Novel Acyclic Nucleoside Phosphonate Analogues with Potent Anti-Hepatitis B Virus Activities

Abstract

Novel acyclic nucleoside phosphonates with a pyrimidine base preferentially containing an amino group at C-2 and C-4 and a 2-(phosphonomethoxy)ethoxy or (R)-2-(phosphonomethoxy)propoxy group at C-6 selectively inhibit the replication of wild-type and lamivudine-resistant hepatitis B viruses. The activity of the most potent compounds was comparable to that of adefovir.

Novel acyclic pyrimidine nucleoside phosphonate (pyrimidine ANP) analogues in which the aliphatic phosphonate side chain is linked to the C-6 position instead of the natural N-1 position of the pyrimidine ring (Fig. 1) (7) were recently shown to possess potent antiretroviral properties (1, 6). It may be surmised that these anti-human immunodeficiency virus (HIV) molecules, akin to adefovir and tenofovir (12), exhibit antiviral activity through inhibition of viral reverse transcriptase by their diphosphorylated metabolites. We therefore investigated whether these novel compounds have activities against hepatitis B virus (HBV) replication.

FIG. 1.

Structural formulae of pyrimidine ANP analogues in which the acyclic side chain is attached to either C-6 or C-1 of the pyrimidine ring and of several PME, PMP, and HPMP (hydroxy-phosphonyl-methoxypropyl) purine analogues.

To this end, we made use of the tetracycline-responsive cell lines HepAD38 and HepAD79 (8, 9). These are hepatoma cells that have been stably transfected with a cDNA copy of the pregenomic RNA of wild-type virus or with cDNA encoding the lamivudine-resistant YMDD (M204V) variant. The withdrawal of tetracycline from the culture medium results in the initiation of viral replication (11). Both cell lines were cultured at 37°C in a humidified 5% CO₂ atmosphere in seeding medium, Dulbecco modified Eagle medium-Ham F-12 medium (50:50) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum, 100 IU of penicillin/ml, 50 μg of streptomycin/ml, 100 μg of kanamycin/ml, 400 μg of G418/ml, and 0.3 μg of tetracycline/ml. When the assay was started, the cells were seeded in 48-well plates at a density of 5 × 10⁵ cells per well. After 2 to 3 days, the cultures were induced for viral production by washing them with prewarmed phosphate-buffered saline and were fed 200 μl of assay medium (seeding medium without tetracycline and G418) with or without the antiviral compounds. The medium was changed after 3 days. The antiviral effect was quantified by measuring the levels of viral DNA in the culture supernatant at day 6 postinduction by real-time quantitative PCR. Cell culture supernatant was collected and centrifuged at 1,500 × g for 10 min. The quantitative PCR was performed with 3 μl of culture supernatant in a reaction volume of 25 μl by using TaqMan Universal PCR Master Mix (Applied Biosystems, Branchburg, N.J.) with a forward primer (5′-CCG TCT GTG CCT TCT CAT CTG-3′; final concentration, 600 nM), a reverse primer (5′-AGT CCA AGA GTY CTC TTA TRY AAG ACC TT-3′, where Y is either C or T and R is either A or G; final concentration, 600 nM), and a TaqMan probe (6-FAM-CCG TGT GCA CTT CGC TTC ACC TCT GC-TAMRA, where FAM is 6-carboxyfluorescein and TAMRA is 6-carboxytetramethylrhodamine; final concentration, 150 nM). The reaction was analyzed with an SDS 7000 (Applied Biosystems, Foster City, Calif.) real-time PCR apparatus. A plasmid containing the full-length insert of the HBV genome was used to prepare the standard curve. The amounts of viral DNA produced in treated cultures were expressed as percentages of that in the mock-treated samples. At day 6 after the induction of the HepAD38 cells, culture supernatant contained large numbers of Dane particles, as was demonstrated by means of electron microscopy (data not shown).

The cytostatic effects of the various compounds were assessed by employing the parent hepatoma cell line, HepG2, as well as Vero cells. The effects of the compounds on exponentially growing HepG2 cells were evaluated by means of the MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] method (Promega). Briefly, cells were seeded at a density of 3,000 cells/well (96-well plates) and allowed to proliferate for 3 days in the absence or presence of compounds, after which time cell density was determined.

