Abstract
Alkoxyalkyl esters of cidofovir (CDV) have substantially greater antiviral activity and selectivity than unmodified CDV against herpesviruses and orthopoxviruses in vitro. Enhancement of antiviral activity was also noted when cyclic CDV was esterified with alkoxyalkanols. In vitro antiviral activity of the most active analogs against human cytomegalovirus (HCMV) and orthopoxviruses was increased relative to CDV up to 1,000- or 200-fold, respectively. Alkyl chain length and linker structure are important potential modifiers of antiviral activity and selectivity. In this study, we synthesized a series of alkoxyalkyl esters of CDV or cyclic CDV with alkyl chains from 8 to 24 atoms and having linker moieties of glycerol, propanediol, and ethanediol. We also synthesized alkyl esters of CDV which lack the linker to determine if the alkoxyalkyl linker moiety is required for activity. The new compounds were evaluated in vitro against HCMV and murine CMV (MCMV). CDV or cyclic CDV analogs both with and without linker moieties were highly active against HCMV and MCMV, and their activities were strongly dependent on chain length. The most active compounds had 20 atoms esterified to the phosphonate of CDV. Both alkoxypropyl and alkyl esters of CDV provided enhanced antiviral activities against CMV in vitro. Thus, the oxypropyl linker moiety is not required for enhanced activity. CDV analogs having alkyl ethers linked to glycerol or ethanediol linker groups also demonstrated increased activity against CMV.
Cidofovir (CDV), an acyclic phosphonate analog of dCMP, is an antiviral agent that is active against all double-stranded DNA viruses including herpes simplex, cytomegalovirus (CMV), orthopoxviruses, adenovirus, Epstein-Barr virus, polyomavirus, and papillomavirus (9). CDV is not orally active but is effective when administered intravenously for CMV retinitis in patients with AIDS (8, 19). Esterification of CDV with certain alkoxyalkanols dramatically increases the antiviral activity and selectivity of CDV in vitro (2, 11, 13, 15) and confers oral bioavailability (4, 5, 6, 14, 20). Hexadecyloxypropyl-CDV (HDP-CDV) and various other alkoxyalkyl CDV esters are orally active in three lethal challenge models of poxvirus disease (5, 20) and in animal models of CMV disease (4, 14).
The alkyl chain length of these CDV analogs is related to solubility and the ability of the compounds to associate with biomembranes. To examine the effect of structure on antiviral activity and selectivity of alkoxyalkanol-CDV analogs against CMV, we synthesized a family of alkoxypropyl-CDVs and -cyclic CDVs (cCDVs) varying in overall chain length from 8 to 24 atoms. The nature of the linker group is also important because it may strongly affect the rate of cellular metabolic conversion to CDV. We synthesized several representative analogs having propanediol, ethandiol, or glycerol linkers to assess the effect of the linker structure. Finally, we also synthesized and evaluated a series of alkyl-CDVs lacking the linker.
The various CDV and cCDV analogs were tested in vitro for antiviral activity against the AD-169 strain of human CMV (HCMV) and the Smith strain of murine CMV (MCMV). Cytotoxicity was assessed and selectivity was determined. The antiviral activity and selectivity of the analogs were highly dependent on the number of atoms in the attached ester. Linkers consisting of glycerol, ethanediol or propanediol were all effective but were not required for enhanced antiviral activity.
MATERIALS AND METHODS
General.
1H and 31P nuclear magnetic resonance (NMR) spectra were recorded on a Varian HG-400 spectrometer with tetramethylsilane (internal) and 85% D3PO4 in D2O (external) as references for 1H and 31P (0.00 ppm), respectively. Electrospray ionization mass spectroscopy (ESI) was performed by HT Laboratories (San Diego, Calif.). Chromatographic purification was done by the flash method with Merck silica gel 60 (240 to 400 mesh). All final products were homogeneous by thin-layer chromatography performed on Analtech 250-μm Silica Gel GF Uniplates visualized under UV light, with phospray (Supelco, Bellafonte, Pa.), and by charring (400°C). Cyclic cidofovir dihydrate was provided by Gilead Sciences, Inc. (Foster City, Calif.). Bromoalkanes and bromoalkoxyalkanes were either commercially available or synthesized from the corresponding alcohol as previously described (2). All other chemicals were of reagent quality and used as obtained from the suppliers. All reactions were carried out in an inert atmosphere (dry nitrogen).
Esterification procedure A.
The appropriate 1-bromoalkane (or 1-bromoalkoxyalkane) (1.2 eq) was added to a suspension of cCDV dihydrate, N,N-dicyclohexyl-4-morpholino-carboxamidine (1.1 eq) and N,N-dimethylformamide (10 ml/mmol), and the mixture was heated to 60°C and stirred magnetically overnight. The solvent was then evaporated under reduced pressure, and the residue was adsorbed onto silica gel and purified by flash column chromatography (elution gradient, CH2Cl2 to 15% EtOH). The cCDV esters were isolated as equimolar mixtures of the axial and equatorial diastereomers. The following compounds were prepared by using procedure A (Fig. 1).
FIG. 1.
Chemical structures of key lipid esters of CDV.
(i) Octyl cyclic cidofovir.
Yield 46%; 1H NMR (CDCl3 + CD3OD) δ 7.00 (dd, J = 7.4 Hz, J = 10 Hz, 1H), 5.44 (d, J = 7.4 Hz, 1H), 3.99 to 4.29 (m, 2H), 3.64 to 3.86 (m, 6H), 3.50 to 3.58 (m, 2H), 3.12 to 3.33 (m, 1H), 1.28 to 1.39 (m, 2H), 0.75 to 1.15 (m, 10H), 0.52 (t, J = 6.8 Hz); 31P NMR δ 12.31, 13.58. MS (ESI): m/z 374 (M + H) +, 372 (M − H)−.
(ii) Dodecyl cyclic cidofovir.
