Targeting nucleotide dimers containing antiviral nucleosides

Curr. Med. Chem. - Anti-Infective Agents, 2005, 4, 37-54

37

Targeting Nucleotide Dimers Containing Antiviral Nucleosides L. Rossi1, S. Serafini1, P. Franchetti2, L. Cappellacci2, A. Fraternale1, A. Casabianca1, G. Brandi3, F. Pierigé1, C-F Perno4,5, E. Balestra6, U. Benatti7, E. Millo8, M. Grifantini2 and M. Magnani1,* 1,*

Institute of Biochemistry “G. Fornaini”, University of Urbino “Carlo Bo”, Urbino, Italy, 2Department of Chemical Sciences, University of Camerino, Camerino, Italy, 3Institute of Hygiene, University of Urbino “Carlo Bo”, Urbino, Italy, 4Department of Experimental Medicine, University of Rome “Tor Vergata”, Rome, Italy, 5IRCCS “L. Spallanzani”, Rome, Italy, 6Institute of Microbiology, University of Rome “La Sapienza”, Rome, Italy, 7Department of Experimental Medicine Biochemistry Section, University of Genoa and 8Institute “Giannina Gaslini”, Genoa, Italy Abstract: Among the antiviral agents developed for the treatment of human viral infections, nucleoside analogs represent the largest group. However, much remains to do to improve their pharmacokinetic properties, to increase their efficacy, to reduce the selection of drug-resistent strains and to reduce their toxic side effects. Towards this end many nucleotide dimers have been synthesized in the last years in several laboratories. Such compounds have several advantages compared to the administration of nucleoside analogs as single drugs: 1) can act as prodrugs for a slow delivery of monomers in circulation; 2) can be encapsulated into autologous erythrocytes to perform as bioreactors converting a non diffusible dimer into a diffusible nucleoside analog to be released in circulation; 3) can be targeted to macrophages by proper drug targeting systems; 4) can overcome the limiting phosphorylating activities of several infectable cell types; 5) can have the advantage of a combination therapy with the administration of a single compound. In this review, dimers developed in our laboratory will be reported. In particular, the heterodinucleotide AZTpPMPA and the homodinucleotide Bis-PMEA are shown to be able to act as prodrugs when administered to mice releasing the single monomer in circulation. The homodinucleotide AZTp2AZT and the dimer AZTp2EMB once encapsulated in human erythrocytes are converted by erythrocyte enzymes into diffusible nucleosides and slowly released from the carrier cells. The dimers AZTp2AZT, AZTp2ACV, ACVpPMPA, AZTpPMPA and Bis-PMEA were targeted to macrophages where a very effective protection against virus replications was obtained. Thus, nucleotide dimers could be used as effective prodrugs for drug delivery in the treatment of viral infections improving the pharmacokinetic of single moieties and can be efficiently targeted to selected cell types with intracellular release of a phosphorylated (active) nucleoside.

Keywords: Homodinucleotide, heterodinucleotide, anti HIV-1 activity, anti HSV-1 activity, pharmacokinetic properties, erythrocytes, human macrophages, murine AIDS. INTRODUCTION The most common therapeutic strategies against viral infections have been based on the administration of nucleoside analogs alone or with other antiviral agents. Once phosphorylated to the 5’-triphosphate form, nucleotide analogs are potent specific inhibitors of viral enzymes necessary for viral replication. However, much remains to be done to improve the pharmacokinetic properties of nucleoside analogs. In addition, at the cellular level, the main problems involved in the use of such drugs are represented by both their limited phosphorylation in some cells (e.g. antiretroviral drugs in macrophages) and the toxic side effects of the corresponding triphosphates usually due to inhibition of not only viral enzymes but enzyme targets in the host cell as well. For this reason, the formation of these molecules should be limited to low concentrations so as to inhibit viral enzymes selectively. Therefore, alternative strategies which reduce drug toxicity, increase antiviral efficacy and improve the pharmacokinetic properties of these *Address correspondence to this author at the Istituto di Chimica Biologica “G. Fornaini”, Università degli Studi di Urbino “Carlo Bo”, Via Saffi, 261029 Urbino, Italy; Tel: +39-722-305211; Fax: +39-722-320188; Email: [email protected] 1568-0126/05 $50.00+.00

drugs allowing prolonged application intervals, are needed. Prompted by these considerations, several homo- and heterodimers of nucleoside analogs have been synthesized as prodrugs in different laboratories. However, some dimers are nonhydrolyzable due to the presence of an aliphatic spacer [1] or to a methylphosphonate linkage [2-4]. In some reports [3, 4], the nucleotide dimers were linked by a 5’,5’phosphate bond and displayed antiviral activity similar to or better than that of the individual nucleoside analogs. However, it is not clear whether these dimers function as such or following conversion to the corresponding moieties, since the cleavage of the phosphodiester bond is very likely to take place. In these studies, the presence of intact dimer in cells was never demonstrated. Recently, amphiphilic heterodinucleotide phosphates containing ddC (dideoxycytidine) and AZT (azidothymidine) have been synthesized. These dimers posses both more favorable pharmacokinetic properties than monomers and an increased antiviral potency [5, 6]. Herein, we report selected examples of antiviral doubledrugs investigated in our laboratory and available methods to improve their delivery. Dimers were designed as metabolically suitable prodrugs, structurally tailored to be © 2005 Bentham Science Publishers Ltd.

38

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

converted in active drugs by endogenous enzymes. In most cases these prodrugs were encapsulated into erythrocytes which performed as slow delivery systems or as drug targeting systems. All the dinucleotides (AZTp2AZT; AZTp2ACV; AZTp2EMB; AZTpPMPA; ACVpPMPA and Bis-PMEA) in fact, contained a pyrophosphate or P-C-P bridge designed to be hydrolyzed by: a) serum enzymes, when dimers were used as free prodrugs; b) erythrocyte enzymes, when dimer-loaded erythrocytes were used as bioreactors for a slow delivery of monomers; c) macrophage enzymes, when dimer-loaded erythrocytes were used as a delivery system to specifically target prodrugs to the macrophagic compartment. At first, chemistry of the nucleotide dimers synthesized by us is briefly described and successively the above mentioned points are discussed. CHEMISTRY OF NUCLEOTIDE DIMERS Some dimers containing nucleosides and other compounds synthesized in recent years as antiviral agents are summarized in Fig. 1. The structures that characterize the reported compounds are: 1.

Nucleotide dimers comprising two nucleoside analogs bound together by a 5’,5’-phosphate bridge (i.e. compounds 7 and 8) [4].

2.

Nucleotide dimers comprising two nucleoside analogs bound together by a 3’,5’-phosphate bridge (i.e. compounds 9 a, b and 10 a, b) [5].

3.

Nucleotide dimers comprising two nucleoside analogs bound together by a 5’,5’-pyrophosphate bridge (i.e. compounds 1, 2, 4-6) [8, 9, 11-13].

4.

Dimers containing a nucleoside analog and a further antimicrobial agent bound by a pyrophosphate bridge (i.e. compound 3) [10] or a nucleoside analog and a nonnucleoside reverse transcriptase inhibitor bound by a 1,n-diaminoalkane linkage linked at the C-5 position of the nucleoside analog and at the C-2 position of the nonnucleoside inhibitor (i.e. compound 11) [7].

