|Home | About | Journals | Submit | Contact Us | Français|
Triciribine (TCN) is a tricyclic nucleoside that inhibits human immunodeficiency virus type 1 (HIV-1) replication by a unique mechanism not involving the inhibition of enzymes directly involved in viral replication. This activity requires the phosphorylation of TCN to its 5′ monophosphate by intracellular adenosine kinase. New testing with a panel of HIV and simian immunodeficiency virus isolates, including low-passage-number clinical isolates and selected subgroups of HIV-1, multidrug resistant HIV-1, and HIV-2, has demonstrated that TCN has broad antiretroviral activity. It was active in cell lines chronically infected with HIV-1 in which the provirus was integrated into chromosomal DNA, thereby indicating that TCN inhibits a late process in virus replication. The selection of TCN-resistant HIV-1 isolates resulted in up to a 750-fold increase in the level of resistance to the drug. DNA sequence analysis of highly resistant isolate HIV-1H10 found five point mutations in the HIV-1 gene nef, resulting in five different amino acid changes. DNA sequencing of the other TCN-resistant isolates identified at least one and up to three of the same mutations observed in isolate HIV-1H10. Transfer of the mutations from TCN-resistant isolate HIV-1H10 to wild-type virus and subsequent viral growth experiments with increasing concentrations of TCN demonstrated resistance to the drug. We conclude that TCN is a late-phase inhibitor of HIV-1 replication and that mutations in nef are necessary and sufficient for TCN resistance.
The virally encoded accessory proteins (Vpr, Vpu, Vif, Nef) are not required for replication in certain cell lines in vitro, but they are important factors causing the clinical manifestations of the pathogenesis of human immunodeficiency virus (HIV) (59). Viral protein R (Vpr) plays an important role in regulating nuclear import for the preintegration complex (PIC), is required for integration/replication in nondividing cells, induces cell cycle arrest in proliferating cells, stimulates viral transcription, and regulates apoptosis in infected cells (9). Viral protein U (Vpu) is an integral membrane phosphoprotein responsible for the degradation of CD4 in the endoplasmic reticulum and enhances virion release from the cell surface by antagonizing the cellular protein tetherin (41, 57). The viral infectivity factor (Vif) is required for replication in nonpermissive cell lines that express APOBE3G but is dispensable for replication in permissive cell lines that do not express APOBEC3G (26, 58). The “negative” factor (Nef) actually enhances viral replication and is necessary for pathogenesis to AIDS (32). Nef is the first viral protein to accumulate to detectable levels in cells following infection and is the most versatile of the accessory proteins (53). It interacts with several cellular pathways that influence HIV replication and immune evasion, resulting in the downregulation of CD4 and major histocompatibility complex class I molecules (25, 38, 51), promotion of the synthesis of FasL (a proapoptotic ligand that initiates apoptosis in nearby cells) (67), and the inhibition of both p53 (a regulator of apoptosis) and ASK-1 (the kinase involved in the initiation of apoptosis) (16, 20, 24). Nef also binds to several signaling proteins, including the Src family of tyrosine kinases (Erk-1, Raf1, and PKC theta), resulting in the activation of the cellular pathways responsible for proliferation and survival (22, 50). As important factors for the manifestation of HIV pathogenesis, these proteins are potential targets for antiretroviral therapy.
Triciribine (TCN; Fig. Fig.1)1) is a tricyclic nucleoside originally synthesized by Schram and Townsend as a potential anticancer drug (56). It is unique compared to naturally occurring purine nucleosides, in that it contains a tricyclic base instead of a bicyclic base. It has been studied broadly both as an antineoplastic agent and as an antiviral active against HIV (37, 65). Recent reports have provided new and unique insight into the antineoplastic activity of TCN. The studies described in those reports demonstrated that inhibition of the kinase activity and the level of phosphorylation of protein kinase B (Akt) are factors that result in the suppression of cell growth and the induction of apoptosis in human cancer cells that harbor constitutively overexpressed levels of Akt (68). Both the antineoplastic and antiviral activities of TCN require that the compound be phosphorylated to its corresponding 5′ monophosphate (TCN-P) (45-48, 64). This phosphorylation occurs intracellularly and is mediated by adenosine kinase (4). TCN is not converted to higher phosphates, nor is it incorporated into nucleic acids (66). Therefore, TCN-P must be the active form of the compound.