Yield 45%; 1H NMR (CDCl3 + CD3OD) δ 7.02 (dd, J = 7.1 Hz, J = 10 Hz, 1H), 5.47 (d, J = 7.4 Hz, 1H), 4.03 to 4.07 (m, 2H), 3.69 to 3.89 (m, 6H), 3.53 to 3.60 (m, 2H), 3.15 to 3.35 (m, 1H), 1.32 to 1.42 (m, 2H), 0.86 to 1.15 (m, 14H), 0.54 (t, J = 6.0 Hz); 31P NMR (CDCl3 + CD3OD) δ = 12.22, 13.51. MS (ESI): m/z 430 (M + H) +, 452 (M + Na)+; 428 (M − H)−.
(iii) Hexadecyl cyclic cidofovir.
Yield 55%; 1H NMR (dimethyl sulfoxide [DMSO]-d6) δ 7.48 (dd, J = 31.2, 7.2 Hz, 1H), 7.108 (br d, J = 39.9 Hz 2H), 5.64 (t, J = 6.8 Hz, 1H), 3.5 to 4.9 (m, 7 H), 1.60 (br s, 2H), 1.24 (br s, 28H), 0.86 (t, J = 6.5 Hz, 3H); 31P NMR (CDCl3 + CD3OD) δ 12.18, 13.58. MS (ESI): m/z 486 (MH)+, 484 (M − H)−.
(iv) Eicosyl cyclic cidofovir (EC-cCDV).
Yield 8%; 1H NMR (DMSO-d6) δ 7.43 (d, J = 7.2 Hz, 1H), 7.08 (br d, J = 30.9 Hz 2H), 5.61 (d, J = 6.9 Hz, 1H), 3.5 to 4.9 (m, 7 H), 1.60 (br s, 2H), 1.24 (br s, 36H), 0.86 (t, J = 6.5 Hz, 3H); 31P NMR δ = 14.02, 12.81; MS (ESI): m/z 564 (M + Na) +, 542 (M + H)+.
(v) Tetracosyl cyclic cidofovir.
Yield 17%; 1H NMR (DMSO-d6) δ 7.43 (d, J = 6.9 Hz, 1H), 7.07 (br d, J = 25.5 Hz 2H), 5.63 (d, J = 6.9 Hz, 1H), 3.1 to 4.0 (m, 7 H), 1.60 (br s, 2H), 1.22 (br s, 44H), 0.84 (t, J = 6.5 Hz, 3H); 31P NMR (CDCl3 + CD3OD) δ 12.66, 13.86. MS (ESI): m/z 620 (M + Na) +, 598 (M + H)+.
(vi) Octyloxypropyl cyclic cidofovir.
Yield 18%; 1H NMR (CDCl3) δ 7.44 (dd, J = 24.15, 7.5 Hz, 1H), 6.08 (br d, 1H), 3.5 to 4.8 (m, 10 H), 1.91 to 2.1 (m, 2H), 1.56 (br s, 2H), 1.27 (br s, 10H), 0.88 (t, J = 6.5 Hz, 3H); 31P NMR δ = 13.58, 11.95; MS (ESI): m/z 454 (M + Na,) +, 432 (M + H)+.
(vii) Dodecyloxypropyl cyclic cidofovir.
Yield 52%; 1H NMR (CDCl3) δ 7.30 (m, 1H), 5.83 (t, J = 7.2 Hz, 1H), 3.36 to 4.44 (m, 10 H), 1.78 to 2.20 (m, 2H), 1.44 to 1.70 (m, 2H), 1.26 (br s, 18H), 0.88 (t, J = 6.5 Hz, 3H); 31P NMR δ 13.20, 11.61; MS (ESI): m/z 488 (M + H) +.
(viii) Oleyloxypropyl cyclic cidofovir (OLP-cCDV).
Yield 45%; 1H NMR (DMSO-d6) δ 7.48 (dd, J = 39.4, 7.2 Hz, 1H), 7.08 (br d, J = 39.4,2H), 5.64 (t, J = 7.2 Hz, 1H), 5.31 (t, J = 4.8 Hz, 2H, vinyl), 3.4 to 4.5 (m, 10 H), 1.97 (m, 2H), 1.83 (m, 2H), 1.47 (m, 2H), 1.26 (br s, 28H), 0.84 (t, J = 6.8 Hz, 3H); 31P NMR δ 14.11, 12.85; MS (ESI): m/z 570 (M + H) +.
Esterification procedure B.
Some of the analogs were made by using an alternative procedure. Anhydrous cCDV (1 eq), the appropriate alkoxyalkanol (or alkanol; 2 eq), and triphenylphosphine (2 eq) were dissolved or suspended in anhydrous N,N-dimethylformamide (6.5 ml per mmol of cCDV) and stirred vigorously under a nitrogen atmosphere. Diisopropyl azadicarboxylate (2 eq) was then added in three portions over 15 min before the mixture was allowed to stir overnight. The solvent was then evaporated under vacuum, and the residue was adsorbed onto silica gel and purified by column chromatography. Gradient elution from 100% CH2Cl2 to 15% EtOH was followed by recrystallization from p-dioxane. The coupled products were isolated as equimolar mixtures of axial and equatorial diastereomers. The following compounds were prepared by using procedure B.
(i) Docosyl cyclic cidofovir.
Yield 26%; 1H NMR (DMSO-d6) δ 7.43 (d, J = 7.2 Hz, 1H), 7.07 (br d, J = 25.5 Hz 2H), 5.61 (d, J = 6.9 Hz, 1H), 3.3 to 4.4 (m, 7 H), 1.60 (br s, 2H), 1.23 (br s, 40H), 0.86 (t, J = 6.5 Hz, 3H); 31P NMR δ 12.80; MS (ESI): m/z 564 (M + Na) +, 570 (M + H)+, 604 (MCl)−, 568 (M − H)−.
(ii) Tetradecyloxypropyl cyclic cidofovir.