Synthesis of selected dinucleotides obtained in our laboratories is reported below. The homodinucleotide P1,P2-bis[thymine-3’-azido-2’,3’dideoxy-β-D-riboside-5’] pyrophosphate (AZTp2AZT) consisting of two molecule AZT bound together by a pyrophosphate bridge (Fig. (1), compound 1), was synthesized by coupling the diphenylpyrophosphate derivative of AZT with the mono tri-butylammonium salt of AZT-MP together with hexamethylphosphotriamide in dry pyridine, as reported in [8]. The heterodinucleotide AZTp2ACV [P1-(thymine-3’azido-2’,3’-dideoxy-β-D-riboside-5’-P2-guanine-9-(2-hydroxyethoxymethyl)pyrophosphate] is a double-drug derived from AZT and 9-(2-hydroxyethoxymethyl)guanine (ACV) bound together by a pyrophosphate bridge (Fig. (1), compound 2). The dinucleotide was synthesized by coupling AZT 5’-monophosphate as an activated phosphoromorpholidate obtained starting from AZT-MP with morpholine and dicyclohexylcarbodiimide (DCCI) in aqueous terbutanol, with the tri-butylammonium ACV-MP salt [9].

Magnani et al.

In AZTp2EMB, AZT and ethambutol (EMB) were bound together by a pyrophosphate bridge (Fig. (1), compound 3). The AZTp2EMB [P1-thymine-3’-azido-2’,3’-dideoxy-β-Driboside-5’-P2-(+)2,2`-(ethylendiimino)di-1-butanol pyrophosphate] was obtained by condensation of the morpholidate derivative of AZT-MP with the trioctylammonium EMBMP salt and the final product was analyzed by mass spectrometry, as reported in [10]. In the Scheme 1, the synthesis of a new double-drug comprising AZT and the nucleoside acyclic phosphonate 9[(R)-2-(phosphonomethoxy)propyl]adenine [(R)-PMPA] was reported. In the heterodinucleotide AZTpPMPA [P 1-(thymine3’-azido-2’,3’-dideoxy-β-D-riboside-5’-phosphate)-P2-(9-(R)2-(phosphonomethoxypropyl)adenine] AZT and (R)-PMPA was bound together by a phosphate bridge (Fig. (1), compound 4) [11]. The dinucleotide was obtained by coupling AZT 5'-monophosphate as an activated derivative with the tri-octylammonium (R)-PMPA salt. AZT 5’monophosphate was synthesized as a morpholidate derivative as described above and purified by flash chromatography on silica gel. This nucleotide was reacted with (R)-PMPA tri-octylammonium salt obtained starting from free acid PMPA with octylammine in dry methanol. Condensation was carried out in anhydrous condition in pyridine. The crude compound was purified by flash chromatography on silica gel and the dimer was obtained as a disodium salt. In ACVpPMPA, Acyclovir and (R)-PMPA were bound by a mixed phosphate-phosphonate anhydride (Fig. (1), compound 5). The heterodinucleotide (P 1-(6H-purin-6-one,2amino-1,9-dihydro-9-[(2-(phosphonooxy)ethoxy]methyl]-P2(9-(R)-2-(phosphonomethoxypropyl)adenine was obtained by condensation of the morpholidate derivative of (R)-PMPA with the tri-butylammonium ACV-MP salt (molar ratio 1/1) in anhydrous pyridine (Scheme 2) [12]. The dimer was purified by chromatography on silica gel and obtained as diammonium salt. The homodinucleotide P1,P2-bis[2-(adenin-9H-yl)ethoxymethyl]phosphonate (Bis-PMEA) consists of two molecules of 9-[2-(phosphonomethoxy)ethyl]adenine (Fig. (1), compound 6) was also developed. This was synthesized starting from the morpholidate derivative of PMEA as N,N’dicyclohexyl-4-morpholine-carboxamidinium salt, by coupling to PMEA as mono tri-butylammonium salt in dry N,N-dimethylformamide at 50 °C. The title dimer was obtained as a diammonium salt after chromatography on silica gel column. In a similar way, the radiolabeled dimer [14C]Bis-PMEA was prepared by coupling the mono-tributylammonium *PMEA(adenine-8-14C) salt to PMEA morpholidate [13]. Structures of the described dinucleotides was assigned on the basis of their analytical and spectroscopic data by 1H-, 31P-NMR and MS spectra. NUCLEOTIDE AGENTS

DIMERS

AS

ANTIMICROBIAL

Nucleotide dimers as such, or dimers containing non hydrolysable bounds, have shown very limited antimicrobial properties. In contrast, when nucleotide dimers dissociate in body fluids by cleavage of their phosphate bounds, they become biologically active and result in improved

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

O

O H3C

O

O

N

CH3

HN

NH

O

O

P

O O

O

P

OH

O

N O

OH

N3

N3 AZTp2AZT (1)

O

O O H2N

O

O

N

N

CH3

HN

N

HN

O

O

P

O

P

OH

O

N O

OH N3

AZTp2ACV (2) O C2H5

CH2OH HC NH

CH2

CH2

NH

C2H5

CH3

HN

CH

O

O

CH2

O

P

O

P

O O

N O

OH

OH

N3

AZTp2EMB (3)

NH2

O H3C

N

NH O

O

N

O

O

O

P

O

NH2 O

N

N

HN

O H2N

N

O

O

P

O

OH

P

N

O

OH CH3

N3

N AZTpPMPA (4)

N

N

N O

N

P

OH

N

O

OH CH3

ACVpPMPA

(5) NH2

NH2 N

N N

N

N

O

O O

P OH

O

P OH

Bis-PMEA (6)

N O

N N

39

40

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Magnani et al.

(Fig. 1). contd….

O

O N

HN

HN O

N

N

O

O

O

P

O

O

NH 2 CH3

N

N

N

O

N

N

O

O

O

OH

O

P

O

N3

AZTpddI (7)

AZTpddA (8) NHR

NHR N

N O

O

N HO

O

NH 2

CH3

HN O

O O P

N O

O HN

HO

N O

OH N3

HO

CH3

HN

O

O

N O

O

a: R =

CO(CH2) 14CH3

b: R =

CH2(CH2) 14CH3

HO

P

O

N O

O N3

N3 N4-pamdC-AZT (9a)

N4-pamdC-ddC (10a)

N4-hxddC-AZT (9b)

N4-hxddC-ddC (10b) CH3

O

O

O HO

N

(CH2) n

HN

N H

N H

N

N

N O [d4T]-NH-(CH2) n -NH-[imidazo[1,5-b]pyridazine (11)

Fig. (1). Structures of the described antiviral double-drugs.

pharmacokinetic properties of their single moieties. Furthermore, the intracellular administration of nucleoside dimers results in the formation of 5’-phosphate derivatives of the single nucleotide, overcoming the limited phosphorylating activities of several cell types and thus increasing their antimicrobial activity (i.e. nucleoside analogues should be phosphorylated by cell resident enzymes to be converted into active, phosphorylated forms). To better describe these concepts we will illustrate the possible advantages of using nucleotide dimers. Nucleotide dimers can be used as: 1. Prodrugs that are converted by body fluids into single moieties. 2. Prodrugs that can be delivered encapsulated into autologous erythrocytes which contain enzymes able to convert the dimers into free-diffusible nucleoside analogues that are slowly released in circulation.