As an antiviral, we have found that TCN is active against laboratory and clinical isolates of HIV type 1 (HIV-1) and HIV-2, including strains that are resistant to the reverse transcriptase (RT) inhibitors zidovudine (AZT) and 4,5,6,7-tetrahydro-5-methylimidazo [4,5,1-jk][1,4]benzodiazepin-2(1H)-one (TIBO) (37). Those studies were performed with a variety of cell lines, including macrophage and human peripheral blood lymphocytes. This activity occurs at concentrations up to 1,000-fold below those that are cytotoxic for uninfected cells (37). Mechanism-of-action studies have demonstrated that TCN is not an inhibitor of the virally encoded enzymes reverse transcriptase (37), integrase, RNase H, or protease (49). These observations establish that TCN inhibits HIV-1 replication by a unique mechanism that is different from the mechanisms of action of the antiviral compounds that are currently available for the treatment of HIV-1-infected individuals.
The spread of HIV and the ever increasing occurrence of resistance to the available drug therapies have increased the need for the development of novel inhibitors of virus replication. Due to the unique nature by which TCN inhibits HIV-1 replication, we have continued to study its effects on virus replication. Herein we describe the antiviral activity of TCN, as tested with peripheral blood mononuclear cells (PBMCs), against a new panel of low-passage-number clinical isolates and describe the observations that have helped to elucidate the mechanism of action by which TCN inhibits virus replication.
AA-5, CEM-SS, H9, U1, and ACH-2 cells were obtained from the NIH/NIAID AIDS Research and Reference Reagent Program (Bethesda, MD) and were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM l-glutamine, penicillin (100 U/ml), and streptomycin (100 μg/ml). HeLa-CD4-LTR-ß-Gal cells were also obtained from the NIH/NIAID AIDS Research and Reference Reagent Program and were maintained in Dulbecco's modified Eagle medium supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), l-glutamine (2 mM), G418 (200 μg/ml), and hygromycin (100 μg/ml). Human PBMCs and monocyte-derived macrophages (MDMs) were isolated from hepatitis virus- and HIV-seronegative donors by Ficoll-Hypaque gradient centrifugation, as described previously (7). Antiviral assays were accomplished with 3-day-old phytohemagglutinin- and interleukin-2-stimulated PBMCs or MDMs cultured for 6 days. All antiviral evaluations were performed in triplicate in RPMI 1640 medium supplemented with 10% FBS, l-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). HIV replication was determined by measurement of the reverse transcriptase activity in the supernatant (6) or by enzyme-linked immunosorbent assay (ELISA; Beckman Coulter, Somerset, NJ) for p24 antigen expression 6 days postinfection, as indicated throughout the text. Cell viability by formazan dye reduction was determined with the CellTiter reagent (Promega, Madison, WI).
The following virus isolates were obtained from the NIH/NIAID AIDS Research and Reference Reagent Program: HIV-1 laboratory isolate IIIB (HIV-1IIIB); monocytropic isolates HIV-1 Ba-L and ADA; clinical isolates RW/92/016 (subgroup A), 302056 (subgroup B), BR/92/025 (subgroup C), UG/92/046 (subgroup D), CMU02 (subgroup E), BR/93/020 (subgroup F), JV 1083 (subgroup G), BCF01 (subgroup O), BCF02 (subgroup O), and BCF03 (subgroup O); HIV-2 isolates ROD, CDC 310319, and CDC 310342; and simian immunodeficiency virus (SIV) Mac251. SIV 69R was isolated from a rhesus macaque infected with SIVmne/e11s and was obtained from Mark Lewis (Henry M. Jackson Foundation). Multidrug-resistant HIV-1 isolate MDR 769 was obtained from Thomas C. Merigan (Stanford University). This isolate is resistant to the reverse transcriptase inhibitors zidovudine (AZT), didanosine, lamivudine, stavudine, foscarnet, and nevirapine (i.e., it has the resistance mutations M41L, K65R, D67N, V75I, F116Y, Q151M, Y181I, L210W, and T215Y) and to the protease inhibitors indinavir, saquinavir, and nelfinavir (i.e., it has resistance mutations L10I, M36M/V, M46I, I54V, L63P, A71V, V82A, I84V, and L90M) (47). Low-passage-number pediatric clinical isolates HIV-1ROJO, HIV-1SLKA, HIV-1WEJO, and HIV-1TEKI were presumed to be subgroup B isolates and were derived in the laboratories of Southern Research Institute, as described previously (8). Clinical isolates ROJO and WEJO have been typed as syncytium inducing (SI) in MT-2 cells, and clinical isolates SLKA and TEKI have been typed as non-syncytium inducing (NSI) in MT-2 cells. The SI and NSI phenotypes have been correlated with lymphocyte (CXCR4) and monocyte (CCR5) tropism, respectively, and these viruses have been found to favor the corresponding coreceptor for infection. Clinical isolate HIV-1SK-1 was isolated in the laboratory of Miles Cloyd (Duke University Medical Center) by cocultivation and serial passage in H9 cells, as described previously (10). Aliquots of all virus isolates at predetermined titers were thawed immediately prior to use by rapid thawing in a biological safety cabinet.