Yield 3.4%; 1H NMR (CDCl3 + CD3OD) δ 6.98 (dd, J = 7.0 Hz, 1H), 5.21 (d, J = 7.2 Hz, 1H), 3.95 to 4.11 (m, 2H), 3.48 to 3.92 (m, 6H), 3.09 to 3.20 (m, 2H), 3.00 to 3.09 (m, 2H), 2.95 to 3.00 (m, 1H), 1.51 to 1.65 (m, 2H), 1.01 to 1.23 (m, 2H), 0.87 (br s, 42H), 0.48 (t, J = 6.8 Hz, 3H); 31P NMR δ 13.67, 12.44. MS (ESI): m/z 516 (M + H)+, 538 (M + Na)+; 514 (M − H)−, 550 (MCl)−.
(iii) Oleyloxyethyl cyclic cidofovir (OLE-cCDV).
Yield 34%; 1H NMR (DMSO-d6) δ 7.48 (m, 1H), 7.19 (br d, J = 42.3,2H), 5.67 (m, 1H), 5.32 (m, 2H, vinyl), 3.2 to 4.4 (m, 10 H), 1.97 (m, 2H), 1.83 (m, 2H), 1.47 (m, 2H), 1.26 (br s, 28H), 0.85 (m, 3H); 31P NMR δ 14.36, 13.05; MS (ESI): m/z 556 (M + H)+.
(iv) Eicosyloxypropyl cyclic cidofovir.
Yield 22%; 1H NMR (CDCl3 + CD3OD) δ = 7.00 (dd, J = 7.4 Hz, J = 10 Hz, 1H), 5.42 (d, J = 7.5 Hz, 1H), 3.91 to 4.17 (m, 2H), 3.46 to 3.91 (m, 6H), 3.21 to 3.38 (m, 2H), 3.08 to 3.21 (m, 2H), 3.00 to 3.08 (m, 2H), 2.92 to 3.00 (m, 1H), 1.51 to 1.62 (m, 2H), 1.10 to 1.23 (m, 2H), 0.80 to 1.10 (m, 34H), 0.47 (t, J = 6.8 Hz, 3H). 31P NMR (CDCl3 + CD3OD) δ = 12.48, 13.72. MS (ESI): m/z 600 (M + H)+; 598 (M − H)−, 634 (MCl)−.
(v) 1-O-Octadecyl-2-O-benzyl-sn-glycero-3-cCDV (ODBG-cCDV).
Yield 45%; 1H NMR (CDCl3) δ 7.27 to 7.38 (m, 7H), 7.16 and 7.30 (pair of doublets, 1H), 5.72 and 5.68 (pair of doublets, 1H), 4.68 and 4.62 (pair of singlets, 2H), 3.97 to 4.40 (m, 9H), 3.44 (t, 2H), 3.41 (t, 2H), 3.25 (m, 1H), 1.54 (m, 2H), 1.26 (br s, 30H), 0.88 (t, 3H); 31P NMR δ 13.72 and 12.01; MS (ESI) m/z 678 (M + H)+, 700 (M + Na)+, 676 (M − H)−.
General ring-opening procedure.
The alkoxyalkyl- or alkyl-cCDV analogs were suspended in 2 M NaOH (25 ml/mmol). The suspensions were heated to 80°C and stirred for 1 h, during which the mixtures became clear. After hydrolysis, the solutions were cooled to 25°C and acidified with glacial acetic acid (pH approximately 5). The resulting precipitates were collected by vacuum filtration and dried under vacuum. The crude products were purified either by flash column chromatography (eluant, CH2Cl2-20% MeOH) or recrystallized to purity from ethanol (2). Yields and analytical results are provided.
(i) Octyl cidofovir, sodium salt.
Yield 46%; 1H NMR (DMSO-d6) δ 7.53 (d, J = 7.5 Hz, 1H), 7.12 (br d, J = 58.5 Hz 2H), 5.64 (d, J = 7.5 Hz, 1H), 3.20 to 3.80 (m, 7 H), 1.42 (br s, 2H), 1.24 (br s, 12H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ = 14.46; MS (ESI): m/z 436 (M + 2Na) +, 414 (M + Na)+; 390 (M − H)−.
(ii) Dodecyl cidofovir, sodium salt.
Yield 32%; 1H NMR (DMSO-d6) δ 7.54 (d, J = 7.2 Hz, 1H), 7.19 (br d, J = 57.3 Hz 2H), 5.67 (d, J = 7.2 Hz, 1H), 3.20 to 3.81 (m, 7 H), 1.42 (br s, 2H), 1.22 (br s, 20H), 0.84 (t, J = 6.5 Hz, 3H); 31P NMR δ 14.60; MS (ESI): m/z 492 (M + 2Na) +, 470 (M + Na)+; 446 (M − H)−.
(iii) Hexadecyl cidofovir, sodium salt.
Yield 43%; 1H NMR (DMSO-d6) δ 7.52 (d, J = 7.2 Hz, 1H), 7.19 (br d, J = 57.3 Hz 2H), 5.65 (d, J = 7.2 Hz, 1H), 3.2 to 3.9 (m, 7 H), 1.45 (br s, 2H), 1.24 (br s, 28H), 0.86 (t, J = 6.5 Hz, 3H); 31P NMR (D2O) δ = 16.73; MS (ESI): m/z 504 (M + Na)+, 526 (M + 2Na) +, 502 (M − H)−.
(iv) Eicosyl cidofovir, sodium salt (EC-CDV).
Yield 39%; 1H NMR (DMSO-d8) δ = 7.49 (d, J = 7.2 Hz, 1H), 7.01 (br d, J = 45.3 Hz 2H), 5.60 (d, J = 7.2 Hz, 1H), 3.2 to 3.9 (m, 7 H), 1.42 (br s, 2H), 1.26 (br s, 36H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ = 14.22; MS (ESI): m/z 604 (M + 2Na) +, 582 (M + Na)+; 558 (M − H)−.
(v) Docosyl cidofovir, sodium salt.
Yield 57%; 1H NMR (DMSO-d6) δ 7.50 (d, J = 7.2 Hz, 1H), 7.03 (br d, J = 50.4 Hz 2H), 5.60 (d, J = 7.2 Hz, 1H), 3.2 to 3.9 (m, 7 H), 1.42 (br s, 2H), 1.22 (br s, 40H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ 14.21; MS (ESI): m/z 632 (M + 2Na) +, 582 (M − H, sodium salt plus H)+; 540 (M − H)−, 576 (M − Cl).