3. Prodrugs that are first encapsulated into autologous erythrocytes, which are then modified to be recognized and phagocytosed by tissue macrophages. Once the dimer is (by way of erythrocyte phagocytosis) within the macrophage, is cleaved into two phosphorylated nucleoside analogs overcoming the limited drug phosphorylating activity of macrophages. The use of nucleotide dimers according to points 1 to 3 are now illustrated with selected examples Fig. (2). NUCLEOTIDE DIMERS AS PRODRUGS Pharmacokinetic and Antiviral Activity of Bis-PMEA PMEA is an antiviral drug with activity against herpes viruses (HSV-1), Epstein-Barr virus and retroviruses, including the human immunodeficiency virus (HIV-1). Thus, PMEA is of interest both as a potential antiretroviral drug for HIV-1 infections and also for the treatment of some of the

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1 NH 2

O H3 C

N

NH O

O

N

O

O

O

P

+

OH

- O

Oct 3NH

+

N

P

OH

N N

O

OH

CH3

N3 AZT-MP

(R)-PMPA

i

ii

NH 2

O H3 C

N

NH O

O

N

O

O

O

P

O

N

P

N N

O

OH

OH

CH3

N3

AZTpPMPA

Scheme 1. Reagents: i) morpholine, DCCI, aqueous tert-BuOH, reflux; ii) dry pyridine, r.t.

NH 2

O N

N N

N O

N O

P

OH

N

+ Bu3 NH - O

+

P

O

O

OH

OH

CH3

ACV-MP

(R)-PMPA i

ii

NH 2

O N

N N

NH

O

N O

N

N O CH3

P OH

NH

O O

P

O

O

OH

ACVpPMPA

Scheme 2. Reagents: i) morpholine, DCCI, aqueous tert-BuOH, reflux; ii) dry pyridine, r.t.

N

NH 2

N

NH 2

41

42

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Magnani et al.

Fig. (2). Possible use of nucleotide dimers.

opportunistic infections associated with AIDS [14]. Unfortunately, oral PMEA administration, as required for long-term therapy, is hindered by its low bioavailability due to the negative charge (at physiological pH) of the phosphonyl group. In order to mask the phosphonate charge and enhance the lipophilicity of the molecule, a large number of prodrugs have been designed. Among these, the bis (pivaloyloxymethyl) [bis(POM)] ester prodrug was selected as a potentially useful prodrug of PMEA and clinical trials were performed [15]. However, some adverse effects due to the pivaloyl moiety of the prodrug have been observed. We have suggested a novel approach based on a compound

consisting of two PMEA molecules bound together by a PO-P bond (Bis-PMEA) able to act as prodrug for a slow release of PMEA in circulation. The oral bioavailability of Bis-PMEA in mice and its antiretroviral activity in a murine model of immunodeficiency (murine AIDS, MAIDS) were evaluated. Pharmacokinetic studies were performed in ICR mice and the plasma concentrations of monomer evaluated at different times from Bis-PMEA administration, as described in detail in [13]. Radiolabeled compounds ([14C]PMEA and [14C]Bis-PMEA) were used. Mice received either

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

[14C]PMEA by intravenous (i.v.) bolus injection (via the retro orbital sinus) or by oral gavage (o.g.), or [14C]BisPMEA o.g., all given at equimolar doses of PMEA (28 mg/kg). Pharmacokinetic evaluations were carried out according to the linear trapezoidal rule, evaluating the area under the curve from zero to the time of the last measurable concentration (AUCt0→tlast). As shown in Fig. (3), the drug concentration-time curve after the i.v. bolus injection of [14C]PMEA showed a rapid and biphasic decline, while concentrations of PMEA in plasma were still detectable up to 2 days after o.g. administration of [14C]PMEA or [14C]BisPMEA. Moreover, a significant higher bioavailability of PMEA following o.g. Bis-PMEA administration (50.8%) with respect to the administration of PMEA (13.5%) was obtained, suggesting that Bis-PMEA administered orally (per os, p.o.) can act as a prodrug providing a slow delivery of PMEA in circulation.

Fig. (3). Pharmacokinetic of Bis-PMEA. Concentrations of PMEA in the plasma of mice after intravenous bolus injection of [14C] PMEA or oral gavage of [14C] Bis-PMEA or oral gavage of [ 14C] PMEA. All compounds were given at a dose equivalent to 28 mg of PMEA/kg. Data are the average values ± standard deviation for three independent experiments (three mice per time point in each experiment). Reprinted with permission from: Rossi L., et al. J. Antimicrob. Chemother., 2002, 50, 365. © 2002 The British Society for Antimicrobial Chemotherapy.

The antiretroviral activity of Bis-PMEA administered orally was tested in C57BL/6 mice infected with the LPBM5 retrovirus complex that causes a disease with a pathology that resembles that of human AIDS [16], including abnormal T- and B-lymphocyte functions, polyclonal B-cell proliferation, lymphoadenopathy, splenomegaly, hypergammaglobulinemia and enhanced susceptibility to infections [17, 18]. Bis-PMEA and PMEA were administered to infected mice by oral gavage at equimolar doses of PMEA at a concentration of 50 mg/kg. After 4 and 9 weeks postinfection, mice were sacrificed and several parameters characterizing the progression of the disease were evaluated.

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

43

The results obtained (summarized in Fig. (4)) show that BisPMEA was more effective than PMEA in reducing hypergammaglobulinemia (at 4 weeks post infection) and lymphoadenopathy. The same result was observed following evaluation of lymph node BM5 proviral DNA content after 4 weeks of treatment where a 54 ± 29% and 96 ± 4% of inhibition with PMEA and Bis-PMEA per os were obtained, respectively. Neither PMEA nor Bis-PMEA was able to restore the ability of both T and B cells to proliferate in vitro. Furthermore, the evaluation of percentage of CD19+ lymphocytes in the spleen revealed that Bis-PMEA did not cause a reduction in the number of spleen lymphocytes in infected mice as instead observed by PMEA administration. It is worth noting that PMEA was not approved by the FDA as an anti-HIV-1 drug because its renal toxicity [19-21], however in our experimental model, we were unable to detect alterations in either plasma creatinine concentration or kidney histological examination in all conditions tested, probably as a consequence of the short treatment regimen.

Fig. (4). Effect of oral PMEA or Bis-PMEA administrations on serum IgG levels and lymph node weights in LP-BM5-infected C57BL/6 mice. (A) Serum IgG in control mice, LP-BM5-infected mice and infected mice treated with per os (p.o.) administration of PMEA or Bis-PMEA. Drugs were administered at a dose equivalent to 50 mg of PMEA/kg. Serum IgG levels were determined 4 or 9 weeks post infection by enzyme-linked immunosorbent assay. Values are the means ± standard deviation of duplicate determinations on three mice (4 weeks post infection) and on 7 mice (9 weeks post infection). ap=0.0077 (vs infected); bp=0.0625 (vs infected). (B) Lymph nodes weight. PMEA and Bis-PMEA were given at equimolar doses of PMEA at a concentration of 50 mg/kg. Drugs were administered by p.o. five days per week for a period of 4 or 9 weeks. Values are the means ± S.D. of 3 animals determined 4 weeks post virus inoculation and 7 animals at 9 weeks post virus inoculation.

44

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Magnani et al.