TCN was synthesized in the laboratories of L. B. Townsend, as described previously (56). AZT was purchased from Sigma. Stocks were prepared by dissolving compounds in dimethyl sulfoxide (DMSO) and were stored at −20°C. Compounds were added to the cultures so that the resulting concentrations of DMSO did not exceed 0.5% (by volume).
The reverse transcriptase activity in cell-free supernatants was measured as described previously (6). Briefly, tritiated dTTP ([3H]TTP, 80 Ci/mmol; NEN) was received in 1:1 distilled H2O (dH2O)-ethanol at 1 mCi/ml. Template-primer poly(rA)-oligo(dT) (Pharmacia) was prepared as a stock solution by combining 150 μl poly(rA) (20 mg/ml) with 0.5 ml oligo(dT) (20 units/ml) and 5.35 ml sterile dH2O, followed by aliquoting (1.0 ml) and storage at −20°C. The RT reaction buffer was prepared fresh on a daily basis and consisted of 125 μl 1.0 M EGTA, 125 μl dH2O, 125 μl 20% Triton X-100, 50 μl 1.0 M Tris (pH 7.4), 50 μl 1.0 M dithiothreitol, and 40 μl 1.0 M MgCl2. The final reaction mixture was prepared by combining 1 part [3H]TTP, 4 parts dH2O, 2.5 parts poly(rA)-oligo(dT) stock, and 2.5 parts reaction buffer. Ten microliters of this reaction mixture was placed in a round-bottom microtiter plate, and 15 μl of virus containing supernatant was added and mixed. The plate was incubated at 37°C for 60 min. Following the reaction, the reaction volume was spotted onto DE81 filter mats (Wallac); and the mats were washed five times for 5 min each time in a 5% sodium phosphate buffer, two times for 1 min each time in distilled water, and two times for 1 min each time in 70% ethanol and then dried. The radioactivity incorporated (in counts per minute) was quantified by standard liquid scintillation techniques.
ELISA kits were purchased from Beckman Coulter. The assay was performed according to the manufacturer's instructions. Control curves were generated for each assay to accurately quantify the amount of p24 antigen in each sample. Data were obtained by spectrophotometric analysis at 450/570 nm with a Molecular Devices plate reader for determination of the maximum rate of metabolism (Vmax). The final concentrations were calculated from the optical density values by using the Molecular Devices SoftMax Pro software package.
At assay termination, the assay plates were stained with the soluble tetrazolium-based dye 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS; CellTiter reagent; Promega), to determine cell viability and quantify compound toxicity. MTS is metabolized to a soluble formazan product by mitochondrial enzymes of metabolically active cells, thereby allowing the rapid quantitative analysis of cell viability and compound cytotoxicity. This reagent is a single stable solution that does not require preparation before use. At the termination of the assay, 20 μl of the MTS reagent was added to each well. The wells were incubated for 4 h at 37°C. Adhesive plate sealers were used in place of lids, the sealed plate was inverted several times to mix the soluble formazan product, and the plate was read spectrophotometrically at 490/650 nm with a Molecular Devices Vmax plate reader.
Virus infectivity was assayed by titration on HeLa-CD4-LTR-ß-Gal cells (34). Following incubation of the HeLa-CD4-LTR-ß-Gal cells with cell-free supernatants collected from drug-treated, HIV-infected cells, virus infectivity was detected as the transactivation of the long terminal repeat (LTR)-driven ß-galactosidase (ß-Gal) gene by chemiluminescence by use of a single-step lysis and detection method (Gal-screen; Tropix, Bedford, MA).