(vi) Tetracosyl cidofovir, sodium salt.
Yield 35%; 1H NMR (CDCl3 + methanol d4) δ 7.19 (d, J = 7.4 Hz, 1H), 5.41 (d, J = 7.4 Hz, 1H), 3.03 to 3.62 (m, 7H), 2.90 to 2.91 (m, 2H), 1.10 to 1.22 (m, 2H), 0.83 (br s, 42H), 0.45 (t, J = 6.9 Hz). 31P NMR (CDCl3 + methanol d4) δ = 16.14. MS (ESI): m/z 638 (M + H, sodium salt plus H)+, 660 (M + Na, sodium salt of sodium salt) +; 614 (free acid)−.
(vii) Octyloxypropyl cidofovir, sodium salt.
Yield 62%; 1H NMR (DMSO-d8, 400 MHz) δ = 7.51 (d, J = 6.8 Hz, 1H), 6.94 (br d, J = 94.5 Hz 2H), 5.65 (d, J = 6.8 Hz, 1H), 3.2 to 3.9 (m, 10 H), 1.70 (m, 2H), 1.45 (br s, 2H), 1.23 (br s, 10H), 0.84 (t, J = 6.5 Hz, 3H); 31P NMR δ = 14.81; MS (ESI): m/z 494 (M + H, sodium salt plus H)+, 386.
(viii) Dodecyloxypropyl cidofovir, sodium salt.
Yield 69%; 1H NMR (DMSO-d8) δ = 7.51 (dd, J = 6.8, 4.0 Hz 1H), 7.09 (br d, J = 57.9 Hz 2H), 5.83 (dd, J = 6.8, 4.0 Hz, 1H), 3.2 to 3.8 (m, 10 H), 1.66 (m, 2H), 1.5 (m, 2H), 1.23 (br s, 18H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ = 14.35; MS (ESI): m/z 550 (M + Na, sodium salt of sodium salt) +, 528 (M − H, sodium salt plus H)+; 504 (M − H, free acid minus H)−.
(ix) Tetradecyloxypropyl cidofovir, sodium salt.
Yield 62%; 1H NMR (DMSO-d6) δ 7.51 (d, J = 7.2 Hz 1H), 7.08 (br d, J = 53.4 Hz 2H), 5.63 (d, J = 7.2 Hz, 1H), 3.2 to 4.2 (m, 10 H), 1.68 (m, 2H), 1.45 (m, 2H), 1.23 (br s, 22H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ 14.52; MS (ESI): m/z 578 (M + 2Na) +; 532 (M − H)−.
(x) Oleyloxypropyl cidofovir, sodium salt (OLP-CDV).
Yield 42%. 1H NMR (DMSO-d6) δ 7.51 (d, J = 6.9 Hz 1H), 7.02 (br d, J = 50.4 Hz 2H), 5.63 (d, J = 7.2 Hz, 1H), 5.31 (t, J = 5.0, 2H, vinyl), 3.2 to 3.8 (m, 10 H), 1.97 (m, 2H), 1.68 (m, 2H), 1.47 (m, 2H), 1.23 (br s, 28H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ 52.83; MS (ESI): m/z 632 (M + 2Na) +, 610 (M + Na)+; 586 (M − H)−.
(xi) Oleyloxyethyl cidofovir, sodium salt (OLE-CDV).
Yield 77%; 1H NMR (DMSO-d6) δ 7.51 (d, J = 7.2 Hz 1H), 7.01 (br d, J = 77.6 Hz 2H), 5.63 (d, J = 7.2 Hz, 1H), 5.32 (t, J = 5.2, 2H, vinyl), 3.2 to 4.4 (m, 10 H), 1.98 (m, 2H), 1.45 (m, 2H), 1.24 (br s, 30H), 0.85 (t, J = 6.5 Hz, 3H); 31P NMR δ 14.16; MS (ESI): m/z 618 (M + 2Na) +, 596 (M + Na)+; 572 (M − H)−.
(xii) Eicosyloxypropyl cidofovir, sodium salt.
Yield 66%; 1H NMR (CDCl3 + CD3OD) δ 7.46 (d, J = 7.0 Hz, 1H), 5.77 (d, J = 7.2 Hz, 1H), 3.79 to 3.86 (m, 4H), 3.60 to 3.75 (m, 1H), 3.53 to 3.59 (m, 2H), 3.40 to 3.44 (m, 2H), 3.30 to 3.32 (m, 2H), 3.24 to 3.26 (m, 2H), 1.74 to 1.78 (m, 2H), 1.40 to 1.47 (m, 2H), 1.17 (br s, 34H), 0.79 (t, J = 6.8 Hz, 3H). 31P NMR (CDCl3 + CD3OD) δ 16.87. MS (ESI): m/z 662 (M + 2Na) +, 640 (M + Na)+; 616 (M − H)−.
(xiii) 1-O-Octadecyl-2-O-benzyl-sn-glycero-3-CDV, sodium salt (ODBG-CDV).
Yield 85%; 1H NMR (CD3OD) δ 7.59 (d, 1H), 7.37 to 7.23 (m, 5H), 5.79 (d, 1H), 4.69, (d,2H), 3.98 (t, 2H), 3.97 (t, 2H), 3.82 to 3.51 (m,6H), 3.45 (d, 2H), 3.43 (t, 2H), 1.57 (m, 2H), 1.26 (br s, 30H), 0.88 (t, 3H); 31P NMR δ 17.11. MS (ESI): m/z 718 (M + H)+, 740 (M + Na)+, 694 (M − Na)+.
Determination of antiviral activity and drug cytotoxicity. (i) Cell cultures and viruses.