In conclusion, our results show that Bis-PMEA administered by oral gavage can act as a prodrug for a better delivery of PMEA in circulation than that obtained with oral gavage PMEA, assuring higher antiviral efficacy without causing significant toxicity. Pharmacokinetic and Antiviral Activity of AZTpPMPA The PMEA-related compound (R)-PMPA (tenofovir) is a highly potent and selective antiretroviral agent that received U.S. Food and Drug Administration (FDA) approval (as tenofovir disoproxil fumarate, Viread) [22] for the treatment of HIV-1 infection when taken in combination with other antiretroviral agents. PMPA is phosphorylated by cellular enzymes to form the active metabolite PMPA diphosphate [23] that exhibits a long intracellular half-life in both resting and activated peripheral blood mononuclear cells, allowing thus once-daily dosing. Moreover, it was reported that PMPA demonstrated strong synergistic anti-HIV-1 activity in combination with AZT in vitro, but minor synergistic inhibition of HIV-1 replication in combination with other antiretroviral drugs such as didanosine (ddI) and nelfinavir [24]. However, it is possible that during combination therapy drug interactions may arise in vivo that would influence their antiviral activity. To investigate in vivo PMPA synergistic antiretroviral activity in combination with AZT, experiments involving combination of PMPA and AZT were performed in a murine model of immmunodeficiency (MAIDS). Moreover, since AZT has a low serum elimination half-life (t1/2) of about 1 hour and frequent daily administrations of this drug are needed in humans to maintain therapeutically useful drug levels (thus causing minor patients compliance) we evaluated if the single drug AZTpPMPA could act as a prodrug for a slow delivery of AZT and PMPA in circulation. Subsequently, the antiretroviral activity of AZTpPMPA was evaluated and compared with that obtained following both AZT plus PMPA and AZT or PMPA, administered as single drugs [24]. The intraperitoneal and the oral route of administration of the single drugs (AZT plus PMPA) and of the heterodimer (AZTpPMPA) were first compared in order to choose the one giving the highest and/or maintained levels of drugs in circulation. Two groups of ICR mice were treated intraperitoneally with 7.5 µmoles of AZTpPMPA (0.3 g/kg) or of AZT plus PMPA as single drugs (0.12 and 0.13 g/kg, respectively), while other two groups were treated orally with the same amounts of drugs. Values for the area under the curve, calculated in the range 0.25-6 h show the advantage of intraperitoneal administration compared with oral gavage (Table 1) [25]. Table 1.

The ability of the heterodinucleotide AZTpPMPA to act as an antiretroviral prodrug and thus to inhibit MAIDS, was tested and compared with the administration of free drugs AZT and PMPA alone or in combination. The drugs were administered intraperitoneally at equimolar dosage (2.8 µmol/mouse) to four groups of infected mice: the first treated with AZTpPMPA, the second with AZT plus PMPA, the third with AZT and the last one with PMPA. As control, uninfected/untreated and infected/untreated mice were used. After 10 weeks of infection, mice were sacrified and several parameters characterizing the progression of the disease were evaluated. AZTpPMPA in MAIDS was able to markedly reduce lymphoadenopathy, splenomegaly and lymph node BM5 proviral DNA content (88%, 64% and 49%, respectively) and partially (40%) hypergammaglobulinemia. Moreover, the tissue histological examinations have shown that upon AZTpPMPA administration, liver and spleen were as controls and no signs of drug toxicity were found. Furthermore, an increase in red blood cell number, hematocrit, hemoglobin concentration and white cell count was observed compared with infected/untreated mice. However, the antiviral efficacy of AZTpPMPA was similar to that obtained upon administration of AZT in combination with PMPA and, in addition, with PMPA alone, results similar to those of the combined therapy (AZT plus PMPA) were obtained, excluding possible synergic effects between AZT and PMPA. In conclusion, the results obtained show that AZTpPMPA is able to perform as a prodrug for slow delivery of AZT and PMPA in circulation but it does not appear to be more efficient than AZT plus PMPA or even PMPA alone, at least in murine AIDS. However, the favorable pharmacokinetics, absence of toxicity and significant antiviral activity of this new heterodinucleotide suggest a possible alternative to the use of new combination therapies based on single molecules consisting of two different antiviral drugs each with different intracellular pharmacokinetics. NUCLEOTIDE DIMER-LOADED ERYTHROCYTES AS BIOREACTORS FOR A SLOW MONOMERS DELIVERY Properly engineered erythrocytes (RBC) can behave as slow delivery systems for drugs, improving their pharmacokinetic patterns and therapeutic performances [2630]. Red blood cells possess unique characteristics that make them useful as drug delivery system. Erythocytes are readily available in great quantities, are biocompatible (when

AUC Determination upon AZTpPMPA or AZT plus PMPA Intraperitoneally or Orally Administered AUC(0.25-6h) (nmol x h/ml) Drug

AZTpPMPA

AZT plus PMPA

AZTpPMPA

INTRAPERITONEAL (i.p.)

AZT plus PMPA

ORAL GAVAGE (p.o.)

AZT

23,739

34,375

41,477

27,326

PMPA

32,333

31,422

4,005

1,696

Reprinted with permission from: Rossi L., et al. J. Antimicrob. Chemother., 2002, 50, 639. © 2002 The British Society for Antimicrobial Chemotherapy.

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

autologous RBC are used), are completely biodegradable and have a large capacity so that a high percentage of encapsulation can be obtained. Furthermore, RBC are not inert carriers but active ones, being endowed with several enzymatic activities that can act directly to the loaded molecules, modifying them. Several methods are used to encapsulate molecules in RBC. Critically comparing these encapsulation methods to evaluate their loading efficiency [31], that one based on hypotonic hemolysis has shown several advantages: high total amount of drug encapsulated, high percent ratio of the loaded amount and high cell recovery. The procedure is based on the remarkable property of the RBC to increase in volume when placed in the presence of a hypotonic solution; during this step membrane pores opening occurs and externally added drugs can cross the pores and enter in RBC. Finally, by raising the salt concentration to its original level, the membrane resealed and the RBC can reassume their normal biconcave shape and impermeability features. Then, the non-entrapped substances are washed out and loaded RBC are ready to be used as carriers for the delivery of the encapsulated drug. Here, we describe the ability of erythrocytes loaded with impermeant dimers to slowly release the corresponding diffusible monomers. Two AZT analogs were synthesized and encapsulated in RBC: the heterodimer AZTp2EMB designed for HIV-1 and MAC (Mycobacterium avium complex) coinfection, consisting of an antiretroviral (AZT) and an antimicobacterial (EMB, ethambutol) agent, and the AZT homodinucleotide AZTp2AZT designed for HIV-1 infection. Antimicrobial Erythrocytes

Activity

of

AZTp2EMB-loaded

Disseminated infection with Mycobacterium avium complex remains the most common and serious bacterial infection in patients with advanced AIDS. With the advent of highly active antiretroviral therapy (HAART, the therapy usually includes one nucleoside analog, one protease inhibitor and either a second nucleoside analog or a nonnucleoside reverse transcription inhibitor) and restoration of CD4+T counts, a decrease in the incidence of opportunistic infections in AIDS patients was observed [32, 33]. However, HAART is not universally available, does not eliminate lymph node MAC [34, 35] and moreover, newly diagnosed AIDS patients, those for whom HAART has been ineffective or those who are unable to afford AIDS, remain at risk for opportunistic infections. Thus, M. avium is still a significant pathogen. Hence, therapeutic strategies able to inhibit both mycobacterial and retroviral infection are needed. The nucleoside analog AZT and the antimicrobial agent EMB are among the drugs of choice against HIV-1 and MAC infection, respectively. Unfortunately, because of the rapid plasma elimination and toxicity of these drugs, daily multiple-drug therapies must often be continued throughout life, frequently causing major side effects and, as a consequence, poor patient compliance. Therefore, alternative strategies which reduce the toxicity of the drugs and allow prolonged application intervals are sorely needed. The administration of a single molecule featuring activity against both of the two infections (HIV-1 and MAC) could be advantageous, specially if this treatment was able to improve the pharmacokinetics and decrease toxic side effects of these