Supernatant stocks of wild-type HIV-1IIIB were used to infect CEM-SS cells, and TCN was added at a starting concentration of 0.05 μM (about twice the concentration of the drug required to inhibit virus production by 50% [IC50]). The supernatant virus produced in this culture was monitored for RT activity over time. When it reached >40,000 cpm, it was used to infect fresh CEM-SS cells and the concentration of TCN was increased to 0.25 μM. The supernatant virus was sequentially passaged in this manner, and the TCN concentration was increased from two- to fivefold with each passage until, after 18 months, the virus was successfully growing in 160 μM TCN. Several drug-resistant clones were isolated at selected levels of resistance by the method of Klein (35). Clone H10 was the last clone selected, which was at approximately 18 months. All clones were tested for TCN resistance by a standard assay that measured the supernatant RT after 6 days of incubation.
The HIV-1H10 nef-coding region containing mutations associated with TCN resistance was transferred into plasmid pNL4-3 by using the unique XhoI restriction enzyme site located at the 5′ end of nef in the NL4-3-coding sequence. pNL4-3 was digested with XhoI and BsrBI (which cuts in the vector 3′ to the viral coding sequence). The analogous section of HIV-1H10 (previously cloned into another plasmid) was generated by digestion with XhoI and HpaI, and the resulting fragment was ligated to the XhoI- and BsrB1-digested pNL4-3 backbone to generate a new plasmid, pTAK-1. Plasmids pNL4-3 and pTAK-1 were transfected into CEM-SS cells by using DEAE-dextran76, and the resultant viruses (HIV-1pNL4-3 and HIV-1pTAK-1) were tested for resistance to TCN.
Antiviral data and toxicity data are reported as the IC50 and the concentration of drug required to reduce cell viability by 50% (TC50), respectively, as calculated by linear regression analysis with an in-house computer program. Unless otherwise indicated, values represent the results from a single assay. For experiments performed more than once, values represent the mean ± 1 standard deviation. Except in assays with chronically/latently infected cell lines and in assays for HIV-1 MDR 769, in which AZT is ineffective, the HIV reverse transcriptase inhibitor AZT (Sigma) was used as a positive control for the validation of all assays. Dextran sulfate (Sigma) was used as a positive control for the experiments with HIV-1 MDR 769.
In order to extend previous evaluations of TCN for potential therapeutic use, it was tested against a panel of low-passage-number clinical isolates of HIV-1, HIV-2, and SIV by using PBMC cultures. Included in this panel were representative isolates from HIV-1 group M subtypes A, B, C, D, E, F, and G; HIV-1 group O; as well as a multidrug-resistant HIV-1 isolate. With the exception of two of the HIV-1 group O isolates (isolates BCF01 and BCF02), TCN was highly active against all the virus isolates tested, with the IC50s ranging from 0.002 to 0.3 μM (Table (Table1).1). Most notably, TCN was active against multidrug-resistant isolate MDR 769 (IC50, 0.08 μM), indicating its possible utility for the treatment of patients that have failed other therapies due to HIV-1 drug resistance. Furthermore, TCN was found to be nontoxic to PBMCs up to the high-test concentration tested (10 μM) in these assays. In fact, in earlier studies with PBMC cultures, TCN was found to have a TC50 of >31 μM (data not shown), indicating a potential therapeutic index (TI) of >95 to >15,625, depending on the virus isolate.
Since previous studies indicated that TCN is not an inhibitor of any of the well-characterized, virally encoded enzymes (RT, RNase H, integrase, and protease) (37, 49), TCN was further tested in CEM-SS and H9 cells chronically infected by HIV-1. TCN had significant activity against HIV-1 clinical isolate SK-1 in both CEM-SS and H9 cells chronically infected with the virus (Table (Table2).2). Furthermore, TCN also inhibited HIV-2ROD replication in chronically infected CEM-SS cells. The fact that TCN retained significant activity against virus in these chronically infected cell lines is consistent with previous results (37) and suggests that TCN may inhibit a late-stage process in HIV-1 replication at a point after provirus integration.
The infectivity of virus particles released from chronically infected cells treated with TCN was also studied. H9 cells chronically infected with HIV-1SK-1 were treated with TCN, and supernatants from the cells were collected and assayed for RT and p24. The samples were normalized for p24, and the infectivity of the virus in the supernatants was determined. The data in Fig. Fig.22 demonstrate that TCN inhibited virus production (RT and p24) but did not have a significant effect on the infectivity of the reduced amount of HIV-1 that was released from chronically infected cells. This strongly suggests that the mechanism of action of TCN is independent of any direct effect on the infectivity of the virus produced in the presence of drug.
