat (Yedavalli et al., 2003), the inactivation of CDK9/cyclin T1 might affect both NF-κB and Sp1-driven transcription of the HIV-1 promoter.
Considering the potential mechanism of the anti-HIV activity of the chelators and the possible role of CDK2 as their molecular target, it is significant to consider the following studies. CDK2 knockout mice are viable, although CDK2 is required for germ cell development, because CDK2(−/−) mice are sterile (Berthet et al., 2003). In the absence of CDK2, CDK1 compensates for its loss (Satyanarayana et al., 2008). However, the impairment of DNA repair activity makes CDK2(−/−) mice more sensitive to lethal irradiation (Satyanarayana et al., 2008). It is noteworthy that although CDK2 activity is needed for proliferation of cultured lymphocytes, proliferation of human marrow cells is driven by CDK1 combined with either cyclin A or cyclin B (Xie et al., 2008). Thus, CDK2 may be dispensable for proliferation and survival of some cells in vivo and, if so, inhibitors of CDK2 might be considered as a novel class of anti-HIV-1 therapeutics.
In addition to the inhibition of HIV-1 transcription, iron depletion may affect other steps in HIV-1 replication. During viral entry, HIV-1 replication is dependent on the activity of host cell ribonucleotide reductase that contains nonheme iron, which is important for enzymatic activity (Tsimberidou et al., 2002). Export of unspliced HIV-1 mRNA requires HIV-1 Rev protein and host elongation factor 5A (eIF5α). The eIF5α protein contains N-ε-4-amino-2-hydroxybutyl-lysine (hypusine) that is produced by deoxyhypusine hydroxylase, an iron-containing enzyme (Kim et al., 2006). The topical fungicide ciclopirox (which also is an iron chelator) and the orally active chelator deferiprone inhibit HIV-1 gene expression interfering with the hydroxylation step in the hypusine modification of eIF5α (Hoque et al., 2009). Furthermore, assembly of the HIV capsid requires an ATP-binding protein, ABCE1, which contains iron-sulfur clusters (Barthelme et al., 2007) and binds to HIV-1 Gag protein (Zimmerman et al., 2002). Thus, iron chelators may have a wider spectrum of molecular targets and further studies are required to analyze these.
To assess their possible utility as pharmaceuticals against HIV-1, we compared the antiviral activity of the most efficient chelators, Bp4aT and Bp4eT, with the previously reported HIV-1 inhibitors, ARC, and roscovitin (Agbottah et al., 2005; Nekhai et al., 2007). These studies showed that the ligands were approximately 4- to 10-fold less efficient inhibitors than ARC (Fig. 3) but very similar to roscovitin. However, the fact that these chelators were not specifically designed for the treatment of HIV-1 demonstrates their clear potential. Further chemical optimization of the design of these ligands for the purpose of inhibition of HIV-1 is vital. Clearly, the good therapeutic index in short-term infectivity assays shown in the current studies must be followed with extensive in vivo assessment to ensure high tolerability, which is critical for human use (Biglione et al., 2007). Finally, it must be noted that the very different mechanism of action of these chelators compared with other established antiviral therapies is an advantage of these agents. Hence, they may be useful in terms of treating resistance to AZT and other commercially available antiviral drugs that is a major emerging problem.
Significantly, previous in vivo studies in mice examining the effects of Dp44mT have shown that it neither leads to decreased levels of tissue iron nor induces alterations in multiple hematological parameters (Yuan et al., 2004; Whitnall et al., 2006). This finding is probably due to the very low doses of the agents (0.4–0.7 mg/kg) required for efficacy. Moreover, Dp44mT was well tolerated at optimal doses, demonstrating its potential for clinical use (Yuan et al., 2004; Whitnall et al., 2006). Given the levels of cytotoxicity observed after 3 days in vitro in the current studies, questions could arise as to the feasibility of the development of these compounds as antivirals. However, many common chemotherapeutic drugs currently in clinical use (e.g., doxorubicin) display marked cytotoxicity profiles in vitro (Yuan et al., 2004) but are tolerated at appropriate doses in vivo and have led to vast improvements in a variety of disease states. Hence, clearly, the effect in vitro is difficult to directly translate to the in vivo situation and, as observed with many other drugs, appropriate dosing is key to optimizing antiviral activity and minimizing toxicity in vivo.