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

45

drugs. Prompted by these considerations, we developed a heterodimer consisting of AZT and EMB interconnected by a pyrophosphate bridge (AZTp2EMB) [10]. Unfortunately, this heterodimer cannot be directly administered as a free drug because it is unable to cross the cell membrane and, in addition, is quickly degraded in plasma (t1/2 approx. 1h) with a corresponding immediate production of AZT. This limitation could, however, be overcome by administering AZTp2EMB encapsulated in RBC. Human erythrocytes possess a dinucleotide pyrophosphatase able to cleave the pyrophosphate bridge of AZTp2EMB with subsequent production of AZT-MP and EMB-MP which are then converted to AZT and EMB. The intraerythrocytic metabolism of AZTp2EMB for the production of the diffusible monomers was evaluated over a long period of time. The heterodimer resulted stable enough in RBC to allow their use for the proposed objectives Fig. (5). As an evidence of this, the antimycobacterial activity of AZTp2EMB-loaded RBC was evaluated by administering the heterodimer-loaded erythrocytes to human macrophages infected with a strain of Mycobacterium avium. Our results demonstrate that AZTp2EMB-loaded RBC were able to inhibit the growth of bacilli in infected macrophages in a time-dependent manner. The administration of loaded RBC to infected macrophages was able, in fact, to inhibit the intracellular replication of M. avium by about 22% in one day and by more than 90% in 6 days (Table 2). The efficacy of AZTp2EMB-loaded RBC against HIV-1 replication was not evaluated because the evidence of AZT production (4 mM see Fig. (5)), should be sufficient to assure its antiretroviral activity (0.05-0.5 µM) [36]. In conclusion, these data prove that erythrocytes loaded with AZTp2EMB act as bioreactors for the slow delivery of the antiviral drug AZT and the antimicrobial drug ethambutol. Table 2.

Activity of Mycobacterium Macrophages

AZTp2EMB-loaded avium Replication

RBC on in Human

Days of incubation

CFU/ml (%)

1

77.8

2

22.2

3

16

6

6.7

Macrophages were infected with M. avium and cultured for six days. Loaded RBC were incubated with infected macrophages for different periods: 1, 2, 3 and 6 days. The data represent the percentage of total viable counts (CFU, colony forming units) on supernatants with respect to infected untreated macrophages and are the mean of two experiments, each performed in duplicate.

Kinetic of AZT Erythrocytes

Release

from

AZTp2AZT-loaded

Homodinucleotide AZTp2AZT seems to have chemical and biochemical properties enabling its profitable utilization in the erythrocyte-encapsulated form. Once the homodimer was encapsulated in RBC by a loading procedure [37], erythrocyte enzymes were shown to metabolize AZTp2AZT to yield AZT that is eventually released by simple diffusion. This could be useful since sustained release of AZT in

46

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Fig. (5). Metabolism of AZTp2EMB in intact erythrocytes. AZTp2EMB was encapsulated in human erythrocytes by a procedure of hypotonic dialysis and isotonic resealing to a final concentration of 6.5 mM. These cells were then incubated at 37°C and at different times of incubation aliquots were processed to determine the concentration of AZTp 2EBM and its metabolites. The results shown are from one experiment representative of three that agreed within 10% of the reported values. Reprinted with permission from: Rossi L., et al. AIDS Res. Hum. Retrovir., 1999, 15, 345. © 1999 Mary Ann Liebert, Inc.

circulation is advocated to avoid some toxic effects, mostly myelosuppression, which are induced by conventional (oral or intravenous) administration of this molecule. In order to understand the kinetic of the disappearance of AZTp2AZT in the presence of erythrocyte enzymes, studies on its stability in erythrocyte lysate were conducted. These studies showed that AZTp2AZT has a considerable stability in the erythrocyte lysates, particularly in the presence of 1 mM ATP. Thus, after two hours at 37°C, 90% of AZTp2AZT was still present while without addition of ATP, AZTp2AZT decayed down to 62% at two hours. Once made sure that the AZT dimer could be metabolized by erythrocyte enzymes, we evaluated its intraerythrocytic metabolism over long incubation times. As shown in Fig. (6), a slow and linear decrease of intraerythrocytic AZTp2AZT was observed in these conditions, down to 12% of the starting levels at 140 hours of incubation. The production of AZT-MP within red cells started almost immediately, peaking at 50 hours. The output of AZT from red cells began at 25 hours and progressed linearly over the incubation time. The major pathway of such bioconversion is the symmetrical degradation to two AZT-MP molecules that are then dephosphorylated by a 5’-nucleotidase [38]. Hydrolysis of the pyrophosphate bond of AZTp2AZT is most probably catalyzed by a dinucleotide pyrophosphatase previously identified and characterized in human erythrocytes [39], as suggested by inhibition of AZTp2AZT degradation by ATP. However, since human erythrocytes are unable to phosphorylate AZT-MP to AZT-TP [38], the appearance of

Magnani et al.

Fig. (6). AZTp2AZT metabolism in intact erythrocytes. Human red blood cells were loaded with AZTp2AZT by a procedure of hypotonic dialysis and isotonic resealing to a final concentration of 0.15 mM. At the time intervals indicated, AZTp2AZT and its metabolites were determined. AZT and low amounts of AZT-MP (5-10 % of the total) were found in the incubation medium, while all other metabolites were present in the erythrocyte extracts. Results of a representative experiment are shown, out of three different ones in which variability never exceeded 12 %. Reprinted with permission from: Benatti U., et al. Biochem. Biophys. Res. Commun., 1996, 220, 20. © 1996 Academic Press., Inc.

the latter metabolite should due to an asymmetrical hydrolysis of AZTp2AZT that becomes evident upon prolonged incubation and that can produce AZT and AZTDP (the latter compound being then phosphorylated to AZTTP by a kinase activity). As observed incubating AZT-DP in erythrocyte lysate, AZT-DP decayed progressively while AZT-TP was formed but only in the presence of ATP, the phosphate groups donor. AZT-DP and AZT-TP might then represent storage metabolites for delayed formation and sustained delivery of AZT. Moreover, permeation of AZTMP across the erythrocyte membrane has already been observed in AZT-MP-loaded human red cells and human plasma is able to dephosphorylate AZT-MP, thus contributing to the total AZT pool present in plasma. The potential of the AZTp2AZT-loaded erythrocytes to behave as bioreactors suited to the delivery of AZT in circulation was explored by upgrading the starting intraerythrocytic AZTp2AZT levels and following the release of AZT in the surrounding plasma (Fig. (7)). Indeed, the loaded red cells showed a high efficiency of AZT formation and export, but a significant release of AZT-MP was also observed. Therefore, a single batch of AZTp2AZT-loaded erythrocytes might be used as an AZT-releasing bioreactor. In fact, the efficiency of the bioreactor is such that less than two ml of AZTp2AZT-overloaded erythrocytes would be sufficient to produce and release within 24 hours the same