In the course of evaluating TCN as an inhibitor of HIV-1 replication in acutely infected cells, a discordance between the results obtained by monitoring the supernatant RT and by monitoring the supernatant capsid protein (p24) as markers for virus replication was observed in some experiments. When RT was monitored, TCN exhibited significant antiviral activity, whereas it appeared to be inactive when p24 was monitored in the same experiments. This discordance led to the hypothesis that TCN prevents RT from packaging into virus particles and, therefore, is an inhibitor of HIV-1 assembly and release. This hypothesis was consistent with earlier observations that TCN has a unique mechanism of action that does not involve any of the virally encoded enzymes.
Since the discordance between the RT and the p24 data was observed in only a subset of experiments, it was hypothesized that the multiplicity of infection (MOI) used to infect the cells for this subset of experiments may have been significantly different from that used in experiments in which the discordance was not observed. Therefore, experiments were performed with HIV-1IIIB to infect a number of cell lines at several MOIs. The results obtained with AA-5 cells demonstrated little effect of the MOI on the discordance between the RT and the p24 data within the range of MOIs evaluated, whereas the results obtained with H9 cells indicated that the MOI might have an effect (Table (Table3).3). The discordance was evident at higher MOIs but not at lower MOIs. Finally, the discordance was not observed in CEM-SS cells at any of the MOIs tested. This suggests that TCN might have affected virus assembly and release differently depending upon the cell line used. Results consistent with those obtained with CEM-SS cells were also obtained by using macrophage-tropic HIV-1 isolates ADA and Ba-L in MDM cultures. Although RT data were not collected, TCN exhibited significant activity when supernatant p24 was monitored (IC50s = 0.006 to 1.94 μM for HIV-1ADA and 0.018 μM for HIV-1Ba-L). The toxicity of TCN toward the MDM cultures was well separated from the antiviral activity in these experiments (TC50s = 0.66 to >31.25 μM for studies with HIV-1ADA and 0.66 μM for studies with HIV-1Ba-L), resulting in therapeutic indices ranging from >16 to 110.
As a means of understanding the mechanism of action, HIV-1IIIB isolates resistant to TCN were selected. Isolates HIV-1D1, HIV-1A7, HIV-1A12, and HIV-1H10 demonstrated various degrees of resistance to TCN, with the IC50s ranging from 0.33 μM (approximately 5- to 10-fold increase in resistance) to >30 μM (>750-fold increase in resistance) (Table (Table4).4). Interestingly, the isolates that were resistant to TCN also exhibited some resistance to AZT, albeit to a lesser extent. In order to determine the gene responsible for the resistance of HIV-1IIIB to TCN, the viral genome was sequenced. Although mutations were discovered in vif, env, vpr, tat, and nef, the mutations in nef were the only ones consistent in all four isolates and, therefore, the only ones examined further. Five point mutations resulting in amino acid changes in nef were found in highly TCN-resistant (TCNr) isolate HIV-1H10 (Table (Table4).4). Sequencing of the other TCN-resistant isolates established that they all contained at least one and up to three of the same mutations observed in isolate HIV-1H10. The total number of changes in any one isolate correlated to the degree of resistance exhibited by that isolate (Table (Table4).4). Compared to the nef sequence alignments available from the Los Alamos HIV sequence database (http://hiv-web.lanl.gov/), the one mutation observed in all four resistant isolates (Tyr127His) was identified as a highly conserved tyrosine residue (99.8%) in the HIV-1, HIV-2, and SIV nef genes. Amino acids 140, 149, and 174 are also highly conserved, with greater than 95% of the sequences containing the amino acid found in HIV-1IIIB. Interestingly, >95% of the HIV-1 group O isolates contained either a Ser or a Thr (instead of Glu) at amino acid 149, which may explain the resistance of this group of isolates to TCN (Table (Table1;1; Nef sequences for isolates BCF01, BCF02, and BCF03 were not available). The sequence alignment for amino acid 178 demonstrated three polymorphisms, one of which was the amino acid found in HIV-1IIIB (Arg) and another of which was the amino acid found in both the HIV-1A10 and HIV-1H10 isolates (Lys). We therefore hypothesize that this mutation is not as important as the other four for HIV resistance to TCN.