Human foreskin fibroblast (HFF) cells and mouse embryo fibroblast (MEF) cells were prepared as primary cultures and used in the HCMV and MCMV assays. Cells were propagated in minimal essential medium (MEM) containing 10% fetal bovine serum (FBS), 2 mM l-glutamine, 100 U of penicillin per ml and 25 μg of gentamicin per ml in T-175 cm2 tissue culture flasks (BD Falcon, Bedford, Mass.) until used in antiviral assays. HCMV strain AD-169 and MCMV strain Smith were propagated by using standard virological techniques.
(ii) Neutral red uptake assay for cytotoxicity.
HFF cells were seeded into 96-well tissue culture plates at 2.5 × 104 cells/well. After a 24-h incubation, medium was replaced with MEM containing 2% FBS, and drug was added to the first row and then diluted serially fivefold from 100 to 0.03 μM. The plates were incubated for 7 days, and cells were stained with neutral red and incubated for 1 h. Plates were shaken on a plate shaker for 15 min, and neutral red was solubilized with 1% glacial acetic acid-50% ethanol. The optical density was read at 540 nm. The concentration of drug that reduced cell viability by 50% (CC50) was calculated by using computer software. MEF cells were stained with neutral red and evaluated visually with a stereomicroscope at ×10 magnification. No toxicity was observed in the MCMV assays at the concentrations tested.
(iii) Plaque reduction assay for antiviral activity.
HFF or MEF cells were placed into 6- or 12-well plates and incubated at 37°C for 2 days (HFF) or 1 day (MEF). Ganciclovir and CDV were used as positive controls. Virus was diluted in MEM containing 10% FBS to provide 20 to 30 plaques per well. The medium was aspirated, and 0.2 ml of virus was added to each well in triplicate with 0.2 ml of medium added to drug toxicity wells. The plates were incubated for 1 h with shaking every 15 min. Drug was serially diluted 1:5 in MEM with 2% FBS starting at 100 μM and added to appropriate wells. Following a 7-day incubation for MCMV, or 8 days for HCMV, cells were strained for 6 h with 2 ml of 5.0% neutral red in phosphate-buffered saline. The stain was aspirated, and plaques were counted by using a stereomicroscope at ×10 magnification. By comparing drug-treated with untreated wells, 50% effective concentrations (EC50s) were calcuated in a standard manner.
RESULTS
Effect of alkyl chain length.
We synthesized a series of ether lipid CDV esters based on the propanediol linker with alkyl chains ranging from 8 to 20 carbon atoms (in addition to the 3 carbons of propyl and the oxygen heteroatom), making the total number of atoms, beyond the CDV phosphonate oxygen, range from 12 to 24. The antiviral activities against HCMV and MCMV were assessed for CDV and cCDV analogs (Table 1). In Table 1, note that the EC50s are in nanomoles/liter while the CC50 values are in micromoles/liter. Analogs having short alkyl chains of 12 to 16 atoms were not highly active, with EC50s of >1,000 to 10 nM. The most active compound was HDP-CDV with an EC50 versus HCMV and MCMV of 0.9 nM (20 atoms). As the alkyl chain is lengthened beyond 20 atoms, antiviral activity declines sharply. For example, the longest chain tested, eicosyloxypropyl-CDV (24 atoms), was roughly 1.5 logs less active than HDP-CDV, with an EC50s of 34 (HCMV) and 67 nM (MCMV). Against both HCMV and MCMV, a clear optimum in antiviral activity is seen at 20 atoms, with antiviral activity declining markedly with either shorter or longer chains. Generally similar results were noted with cCDV except that cCDV analogs are generally less active than CDV derivatives (Table 1).
TABLE 1.
Structure-activity assessment of alkoxylalkyl and alkyl esters of CDV and cCDV against HCMV and MCMV in vitroa
| Compound (no. of atoms: no. of double bonds) | Abbreviation | CC50 (μM/liter) | HCMV EC50 (nM/liter) | HCMV SI | MCMV EC50 (nM/liter) |
|---|---|---|---|---|---|
| Cidofovir Series | |||||
| Cidofovir | CDV | >317 ± 0 | 1,200 ± 430 | >264 | 40 ± 30 |
| Propanediol linkers | |||||
| Octyloxypropyl- (12) | OP-CDV | >1000 ± 0 | |||
| Dodecyloxypropyl- (16) | DDP-CDV | >190 ± 0 | 10 ± 6 | >19,000 | 15 ± 6 |
| Tetradecyloxypropyl- (18) | TDP-CDV | >100 | 2.0 ± 2.0 | >50,000 | 4.0 ± 0 |
| Hexadecyloxypropyl- (20) | HDP-CDV | 31 ± 21 | 0.9 ± 0.1 | 34,400 | 0.9 ± 0 |
| Octadecyloxypropyl- (22) | ODP-CDV | 44 ± 14 | 1.5 ± 0.7 | 29,300 | 3.0 ± 0.7 |
| Oleyloxypropyl- (22:1) | OLP-CDV | 87 ± 15 | 1.5 ± 0.7 | 58,000 | 4.0 ± 3.0 |
| Eicosyloxypropyl- (24) | ECP-CDV | 90.6 | 34 ± 29 | 2,670 | 67 ± 26 |
| Ethanediol linkers | |||||
| Octadecyloxyethyl- (21) | ODE-CDV | 18 ± 5 | 0.9 ± 0.1 | 2,000 | 1.0 ± 0 |
| Oleyloxyethyl- (21:1) | OLE-CDV* | 56 ± 29 | 1.2 ± 0.5 | 46,700 | 4.0 ± 4.0 |
| Glycerol linker | |||||
| 1-O-Octadecyl-2-O-benzyl-sn-glyceryl- | ODGB-CDV | 47 ± 24 | 3.0 ± 0.1 | 15,700 | 4.5 ± 0.7 |