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

47

Fig. (7). Patterns of AZT release in plasma from AZTp2AZT-overloaded human erythrocytes. The methodology of encapsulation [8] was upgraded to result in starting concentrations of intraerythrocytic AZTp2AZT as high as 3.9 µmoles/ml packed red cells. Incubation was at 37°C in autologous plasma at a 10% hematocrit. The stably low levels of AZTp2AZT in plasma are apparently due to microhemolysis of the loaded red cells, while the concentrations of AZT-MP and of AZT reflect output from the cells. Data are the average values ± standard deviation for three independent experiments.

total amount of plasma AZT detected after 1 hour in subjects receiving a bolus of 100 mg of AZT [40]. NUCLEOTIDE DIMER-LOADED ERYTHROCYTES AS DRUG TARGETING SYSTEM In order to be active as reverse transcriptase inhibitors, most antiviral nucleoside analogs must first be phosphorylated to the corresponding 5’-triphosphate derivatives by cellular kinases. Unfortunately, the activity levels of these enzymes depend on the cell type and on the cell activation state. Quiescence cells usually have low levels of these enzymes responsible for nucleoside analog phosphorylation, whereas activation results in increased activity levels [41-45]. Thus, macrophages (important in vivo reservoirs for different kind of viruses, including HIV-1 and HSV-1) phosphorylate several antiviral nucleoside analogs at a rate lower than that of peripheral blood mononuclear cells [41, 42]. For this and other reasons, over the last few years we have explored systems that are able to selectively target phosphorylated nucleoside analogs to macrophages. In particular, we have developed a drug targeting system [46] that allows the direct administration of nucleoside analogs in phosphorylated form to macrophages. Toward this end, we developed a loading procedure able to encapsulate in erythrocytes non diffusible prodrugs followed by a procedure which mimic red cells senescence, thus allowing erythrocyte recognition and phagocytosis by macrophages as usually occurs for the removal of senescent cells from circulation. It is known that erythrocytes subjected to the procedure of encapsulation by hypotonic dialysis and isotonic resealing show a normal survival in circulation, however, this survival can be shortened by a number of

treatments [47] promoting the sequestration of the drugloaded erythrocytes in the reticuloendothelial system and thus allowing the delivery of the encapsulated drug to the phagocytic cells. Furthermore, we reasoned that the delivery of drugs already in a phosphorylated form would overcome the reduced ability of macrophages to phosphorylate several antiviral nucleoside analogs. However, although an antiviral drug designed to protect macrophages ideally should be administered in a phosphorylated form, it should not be administered as a nucleoside 5’-triphosphate but rather as a close precursor of it to avoid toxicity. For these considerations, we used the described homo- and heterodinucleotides (AZTp2AZT, AZTpPMPA, AZTp2ACV, ACVpPMPA, Bis-PMEA) to be administered specifically by this system. In addition, by the administration of a doubledrug consisting of compounds with different antiviral activity, the advantage of a combination therapy could be obtained. Herein, the antiviral activity of nucleotide dimers selectively delivered to macrophages by RBC is described. Targeting of AZTp2AZT-loaded Eryhtocytes to Infected Macrophages: In Vitro and In Vivo Studies The homodinucleotide AZTp2AZT was evaluated as prodrug to confer protection to macrophages against “de novo” retroviral infection [8]. AZTp2AZT was encapsulated into erythrocytes where it was remarkably stable within the time of phagocytosis (erythrophagocytosis is allowed for approx. 20 h). Induction of erythrocyte membrane protein clusterization and phagocytosis allowed the targeted delivery of this drug to macrophages, where it progressively decayed until becoming undetectable after three days, probably converting itself to the active AZT-5’-triphosphate. Indeed,

48

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

addition of AZTp2AZT-loaded red blood cells to human macrophages inhibits HIV-1 proviral DNA formation by 9397% and p24 production by more than 80%. AZT at 0.1 µM in the culture medium for 20 h had no effect, whereas 0.1 µM maintained in the culture medium for the duration of the experiment inhibits p24 production by 58%. Similar results were obtained with feline macrophages infected with FIV (feline immunodeficiency virus). Protection of murine macrophages from LP-BM5 infectivity was somewhat less efficient (71 ± 1.2 % inhibition of proviral DNA formation). AZTp2AZT-loaded erythrocytes were also in vivo evaluated in a murine retrovirus-induced immunodeficiency model of AIDS (MAIDS), both alone or in combination with oral AZT [48]. C57BL/6 mice were infected with the LPBM5 retroviral complex and treated for 3 months with weekly administrations of 15 nmol AZTp2AZT encapsulated in autologous erythrocytes. AZTp2AZT treatment was found to reduce lymphoadenopathy (48%), splenomegaly (26%) and BM5d proviral DNA content in lymph nodes, spleen and brain by 37, 40 and 36% with respect to the values found for untreated animals. The combined treatment with AZT in drinking water (0.25 mg/ml) and AZTp 2AZT encapsulated in erythrocytes did not provide additive responses in several of the parameters investigated. However, it was found to be much more effective than single treatments in reducing the proviral DNA content in brain (67%). Furthermore, no apparent signs of hematoxicity were observed. Thus, macrophage delivery of antiviral drugs may contribute to brain protection from retroviral infections by mechanisms other than those exerted by the oral AZT administration. Targeting of AZTpPMPA-loaded Erythrocytes to HIV-1Infected Macrophages To the same extent of inhibiting HIV-1 production in human macrophages, we studied the simultaneously administration of AZT and PMPA, trying to overcome both the low ability of macrophages to phosphorylate AZT and the low cellular permeability of PMPA. In addition, since it was reported that AZT demonstrated strong synergistic antiHIV-1 activity in combination with PMPA [24], we thought to target selectively to macrophages AZTpPMPA featuring anti-HIV-1 activity after intracellular cleavage into AZT-MP and PMPA [11]. When AZTpPMPA was encapsulated in RBC, 50% of the compound was still present inside cells after 36 h of incubation (erythrophagocytosis is allowed for 18 h) suggesting that the heterodimer is stable enough in RBC to allow their use as a drug delivery system. Once delivered to macrophage, AZTpPMPA degradation occurs with conversion into active metabolites. The possible mechanism involves first the cleavage of AZTpPMPA into AZT-MP and PMPA, which are then phosphorylated to AZT-TP and PMPA-DP, respectively by cellular kinases. Protection of macrophages against “de novo” HIV-1 infection is almost complete (97%) upon the administration of erythrocytes loaded with the highest AZTpPMPA concentration tested (0.25 mM). When AZTpPMPA, AZT and PMPA were added as free drugs for the same time as in RBC (18 h), only a low inhibition in p24 production was obtained. In conclusion, when selectively delivered to macrophages, AZTpPMPA is very effective in inhibiting HIV-1 production

Magnani et al.