Selected pyrimidine ANP analogues were evaluated for their potential activities against HBV replication in HepAD38 and HepAD79 cultures. Characteristic structural features of the pyrimidine ANP analogues included the presence of amino groups at positions C-2 and C-4 and the 2-(phosphonomethoxy)ethoxy (PMEO) or 2-(phosphonomethoxy)propoxy (PMPO) group linked to the C-6 position of the pyrimidine ring. The compounds were thus designated 6-PMEO and 6-PMPO, respectively (Fig. 1). Depending on the additional ring substitution, these selected compounds exhibited differential anti-HBV activities (Table 1). The 6-PMEO derivative, 2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]-pyrimidine (compound1), and its 5-cyano analogue (compound 8) exhibited the most potent anti-HBV activities, with 50% effective concentrations (EC₅₀s) ± standard deviations (SD) of 0.3 ± 0.2 μM and 0.14 ± 0.01 μM, respectively, against wild-type HBV and 0.25 ± 0.2 μM and 0.5 ± 0.35 μM, respectively, against the YMDD (lamivudine-resistant) variant. These activities are very similar to the anti-HBV activities of the reference compounds PMEA [9-(2-phosphonylmethoxyethyl)adenine] and PMPA [9-(2-phosphonylmethoxypropyl)adenine] (Fig. 2 and Table 2).

TABLE 1.

Anti-HBV activities of pyrimidine ANP analogues with C-6-PME or C6-PMP side chains^a

Compound no.b	Base at position:			Side chain		EC50 (μM)c		HepG2 or Vero cell CC50 (μM)d
	R1	R2	R3	Y	Z	HepAD38	HepAD79
1	NH2	NH2	H	O	H	0.3 ± 0.2	0.25 ± 0.2	>379
2	NH2	NH2	CH3	O	H	3.2 ± 1.8	2 ± 0.7	>360
3	H	NH2	H	O	H	120 ± 80	80 ± 60	>402
4	NH2	HO	H	O	H	1.3 ± 0.7	0.9 ± 0.6	>329
5	NH2	NH2	Br	O	H	8.7 ± 3	17.5 ± 1.5	>292
6	NH2	NH2	CHO	O	H	21 ± 10	9 ± 1.2	>342
7	NH2	CH3	H	O	H	>300	>300	>380
8	NH2	NH2	CN	O	H	0.14 ± 0.01	0.5 ± 0.35	>346
9	NH2	NH2	CH2CN	O	H	>300	>300	>330
10	HO	NH2	H	S	H	>300	>300	>356
11	NH2	NH2	H	S	H	1.8 ± 1	1 ± 0.7	>357
12	NH2	NH2	H	O	CH3 (R)	4.3 ± 1	1.1 ± 2	>360
13	NH2	NH2	H	O	CH3 (S)	61 ± 14	>300	>360

^aSee Fig. 1 for chemical structures.

^bCompound 1, 2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 2, 5-methyl-2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 3, 4-amino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 4, 2-amino-4-hydroxy-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 5, 5-bromo-2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 6, 5-formyl-2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 7, 2-amino-4-methyl-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 8, 5-cyano-2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 9, 5-cyanomethyl-2,4-diamino-6-[2-(phosphonomethoxy)ethoxy]pyrimidine; 10, 4-amino-2-hydroxy-6-{[2-(phosphonomethoxy)ethyl]sulfanyl}pyrimidine; 11, 2,4-diamino-6-{[2-(phosphonomethoxy)ethyl]sulfanyl}pyrimidine; 12, 2,4-diamino-6-[(R)-2-phosphonomethoxy)propoxy]pyrimidine; 13, 2,4-diamino-6-[(S)-2-phosphonomethoxy)propoxy]pyrimidine.

^cEach value represents the mean ± SD for three independent experiments.

^dCC₅₀, 50% cytostatic concentration. Data are from two independent experiments.

FIG. 2.

Dose-dependent anti-HBV activity (in HepAD38 cells) of compound 1 (grey bars) and PMEA (white bars) (data are mean values ± SD for at least three independent experiments).

TABLE 2.

Anti-HBV activities of lamivudine and purine-based ANP analogues

Compound	EC50 (μM)a
	HepAD38	HepAD79
PMEA	0.07 ± 0.04	0.1 ± 0.04
PMEDAP	0.4 ± 0.1	0.5 ± 0.3
PMPA	0.17 ± 0.07	0.3 ± 0.17
(S)-PMPDAP	7.6 ± 6.6	14.5 ± 12
(R)-PMPDAP	0.3 ± 0.3	1.0 ± 0.7
3TCb	0.02 ± 0.009	0.65 ± 0.3

^aEach value represents the mean ± SD for three independent experiments.

^b3TC, lamivudine ((−)-β-2′,3′-dideoxy-3′-thiacytidine).

A prerequisite for potent anti-HBV activity was the presence of an amino group at C-2 (most preferentially) of the pyrimidine ring (compare compounds 1 and 3) together with an amino or hydroxyl group at the C-4 position (compare compounds 1 and 7) (Table 1). A substitution at the C-5 position (R³) of the pyrimidine ring (4, 5) resulted in various influences on the anti-HBV efficacy (Table 1). A methyl group in this position (compare compounds 1 and 2) resulted in a 10-fold decrease in antiviral activity. Replacement with 5-bromo (compound 5) or 5-formyl (compound 6) resulted in compounds with moderate activities. Interestingly, the introduction of a 5-cyano (5-CN; compound 8) resulted in a compound that had an anti-HBV efficacy equipotent to that of compound 1. In contrast, the introduction of a longer side chain, i.e., 5-cyanomethyl (5-CH₂CN), resulted in a complete loss of anti-HBV activity (compound 9). These findings are in sharp contrast to the results obtained for the anti-HIV and anti-mouse sarcoma virus activities, for which the 5-methyl and 5-bromo derivatives were found to be the most powerful inhibitors (4).