| No linker | |||||
| Octyl- (8:0) | O-CDV | >100 | 5,250 ± 5,700 | >19 | >10,000 |
| Dodecyl- (12:0) | DD-CDV | >100 | 310 ± 190 | >322 | >10,000 |
| Hexadecyl- (16:0) | HD-CDV | >157 ± 58 | 5.0 ± 0 | >31,400 | 20 ± 0 |
| Eicosyl- (20:0) | EC-CDV | 38.5 | 0.7 | 55,000 | 5.0 ± 0 |
| Docosyl- (22:0) | DC-CDV | 79.6 | 4.0 | 19,900 | 85 ± 7 |
| Tetracosyl- (24:0) | TC-CDV | >100 | 60 ± 30 | >1,670 | 7,450 ± 640 |
| Cyclic cidofovir series | |||||
| Cyclic cidofovir | cCDV | >331 ± 0 | 3,100 ± 1,800 | >107 | 230 ± 70 |
| Propanediol linker | |||||
| Dodecyloxypropyl- (16) | DDP-cCDV | 63 ± 2.7 | 6.5 ± 2.0 | 9,690 | 13 ± 10 |
| Hexadecyloxypropyl- (20) | HDP-cCDV | >100 ± 0 | 42.7 ± 30 | 2,340 | 4.0 ± 0 |
| Octadecyloxypropyl- (22) | ODP-cCDV | 77 ± 3.7 | 6.0 ± 1.0 | 12,800 | 50 ± 50 |
| Oleyloxypropyl- (22:1) | OLP-cCDV | 46 ± 0.57 | 2.5 ± 0.7 | 18,400 | 7.0 ± 0 |
| Ethanediol linker | |||||
| Octadecyloxyethyl- (21) | ODE-cCDV | 68 ± 29 | 42.7 ± 4.9 | 1,600 | 3 ± 0 |
| Oleyloxyethyl- (21:1) | OLE-cCDV | 34 ± 11 | 1.5 ± 0.7 | 22,700 | 1 ± 0 |
| Glycerol linker | |||||
| 1-O-Octadecyl-2-O-benzyl-sn-glyceryl- | ODGB-cCDV | 77 ± 32 | 13 ± 4 | 5,920 | 25 ± 7 |
| No linker | |||||
| Hexadecyl- (16:0) | HD-cCDV | 83 ± 20 | 38 ± 16 | 2,180 | 130 ± 10 |
| Eicosyl- (20:0) | EC-cCDV | 86.1 | 14 ± 5 | 6,150 | 230 ± 10 |
| Docosyl- (22:0) | DC-cCDV | >100 | 41 ± 10 | >2,440 | 2,800 ± 1,700 |
| Tetracosyl- (24:0) | TC-cCDV | >100 | 290 ± 40 | >344 | 2,300 ± 1,300 |
Data are recorded as the means ± standard deviations of at least two determinations. Values without a standard deviation represent a single determination. SI, selective index (CC50/EC50).
Effect of introducing an unsaturation in the alkyl chain.
Compounds with long saturated alkyl chains generally lack fluidity which, in turn, may affect the physical and biochemical behavior of the molecule. Alkyl chains having one or more unsaturations are generally highly fluid with transition temperatures substantially lower than their saturated congeners. We synthesized ether lipid analogs having an 18 carbon chain with a 9,10 cis double bond, 1-oleyloxypropyl-CDV (OLP-CDV) or 1-O-oleyloxyethyl-CDV (OLE-CDV), and compared their antiviral activities and selectivities with the corresponding saturated 18 carbon analogs of CDV or cCDV. Against HCMV and MCMV, both ODP-CDV and OLP-CDV were equally active with EC50s of 1.5 to 3.0 nM and 1.5 to 4.0 nM, respectively, while 1-O-octadecyloxyethyl-CDV (ODE-CDV) and OLE-CDV were essentially equally active against both viruses (Table 1). OLE-cCDV was more active than ODE-cCDV against HCMV, with an EC50 of 1.5 versus 43 nM. A similar trend was noted with OLP-cCDV and ODP-cCDV which had EC50s of 2.5 versus 6 nM for HCMV and 7 versus 50 nM for MCMV. The selectivity indexes of the unsaturated cCDV compounds, OLP-cCDV and OLE-cCDV, was generally higher than for their noncyclic counterparts (Table 1).
Effect of linker type.
To assess the effect of glycerol, propanediol, and ethanediol as linkers between alkyl chains and CDV, we synthesized 1-O-octadecyl-2-O-benzyl-sn-glycero-CDV (ODBG-CDV), 1-O-octadecyloxypropyl-CDV, ODE-CDV, and the corresponding cCDV esters. When tested in vitro against HCMV and MCMV, ODP-CDV, ODE-CDV, and ODBG-CDV had EC50s of 0.9 to 1.5 nM and 1 to 4.5 nM compared with CDV, which gave EC50s of 1,200 and 40 nM, respectively (Table 1). The selectivity indexes for ODP-CDV, ODE-CDV, and ODBG-CDV were 29,300, 2,000 and 15,700, respectively, compared with >264 for CDV. Generally similar results were noted for these analogs against MCMV. However, ether lipid analogs of cCDV were slightly less active with EC50s for ODP-cCDV, ODE-cCDV, and ODBG-cCDV ranging from 6 to 43 nM against HCMV and 3 to 50 nM versus MCMV (Table 1). With the cCDV analogs, selectivity indexes were generally lower than observed for the corresponding analogs of CDV. Selectivity index values for ODP-cCDV, ODE-cCDV, and ODBG-cCDV ranged from 2,000 to 29,300 for HCMV. Thus, it seems clear that alkyl ether esters of either propanediol, ethanediol, or 2-sn-substituted glycerol serve equally well as the ether lipid moiety.
Effect of absent linker group.
To assess the importance of the oxygen atom in the alkoxyalkyl chain, we synthesized a matched series of straight chain alkyl ethers of CDV and cCDV and compared their activities with the corresponding alkyloxypropyl-CDVs and alkyloxypropyl-cCDVs. Alkyl esters of CDV were highly active and selective. The most active compound was eicosyl-CDV (EC-CDV) with EC50s versus HCMV and MCMV of 0.7 nM and 5 nM, respectively (Table 1). The cCDV series was less active. EC-cCDV had an EC50 of 14 nM versus HCMV and 230 nM against MCMV. Hexadecyl-cCDV also had significant activity (Table 1).