In conclusion, these iron chelators represent agents with a therapeutic window and novel mechanism of action that deserves further investigation in terms of their use in future anti-HIV-1 therapeutic regimens. Their potential may be particularly significant where resistance to standard antiviral therapy occurs.
Participated in research design: Debebe, Ammosova, Breuer, Jerebtsova, Kashanchi, and Nekhai
Conducted experiments: Debebe, Ammosova, Breuer, Kashanchi, and Nekhai.
Contributed new reagents or analytic tools: Lovejoy, Kalinowski, Jerebtsova, Ray, and Richardson.
Performed data analysis: Debebe, Ammosova, Breuer, and Nekhai.
Wrote or contributed to the writing of the manuscript: Debebe, Gordeuk, Kalinowski, Lovejoy, Richardson, and Nekhai.
Other: Debebe, Gordeuk, Kalinowski, Lovejoy, Richardson, and Nekhai acquired funding for the research.
We acknowledge Dr. Anna Suter (Novartis, Pharma AG, Ltd., Basel, Switzerland) for ICL670. We appreciate the assistance of Keon Combi, Xiomei Niu, and Dillon Robinson in cell counting and toxicity assays. We are also thankful to Dr. Kuan-Teh Jeang (National Institute of Allergy and Infectious Diseases, National Institutes of Health) for critical reading and comments on the manuscript. Finally, Dr. Katie Dixon, Zaklina Kovacevic, and Dr. Helena Mangs (Iron Metabolism and Chelation Program, University of Sydney, Sydney, Australia) are acknowledged for thorough assessment of the manuscript.
This project was supported by the National Institutes of Health National Heart, Lung, and Blood Institute and Office of Research on Minority Health [Grants 2R25-HL003679-08,R-1-HL55605, F31-HL090025]; the National Institutes of Health National Center for Research Resources [Grant 2M01-RR10284]; the National Institutes of Health National Institute of General Medical Sciences [1SC1-GM082325]; the National Institutes of Health Research Centers in Minority Institutions (RCMI) Program of the Division of Research Infrastructure, National Center for Research Resources [Grant 2G12-RR003048]; the National Institutes of Health National Institute of Diabetes and Digestive and Kidney Diseases [R01-DK49419]; the National Institutes of Health National Center for Complimentary and Alternative Medicine [R21AT002278]; and by the National Health and Medical Research Council of Australia [Senior Principal Research Fellowship 571123, Project Grant 570952].
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
- reverse transcriptase
- 1,6-dimethyl-3-hydroxy-4-(1H)-pyridinone-2-carboxy-(N-methyl)-amide hydrochloride
- 2-hydroxy-1-naphthylaldehyde isonicotinoyl hydrazone
- cyclin-dependent kinase
- di-2-pyridylketone thiosemicarbazone
- 2-benzoylpyridine thiosemicarbazone
- di-2-pyridylketone 4,4-dimethyl-3-thiosemicarbazone
- long terminal repeat
- green fluorescent protein
- peripheral blood mononuclear cell
- nuclear factor-κB
- wild type
- acetoxymethyl ester
- propidium iodide
- enhanced green fluorescent protein
- polyvinylidene difluoride
- polyacrylamide gel electrophoresis
- C-terminal domain
- low salt
- high salt
- di-2-pyridylketone 2-methyl-3-thiosemicarbazone
- 2-benzoylpyridine 4,4-dimethyl-3-thiosemicarbazone
- 2-benzoylpyridine 4-ethyl-3-thiosemicarbazone
- 2-benzoylpyridine 4-allyl-3-thiosemicarbazone
- 2-benzoylpyridine 4-phenyl-3-thiosemicarbazone
- 6-dimethyl-3-hydroxy-4-(1H)-pyridinone-2-carboxy-(N-methyl)-amide hydrochloride
- di-2-pyridylketone thiobenzoyl hydrazone
- hyperphosphorylated C-terminal domain of RNA polymerase II
- hypophosphorylated C-terminal domain of RNA polymerase II
- host elongation factor 5A.
- Received September 26, 2010.
- Accepted October 18, 2010.
- Copyright © 2011 The American Society for Pharmacology and Experimental Therapeutics