Because these results demonstrated an association between mutations in nef and TCN resistance, marker transfer experiments were performed to establish that the mutations are necessary and sufficient for TCN resistance. Figure Figure33 demonstrates that both wild-type strains (HIV-1IIIB and HIV-1pNL4-3) were sensitive to TCN (IC50s, ~0.2 μM), although at slightly higher concentrations than those seen in the other experiments. As expected and in contrast, strain HIV-1H10 was approximately sevenfold more resistant to TCN (IC50 > 10 μM) than HIV-1IIIB, from which it had been selected. Likewise, HIV-1pTAK-1, in which the mutations in nef were transferred from HIV-1H10 into the HIV-1pNL4-3 clone (Fig. (Fig.3),3), was resistant to TCN (IC50 > 10 μM), thereby establishing that transfer of the mutant nef gene from HIV-1H10 had transferred drug resistance.
In an attempt to identify the mechanism of action of TCN against HIV, a number of experimental approaches have been utilized. Previous experiments demonstrated that TCN-P is not an inhibitor of HIV-1 reverse transcriptase (37), nor did TCN or TCN-P inhibit RNase H (23, 43) at a concentration of 10 μM (James Peliska, University of Michigan, personal communication). Likewise, neither inhibited HIV-1 protease (44) at concentrations well in excess of the antiviral effect range, although 15% inhibition was observed at a concentration 1,000-fold above the IC50 for the inhibition of HIV-1 replication (Ronald Swanstrom, University of North Carolina—Chapel Hill, personal communication). Inhibition of HIV-1 integrase 3′-end processing and DNA strand transfer activity evaluated by the methodology of Engelman and Craigie (15) also were not affected at a concentration of 10 μM. Taken together, these results led us to conclude that none of the targets of the marketed antiretroviral drugs were the target of TCN. In contrast, our experiments demonstrating that TCN retains activity against HIV-1 in chronically infected cells suggested that TCN inhibits a viral target late in the replication cycle at a point after provirus integration (37). The results presented herein provide evidence for the involvement of Nef in the antiviral action of TCN.
The cell line dependence of the differential effects of TCN against the viral p24 and RT end points observed in this study suggested that the potential antiviral target(s) could involve HIV-1 assembly and release. The requirement for many of the functions of HIV-1 accessory proteins Vif (5), Vpu (13, 21), and Nef (61) during HIV-1 replication is known to be cell line dependent. Furthermore, HIV-1 protein p6 is known to promote virus particle release in a cell type-dependent manner (12). This correlation suggests to us that TCN inhibits virus particle assembly and release and supports the hypothesis that one or more of these viral proteins are inhibited by TCN during HIV-1 replication. Furthermore, even though TCN did not appear to have an effect on the infectivity of HIV-1 released from chronically infected H9 cells, it is possible that this is a cell line-dependent observation because the functions of virally encoded accessory proteins associated with increasing HIV-1 infectivity are cell type-dependent functions.
The cytotoxicity of TCN in some of the experiments also supports the idea of the involvement of viral proteins. When TCN was tested in the U1 and ACH2 latently infected cell lines, it appeared to be cytotoxic rather than antiviral. However, this cytotoxicity was less apparent when the cells were not stimulated for the production of HIV-1. This raises the possibility that the cytotoxic effect was associated with the production of viral proteins. Since TCN prevents the assembly and release of HIV-1 from infected cells, the resulting intracellular buildup of viral proteins could be the cause of this observed cytotoxicity, inasmuch as many HIV-1 proteins are known to be cytotoxic (11, 31, 36, 40). Similarly, the cytotoxicity observed for TCN in MDM cultures as part of the antiviral evaluations with HIV-1 Ba-L and ADA could be the result of a similar effect. This is consistent with other observations that TCN is much less toxic to other uninfected cells (37). Although this is only one possible explanation for the observed cytotoxicity, it represents an intriguing potential mechanism and is consistent with the involvement of virally encoded accessory proteins in the antiviral action of TCN.