to almost undetectable levels, suggesting that these cells can be efficiently protected if proper drug and proper drug delivery system are used. Targeting of AZTp2ACV-loaded Erythrocytes to HIV-1 or HSV-1-Infected Macrophages HSV-1 infection is common among individuals infected with HIV-1. Among immunocompetent patients, herpesvirus infection is in most cases self-limited and usually does not require systemic antiviral therapy. In contrast, in immunocompromised hosts, HSV-1 infections are much more severe, frequent and often life-threatening. In addition, HSV-1 is able to activate and increase HIV-1 replication [49, 50], thereby acceleration the progression of the disease. Furthermore, HIV-1 and HSV-1 are able to mutually activate their replication during coinfection of monocyte/derived macrophages [51]. Hence therapeutic strategies able to inhibit replication of both viruses in macrophages are needed. Until now, the most common therapies against HIV1 and HSV-1 infectivity have been based on the administration of nucleoside analogs. The nucleoside analogs zidovudine (AZT) and acyclovir (ACV) are among the drugs of choice against HIV-1 and HSV-1 infection, respectively. However, to be active, these antiviral drugs must be converted to their phosphorylated derivatives by viral and/or cellular kinases, as previously discussed. Furthermore, the administration of ACV in its phosphorylated form as well could be more advantageous than conventional therapy because the drug would be active before the expression of HSV-1 timidine kinase (TK) activity. Prompted by these considerations, the heterodinucletide AZTp2ACV consisting of both an anti-HIV-1 and an antiherpetic drug, bound by a pyrophosphate bridge, was developed [9]. The heterodimer can not be administered as a free drug because it is not able to permeate the cellular membranes and, in addition, is degraded in plasma where it completely disappeared after 2 h of incubation, with a corresponding immediate production of its phosphorylated metabolites AZT-MP and ACV-MP, which were rapidly dephosphorylated to AZT and ACV. This limitation can be overcome by administering AZTp2ACV encapsulated into RBC to infected macrophages. As concern HIV-1 infection, two different concentration of AZTp2ACV inside RBC were evaluated (1.12 and 3.18 mM) and two different RBC-tomacrophage ratios (100:1 and 500:1) were also tested. The best results (80% inhibition of viral replication) were obtained when macrophages were treated with erythrocytes loaded with highest AZTp2ACV concentration (3.18 mM) and at the highest RBC-to-macrophage ratio (500:1) (Fig. (8)). However, a decreased p24 production was also observed with the administration of 1.12 mM AZTp2ACVloaded RBC (60% inhibition). It is worth noting that the administration of unloaded RBC was able to inhibit the replication of HIV-1 by about 17 and 34% at the two RBCto-macrophage ratios tested, respectively (100:1 and 500:1). As an additional control, macrophages were exposed to free 1.0 µM AZTp2ACV for the same time as loaded RBC or unloaded RBC administration. The overnight exposure to the free heterodimer had little effect in inhibiting HIV-1 replication (confirming the importance of administration of AZTp2ACV by RBC).

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

49

Fig. (8). Virus production by HIV-infected macrophages treated with AZTp2ACV-loaded RBC or free drug. Erythrocytes loaded or free drugs were added to human macrophages 18 h before infection and maintained overnight until infection. All values are the means of quadruplicate cultures and show the p 24 synthesized at the peak of viral replication after 3 weeks of infection. One hundred percent virus production corresponds to 28,000 pg of p24/ml.

AZTp2ACV-loaded RBC, at two different drug concentrations (2.8 and 5 mM) and unloaded RBC were tested also for their anti-HSV-1 activity against infected macrophages. As control, 0.1 µM ACV added for the same time as RBC was used. AZTp2ACV-loaded RBC were able to inhibit HSV-1 replication efficiently (by more than 90%) after 48 h of infection (Fig. (9)) whereas ACV treatment results in about 65% inhibition of viral production. As expected, unloaded RBC were partially effective in inhibiting HSV-1 replication in human macrophages (20% untreated cells). These results were obtained at a RBC-tomacrophage ratio of 50:1. Higher ratio (100:1, 150:1) provided similar results. Furthermore, no cytopatic effect

was observed in macrophages receiving AZTp2ACV-loaded RBC, while ACV treatment, although able to prevent the cytopatic effect, did not succeed in inhibiting the formation of cellular aggregates. Probably, macrophages receiving ACV do not produce infected viral particles but produce early- and some intermediate-phase viral proteins that induce cellular fusions (data not shown). In conclusion, our data prove that AZTp2ACV once encapsulated in autologous erythrocytes modified to increase their recognition and phagocytosis is able to protect macrophages from the “de novo” infection by HIV-1 and HSV-1.

Fig. (9). Inhibition of HSV-1 replication by AZTp2ACV in human macrophages. Monocyte-derived macrophages were cultured for 10 days and then treated for 18 h with both unloaded (UL) and RBC loaded with two different AZTp 2ACV concentrations (5.0 and 2.8 mM, respectively) at a ratio of 50 RBC per macrophage. Non ingested RBC were then removed and macrophages were infected for 2 h at 3 pfu/cell. The cells were extensively washed and then cultured for 2 days. Acylovir (0.1 µM) was added for the same time as RBC. Virus production was assayed in Vero cells, using the plaque assay.

50

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Targeting of ACVpPMPA-loaded Erythrocytes to HIV-1 or HSV-1-Infected Macrophages To the same extent as AZTp2ACV and with the aim to improve the anti-HIV-1 activity, a new heterodimer consisting of ACV and PMPA was synthesized [12]. To evaluate the anti-HIV-1 activity of ACVpPMPAloaded RBC, human macrophages were treated with ACVpPMPA (approx. 1.0 mM) inside RBC or unloaded RBC. As control, the free drugs 1.0 µM ACVpPMPA and 1.0 µM PMPA were used. The results shown in Fig. (10), prove that ACVpPMPA-loaded RBC were able to completely protect macrophages against HIV-1 infection; this protection persisted up to the end of the experiment (34 days). To evaluate the anti HSV-1 activity of ACVpPMPA-

Magnani et al.

loaded RBC, human macrophages were treated with 1.0 mM ACVpPMPA-loaded RBC. As control, 15 µM ACVpPMPA added for the same time as RBC was used. The results obtained show that ACVpPMPA-loaded RBC were able to inhibit HSV-1 replication by almost 70%, compared to < 50 % inhibition induced by unloaded RBC, while the addition of the free drug in the medium reduced HSV-1 production by 36% (Fig. (11)). In all, the results show that the antiviral effect of the dimer is substantially different in HIV-1 infected macrophages, as compared to those infected by HSV-1. These apparently surprising results can be explained by the different virus lifecycles. Indeed, HIV-1-RT works only in one step at the beginning of the virus replication cycle. After proviral integration, RT is no longer necessary for virus

Fig. (10). Percent virus inhibition (p24) in HIV-infected macrophages treated with ACVpPMPA-loaded RBC or free drugs. Erythrocytes loaded or free drugs were added to human macrophages 18 h before infection and maintained overnight until infection. All values are means of quadruplicate cultures. One-hundred-percent virus production corresponds to 22,000 pg of p24/ml 20 days after infection.

Fig. (11). Inhibition percentage of HSV-1 replication by ACVpPMPA-loaded RBC or free drugs. Human macrophages were cultured for 10 days before treatment with the drugs. Cells were infected for 2 h with 3 pfu/cell of HSV-1. The inhibitory effect of the compound was evaluated 48 h after infection by a plaque assay in Vero cells. The values are the mean of three experiments, each performed in duplicate. The titer of HSV-1 production in the supernatants of infected untreated macrophages ranged from 0.5 to 1.0 x 106 pfu/ml.