Replacement of the oxygen atom in the 6-PMEO group at the pyrimidine ring with sulfur resulted in a compound that was fivefold less active (compare compounds 1 and 11). The replacement of the amino group at C-2 by a hydroxyl in this thioether molecule (compound 10) resulted in a totally inactive molecule, a result which is in agreement with the weak activity of compound 3. Introduction of a methyl group on the acyclic chain resulted in a PMPO molecule that was about 10- to 30-fold less active (compare compounds 1 and 12). The S-enantiomer of this compound proved only weakly active (against HepAD38) or almost inactive (against HepAD79) (compare compounds 12 and 13). Also, (S)-9-(2-phosphonomethoxypropyl)-2,6-diaminopurine [(S)-PMPDAP] was >10- to 20-fold less active than (R)-PMPDAP (Table 2). Such an enantiospecificity in the series of purine PMP derivatives was also observed for retroviruses (2, 3). Contrary to the activities of the 6-PMEO and (R)-6-PMPO derivatives, the anti-HBV activities of the N-1 isomers (Fig. 1, compounds 14 through 16) of these pyrimidine ANPs were completely lost (data not shown).

The PMEO and PMPO pyrimidine analogues studied here did not prove to be cytostatic (or cytotoxic) in HepG2 and Vero cells (Table 1). Several compounds in this series thus appear to exhibit potent and selective anti-HBV activities. Molecular modeling provides compelling evidence that these pyrimidine derivatives can be considered mimics of 9-[(2-phosphonomethoxy)ethyl]-2,6-diaminopurine (PMEDAP) in that the 2,4-diamino-substituted pyrimidine ring may be considered an open-ring analogue of the purine system in the 2,6-diaminopurine ANP derivatives (Fig. 3). The structures of compound 1 and PMEDAP were optimized at the B3LYP/6-31G* level by using the Gaussian 03 program suite (revision B.02; M. J. Frisch et al., Gaussian, Inc., Pittsburgh, Pa., 2003). The optimization was started from several initial conformations, and the lowest minimum was selected. The structures were compared. The optimized structures were aligned with the maximum overlap of the pyrimidine ring, and the distance between the phosphorus atoms of both molecules was measured as the similarity index. Finally, the molecules were reoptimized with constraints, forcing the phosphorus atom to be placed at the same position relative to the pyrimidine ring. The energy difference between the free and constrained structures was computed as a measure of how easily the molecule could adopt the structure for good binding. The constraint is defined by the fixed distance of the P atom from C-6 of the pyrimidine ring, the fixed P-C-6-C-5 angle, and the fixed P-C-6-C-5-N-3 torsion. The phosphate group in the optimized conformation of compound 1 was found to be located only 1.68 Å from its position in the optimized conformation of PMEDAP. The energy difference required for the conformational changes of molecule 1 to achieve the phosphate group shift to its relative position in PMEDAP is as low as 3.4 kcal mol⁻¹, a typical energy level for a hydrogen bond. Taken together, these data strongly suggest that compound 1, as a representative (and the most potent) of this novel class of compounds, may indeed function as a diaminopurine mimic.

FIG. 3.

Overlay (C) of the optimized structures of compound 1 (A) and PMEDAP (B). Blue, nitrogen; orange, carbon; red, oxygen; yellow, phosphor.

We assume that the 2,4-diaminopyrimidine ring system of the 6-PMEO and 6-PMPO derivatives retains the most important features of the purine base for recognition by the activating enzymes. It is very likely that these compounds are phosphorylated intracellularly to their 5′-diphosphorylated metabolites, which then inhibit the reverse transcriptase and/or the DNA polymerase activity of the HBV polymerase. Interestingly, akin to PMEA and PMPA, all compounds in this class with activities against wild-type HBV retain their activities against the lamivudine-resistant YMDD variant. Recently, homology modeling with HIV reverse transcriptase was used to predict how the diphosphorylated form of PMEA (PMEApp) may interact with the HBV polymerase, an enzyme for which a three-dimensional structure is not yet available (10). A similar modeling study with the novel ANPs may shed further light on their mechanisms of action. The characteristics, the intracellular pharmacology, and the molecular mechanisms of action of the anti-HBV activities of these novel PMEO and PMPO pyrimidine analogues will be subjects of further investigations.