Sequencing of the nef gene from TCN-resistant isolates of HIV resulted in the discovery of five point mutations in the DNA sequence of nef. The Glu149Lys mutation is in a much less conserved location; however, it falls in the middle of a Pro-x-x-Pro amino acid motif known to be involved in Src homology domain 3 (SH3) binding. This particular SH3 binding motif in Nef is known to be required for Nef binding to Hck (52). The Asp174Asn mutation that we found in the two most resistant isolates maps to the position of the first Asp in an Asp-Asp-x-x-x-Glu motif. This motif is known to be critical for the CD4 downregulation function of Nef (3, 30) and for Nef binding to c-Raf1 kinase (27). Furthermore, the Tyr127His and Gly140Glu mutations map to a location in Nef required for Nef binding to Lck by a novel mechanism that requires Nef phosphorylation (14), for the association of Nef with a cellular serine kinase (54), and for Nef binding to a novel acyl coenzyme A thioesterase (63) and are essential for CD4 downregulation (29). Additionally, the Gly140Glu and the Glu149Lys mutations are located at key amino acids that have been described to form part of a consensus sequence for structural homology between Nef and CTLA-4, a T-cell regulatory protein (62). Finally, all of these mutations are located in the C-terminal portion of Nef, which is known to be a cell surface domain that interacts with CD4+ cells and that plays a role in syncytium formation (42).
One of the more intriguing possibilities for the mechanism of action of TCN is its ability to inhibit downstream effects resulting from the activation of cellular pathways due to the presence of virus. HIV gp120 induces several signaling pathways, including the phosphatidylinositol-3-kinase (PI-3K)/cSrc pathway and its downstream effector, Akt/PKB, a serine/threonine kinase. Although the role of this pathway in HIV replication is not completely understood, it has been demonstrated that the inhibition of PI-3K inhibits HIV replication, as long as the inhibitor is introduced before viral DNA integration into the host genome (19). Nef is a protein that is expressed immediately after integration (38), and exogenous HIV Nef protein induces the phosphorylation (activation) of exogenous Akt/PKB in human macrophages (60). Furthermore, Akt/PKB activates several cellular pathways that promote cell proliferation and survival, which are vital for viral replication (33). Since TCN-P is a known inhibitor of Akt/PKB in several tumor cell lines that over express Akt/PKB (68), it is reasonable to assume that TCN-P acts in a similar manner in HIV-infected cells. Therefore, because the mutations in nef described herein produce resistance to TCN and TCN-P is a known inhibitor of Akt/PKB, we hypothesize that TCN-P interacts with Akt/PKB near the point at which Nef interacts with Akt/PKB, thereby inhibiting viral replication. In this way, the mutations observed in Nef could allow it to overcome the interaction of the enzyme with drug, resulting in an Akt/PKB function necessary for the replication and release of virus. Hence, Nef would not be the target of TCN-P, but, as in cancer cells, Akt/PKB would be the target.
Consistent with this hypothesis, the replication of a nef-null clone described by Aiken et al. (1, 2) was inhibited by TCN-P (data not shown). This would occur because the putative target of TCN-P is endogenous Akt/PKB. In the absence of Nef, the activation of Akt/PKB in HIV-infected cells would occur via gp120 stimulation of the PI-3K/cSrc pathway, thereby maintaining the activity of the target for TCN-P. However, the interaction of TCN-P with other HIV-1 proteins and the direct inhibition of Nef by TCN-P cannot be ruled out, but both of those possibilities appear to be less likely to us.
The potential use of TCN as a clinical antiviral drug will be dictated by human toxicology. A large amount of toxicity data already exist as the result of anticancer clinical trials already performed with TCN (17, 18, 28, 39, 47, 55). Although adverse effects such as hepatic toxicity, hyperglycemia, and thrombocytopenia were observed (17), these toxicities were associated with the aggressive nature of these trials in an attempt to treat cancers through the intravenous infusion of large quantities of the TCN prodrug TCN-P. The information generated from those trials should prove invaluable for the further development of TCN as an agent with activity against HIV-1. On the basis of the broad activity of TCN against many different HIV-1 clinical isolates and subtypes, including multidrug-resistant HIV-1, and its unique mechanism of action involving the viral accessory protein Nef, additional development of this compound for use for the treatment of HIV-1 infections appears to be warranted.
We thank Jim Turpin for his insight regarding these data and Julie Breitenbach for performing certain HIV assays.
This study was supported by research grants U01-25739, R01-AI33332, and R01-AI36872 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, and by internal funding from the Southern Research Institute and the University of Michigan.
Published ahead of print on 19 January 2010.