To compare the antiviral activities of the alkyloxypropyl-CDVs and the alkyl-CDVs (no linker CDVs), we plotted their antiviral activities against HCMV (Fig. 2A) and MCMV (Fig. 2B) versus the number of atoms linked to CDV.
FIG. 2.
Effect of alkyl chain length and linker on the antiviral activity of CDV esters against HCMV (left) and MCMV (right) in vitro. •, alkoxypropyl linker; ▾, alkyl chain without linker.
Surprisingly, the antiviral activities of the two classes of analogs against HCMV were nearly identical. There is a broad optimum between 18 to 22 atoms, with maximal activity noted for both types of compounds at chain lengths of 20 atoms (HDP-CDV and EC-CDV). HDP-CDV, ODE-CDV, and EC-CDV have subnanomolar EC50s versus HCMV (0.9, 0.9, and 0.7 nM). Similar results were noted against MCMV, except that the no linker series is somewhat less active. HDP-CDV was the most active anti-MCMV compound with an EC50 of 0.9 nM.
DISCUSSION
The alkyloxypropyl-, alkyloxyethyl-, and alkylglycerol-CDV and -cCDV analogs described above were designed to resemble lysophosphatidylcholine (LPC), a dietary lipid which is absorbed from the gastrointestinal tract substantially unchanged (21). We have assumed that the oxygen heteroatom that is two or three carbon atoms from CDV is essential to provide a resemblance to LPC and also as a potential metabolic cleavage site. We deliberately changed the alkyl ester at the sn-1 position of LPC to an ether to prevent cleavage by carboxyl ester hydrolase, a pancreatic digestive enzyme with lysophospholipase activity (17).
The structure-activity data clearly show that the antiviral activities against HCMV and MCMV are strongly dependent on the length of the alkyl or alkoxyalkyl ester attached to the phosphonate of CDV with optimal activity obtained at 20 atoms (Table 1 and Fig. 2). In the alkoxyalkyl series, linkers of propanediol, ethanediol, and glycerol having long chain alkyl ether residues are all effective; adding a double bond in the alkyl chain either has no effect or tends to increase antiviral activity and selectivity. Surprisingly, straight alkyl chains are equally active when matched for the number of atoms linked to the phosphonate of CDV (Fig. 2). Similar results were obtained when these analogs of CDV were evaluated in vitro against orthopoxviruses (13) and various strains of adenovirus (11). Similar effects of alkyl chain length were also noted in our previous studies on the antiviral activity of alkyloxypropyl- and alkylthiopropyl-esters of phosphoformate against human immunodeficiency virus type 1, herpes simplex virus type 1, and HCMV in vitro (3, 16).
We designed the alkoxyalkyl CDV analogs to provide a metabolism site and to mimic the ether glycerophospholipids which are naturally occurring components of cell membranes and have well-defined metabolic conversions. Highly effective CDV analogs, such as HDP-CDV, are taken up into cells and metabolized to the active metabolite, CDV-diphosphate, much more avidly than unmodified CDV. Cellular levels of CDV-diphosphate are 100 times greater after 48 h of exposure to HDP-CDV relative to that observed with CDV (1). The data in Table 1 suggest that this is probably also the case with the highly active alkyl CDV analogs such as EC-CDV. In contrast, unmodified CDV gains entry to the cell very slowly by fluid phase endocytosis, a process with a limited transport capacity (7).
Regarding the mechanism of enhanced antiviral action, we hypothesize that the short chain analogs of CDV are relatively inactive because they are water soluble and have a very limited capacity to self-associate with the plasma membrane of the cell. However, the highly active analogs having 16 to 22 atoms in the chain are believed to insert spontaneously into the outer leaflet of the plasma membrane and are subsequently transferred to the inner leaflet either spontaneously or by the action of a “flippase” (10, 12). Once in the inner leaflet of the plasma membrane, the highly active CDV analogs desorb spontaneously or under the action of cellular lipid transfer proteins and are metabolized by cellular esterases in a phospholipase C-like transformation which releases CDV intracellularly (18). We propose that the long chain CDV analogs have less antiviral activity due to one or more of the following: poor cell association, slow transbilayer movement due to reduced fluidity, and slower metabolism to CDV inside the cell. Once released by cellular enzymatic action, CDV is metabolized by cellular enzymes that catalyze anabolic phosphorylation to CDV-diphosphate, the active metabolite (1).
Alkoxyalkyl esters of CDV are orally bioavailable, presumably because of their resemblance to LPC (6, 18), and are orally active in lethal challenge models of orthopoxvirus disease (5, 20) and in models of CMV disease (4, 14). The oral bioavailability of the straight chain alkyl esters of CDV (no linker series) has not yet been evaluated. Clearly, these novel prodrug approaches are worth pursuing for CDV. The same strategy can be readily applied to other antiviral and antiproliferative acyclic phosphonate nucleosides.
Acknowledgments
This work was funded in part by PHS contracts NO1-AI-85347 and NO1-AI-30049 from the National Institutes of Health and the National Institute of Allergy and Infectious Diseases (E.R.K.); National Institutes of Health grants EY11834 and EY07366 from the National Eye Institute and AI29164 from the National Institute of Allergy and Infectious Diseases; and by the Department of the Army, grant DAMD 17-01-2-007 (K.Y.H.). The U.S. Army Medical Research Acquisition Activity, Fort Detrick, Md., is the awarding acquisition office. The content of this article does not necessarily reflect the position or policy of the government, and no official endorsement should be inferred.
REFERENCES
- 1.Aldern, K. A., S. L. Ciesla, K. L. Winegarden, and K. Y. Hostetler. 2003. The increased antiviral activity of 1-O-hexadecyloxypropyl-cidofovir in MRC-5 human lung fibroblasts is explained by unique cellular uptake and metabolism, Mol. Pharmacol. 63:678-681. [DOI] [PubMed] [Google Scholar]
- 2.Beadle, J. R., C. B. Hartline, K. A. Aldern, N. Rodriquez, E. Harden, E. R. Kern, and K. Y. Hostetler. 2002. Alkoxyalkyl esters of cidofovir and cyclic cidofovir exhibit multiple-log enhancement of antiviral activity against cytomegalovirus and herpesvirus replication in vitro. Antimicrob. Agents Chemother. 46:2381-2386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Beadle, J. R., G. D. Kini, K. A. Aldern, M. F. Gardner, K. Wright, and K. Y. Hostetler. 1998. Alkylthioglycerol prodrugs of foscarnet: oral bioavailability and structure-activity studies in human cytomegalovirus,herpes simplex virus type-1 and human immunodeficiency virus type-1 infected cells. Antivir. Chem. Chemother. 9:33-40. [DOI] [PubMed] [Google Scholar]
- 4.Bidanset, D. J., J. R. Beadle, W. B. Wan, K. Y. Hostetler, and E. R Kern. 2004. Oral activity of ether lipid ester prodrugs of cidofovir against experimental human cytomegalovirus infections. J. Virol. 190:499-503. [DOI] [PubMed] [Google Scholar]
- 5.Buller, R. M., G. Owens, J. Schriewer, L. Melman, J. R. Beadle, and K. Y. Hostetler. 2004. Efficacy of oral active ether lipid analogs of cidofovir in a lethal mousepox model. Virology 318:474-481. [DOI] [PubMed] [Google Scholar]
- 6.Ciesla, S. L., J. Trahan, K. L. Winegarden, K. A. Aldern, G. R. Painter, and K. Y. Hostetler. 2003. Esterification of cidofovir with alkoxyalkanols increases oral bioavailability and diminishes drug accumulation in kidney, Antivir. Res. 59:163-171. [DOI] [PubMed] [Google Scholar]
- 7.Connelly, M. C., B. L. Robbins, and A. Fridland. 1993. Mechanism of uptake of the phosphonate analog (S)-1-(3-hydroxy-2-phosphonylmethoxy-propyl) cytosine (HPMPC) in vero cells. Biochem. Pharmacol. 46:1053-1057. [DOI] [PubMed] [Google Scholar]
- 8.Cundy, K. C. 1999. Clinical pharmacokinetics of the antiviral nucleoside analogues cidofovir and adefovir. Clin. Pharmacokinet. 36:127-143. [DOI] [PubMed] [Google Scholar]
- 9.De Clercq, E. 2003. Clinical potential of the acyclic nucleoside phosphonates cidofovir, adefovir and tenofovir in treatment of DNA virus and retrovirus infections. Clin. Microbiol. Rev. 16:569-596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gummad, N., and A. K. Menon. 2002. Transbilayer movement of dipalmitoyl-phosphatidylcholine in proteoliposomes reconstituted from detergent extracts of endoplasmic reticulum. J. Biol. Chem. 277:25337-25343. [DOI] [PubMed] [Google Scholar]
- 11.Hartline, C. B., K. M. Gustin, W. B. Wan, S. L. Ciesla, J. R. Beadle, K. Y. Hostetler, and E. R. Kern. Activity of ether lipid ester prodrugs of acyclic nucleoside phosphonates against adenovirus replication in vitro. J. Infect. Dis., in press. [DOI] [PubMed]
- 12.Kawashima, Y., and R. M. Bell. 1987. Assembly of the endoplasmic reticulum phospholipid bilayer. J. Biol. Chem. 262:16495-16502. [PubMed] [Google Scholar]
- 13.Keith, K. A., W. B. Wan, S. L. Ciesla, J. R. Beadle, K. Y. Hostetler, and E. R. Kern. 2004. Inhibitory activity of alkoxyalkyl and alkyl esters of cidofovir and cyclic cidofovir against orthopoxvirus replication in vitro. Antimicrob. Agents Chemother. 48:1869-1871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kern, E. R., D. J. Collins, W. B. Wan, J. R. Beadle, K. Y. Hostetler, and D. C. Quenelle. 2004. Oral treatment of murine cytomegalovirus infections with ether lipid esters of cidofovir. Antimicrob. Agents Chemother. 48:3516-3522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kern, E. R., C. B. Hartline, E. A. Harden, K. A. Keith, N. Rodriguez, J. R. Beadle, and K. Y. Hostetler. 2002. Enhanced inhibition of orthopoxvirus replication in vitro by alkoxyalkyl esters of cidofovir and cyclic cidofovir. Antimicrob. Agents Chemother. 46:991-995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kini, G. D., J. R. Beadle, H. Xie, K. A. Aldern, D. D. Richman, and K. Y. Hostetler. 1997. Alkoxy propane prodrugs of foscarnet: effect of alkyl chain length on in vitro antiviral activity in cells infected with HIV-1, HSV-1 and HCMV. Antivir. Res. 36:43-53. [DOI] [PubMed] [Google Scholar]
- 17.Lombardo, D., J. Fauvel, and O. Guy. 1980. Studies on the substrate specificity of a carboxyl ester hydrolase from human pancreatic juice. I. Action on carboxyl esters, glycerides and phospholipids. Biochim. Biophys. Acta 611:136-146. [DOI] [PubMed] [Google Scholar]
- 18.Painter, G. R., and K. Y. Hostetler. 2004. Design and development of oral drugs for the prophylaxis and treatment of smallpox infection. Trends Biotechnol. 22:423-427. [DOI] [PubMed] [Google Scholar]
- 19.Plosker, G. L., and S. Noble. 1999. Cidofovir: a review of its use in cytomegalovirus retinitis in patients with AIDS. Drugs 58:325-345. [DOI] [PubMed] [Google Scholar]
- 20.Quenelle, D. C., D. J. Collins, K. Y. Hostetler, J. R. Beadle, W. B. Wan, and E. R. Kern. 2004. Oral treatment of cowpox and vaccinia infections in mice with ether lipid esters of cidofovir. Antimicrob. Agents Chemother. 48:404-412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Scow, R. O., Y. Stein, and O. Stein. 1967. Incorporation of dietary lecithin and lysolecithin into lymph chylomicrons in the rat. J. Biol. Chem. 242:4919-4924. [PubMed] [Google Scholar]