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

production. On the contrary, HSV-1-DNA polymerase works during the whole HSV-1 lifecycle, since production of new viral DNA genome is continuously required. Thus, it is conceivable that the effect of the dimer on HIV-1-RT can be amplified by the limited rate of activity of this enzyme in infected macrophages [52], while, in contrast, the effect of the same dimer can be reduced by the continuous activity of the HSV-1-DNA polymerase. In conclusion, ACVpPMPA has protective activity in human macrophages against HIV-1 and HSV-1 infections. The dinucleotide dimer maintains the high antiretroviral activity of (R)-PMPA and at the same time compensates for the lack of anti-HSV-1 activity of nucleotide. Targeting of Bis-PMEA-loaded Erythrocytes to HIV-1 or HSV-1 Infected Macrophages Since PMEA, differently from PMPA, shows an antiviral activity both against retroviruses and herpesviruses, it is possible to protect macrophages against “de novo” HIV-1 and HSV-1 infections by the administration of a unique molecule as prodrug consisting of two molecules of PMEA bound together (Bis-PMEA) [53]. In fact, Bis-PMEA could be more advantageous than PMEA since, mole per mole, it has a double efficacy and could act as a prodrug for a more sustained delivery. When Bis-PMEA was encapsulated in RBC, 50% of the compound was still present inside cells after 5 days of incubation, suggesting that the homodimer is stable enough in RBC to allow their use as a drug targeting system (Fig. (12)). Moreover, by administering [14C] Bis-PMEA-loaded

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

51

RBC to macrophages, 47% of Bis-PMEA and 28% of PMEA were still present ten days after phagocytosis; differently, only 12% of PMEA was found in macrophages receiving [14C] PMEA-loaded RBC. These results strongly suggest the advantage of the dimer form as prodrug. Thus, once delivered to macrophages, a slow Bis-PMEA conversion to PMEA occurs, yielding its pharmacologically active metabolite, as demonstrated by the reported antiviral activities. The hypothesized mechanism is as follows: BisPMEA is first cleaved into two PMEA monomers by a specific pyrophosphatase then phosphorylated to PMEA diphosphate (PMEApp), the active antiviral drug. The protection of macrophages obtained against HIV-1 infection is remarkable (95%) upon drug-loaded RBC phagocytosis, while 45% inhibition was reached after the administration of 1.0 µM PMEA, a concentration that is ≥ 40-fold the EC50 of PMEA in HIV-1 infected macrophages [54]. These results were obtained 10 days post infection and were confirmed even 35 days after virus challenge. When the antiviral activity of Bis-PMEA-loaded RBC against HSV-1 replication was evaluated in macrophages, a good protection (85%) from the infection was obtained. As expected, the administration of Bis-PMEA as a free drug for the same time as in RBC was effective too (due to PMEA formed by hydrolysis of Bis-PMEA by means of the serum enzymes present in the RPMI complete culture medium), giving 70% inhibition of HSV-1 replication. However, it is noteworthy that this result was observed only during the first 48 h post infection, after that time the anti-herpes efficacy of free Bis-PMEA rapidly decayed, while that of the BisPMEA-loaded RBC persisted for at least 6 days after infection (Fig. (13)). Furthermore, the anti-herpes efficacy of a single administration of drug-loaded RBC before infection was comparable to that of Bis-PMEA maintained throughout the entire experiment. Bis-PMEA can be considered a useful prodrug of PMEA that, once inside macrophages, can be slowly converted into PMEA and protect these cells from “de novo” HIV-1 and HSV-1 infections for a longer period of time. CONCLUSION

Fig. (12). Metabolism of Bis-PMEA in intact erythrocytes. Bis-PMEA was encapsulated into human erythrocytes by a procedure of hypotonic dialysis and isotonic resealing to a final concentration of about 0.5 mM. These cells were then incubated at 37°C for 10 days. Values are the mean ± s.e. of three different experiments. Reprinted with permission from: Rossi L., et al. J. Antimicrob. Chemother., 2001, 47, 819. © 2001 The British Society for Antimicrobial Chemotherapy.

The results summarized in this review suggest that the administration of nucleotide dimers to combat viral infections could be advantageous when compared to nucleoside and nucleotide analogs such as AZT, ACV, PMEA, PMPA, etc. Indeed, favorable pharmacokinetics decreased toxic side effects and increased antiviral activities when using homo- and heterodinucleotides have been observed, suggesting a possible alternative to the use of combination therapy based on single molecules. Furthermore, in the presence of coinfections (as usually occurs in immunocompromised patients), in order to favor compliance, single drugs with multiple activities should be preferred to multiple drugs active against each pathogen. We have also shown that red blood cells loaded with homo- or heterodimers of nucleoside analogs can perform as circulating bioreactor releasing the corresponding monomers in circulation improving their pharmacokinetic patterns and therapeutic performances. In fact, the main advantage of this delivery system is the possibility of maintaining a continuous

52

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1

Magnani et al.

similar aims. The field of application of these new dimers is different from the ones discussed in this paper but support the concept that dimers can represent a new class of drugs with improved pharmacokinetical properties over the parent compounds. This dimer approach will certainly be further developed [58-61]. In conclusion, data discussed herein suggest that new drugs (homo- and heterodinucleotides) and new drug delivery systems (RBC) could offer an additional improvement in the use of conventional and new antiviral double-drugs. ACKNOWLEDGEMENTS This work is partially supported by FIRB founds (PNR 2001-2003, Red blood cells as drug carriers, RBNE01TBTR) and by Ministero della Sanità, Istituto Superiore della Sanità, Progetto AIDS Scientific Collaboration Agreement 31D.65 and 40D.42. Fig. (13). Percentage of HSV-1 inhibition by Bis-PMEA-loaded RBC or free drug administration to infected macrophages. Bis-PMEA-loaded RBC were added for 18 h before virus challenge and their activity compared with that of Bis-PMEA added as for RBC or added during infection and maintained throughout the entire experiment. At different times, supernatants were collected and infectivity titres determined using the plaque assay in Vero cells. One of three similar and independent experiment is shown. The mean titre of virus production in the supernatants of infecteduntreated culture was 2.15 x 106 ± 3.22 x 105 pfu/ ml (mean ± s.e.). Reprinted with permission from: Rossi L., et al. J. Antimicrob. Chemother., 2001, 47, 819. © 2001 The British Society for Antimicrobial Chemotherapy.

drug supply in the bloodstream for several days with a single administration of processed erythrocytes. Furthermore, this method of drug-delivery is also expected to reduce toxicity by avoiding the peak drug concentration usually occurring after intravenous or oral drug intake. It should be emphasized that a new equipment (which we named “Red Cell Loader”) for the encapsulation of non-diffusible drugs in human red blood cells was designed and built [55]. Studies of drug delivery in vivo have already been performed [30] while others are in progress. The same apparatus will also be useful for the administration of drugs that should be targeted to macrophages, which are known “reservoirs” of different kinds of viruses. Indeed, the results herein summarized show that homo- and heterodinucleotide-loaded erythrocytes selectively delivered to macrophages are able to protect these cells from HIV-1 and HSV-1 “de novo” infection. In HIV-1 infection, this delivery could be used efficiently in combination with drugs having a recognized lymphocyte-protecting activity (nucleoside analogs, protease inhibitors, etc.). Interestingly, the same drug targeting system could be used for the delivery of other heterodimers, possibly including both a reverse transcriptase inhibitor and a protease inhibitor (such heterodimers have already been synthesized and their antiviral activity as free drugs investigated [56, 57]). If this is feasible, we will be able to protect both non-infected macrophages from new infection and infected macrophages from viral replication. Recently, other homo- and heterodimers have been synthesized with

LIST OF ABBREVIATIONS ACV

=

Acyclovir

ACV-MP

=

Acyclovir monophosphate

ACVpPMPA

=

(P1-(6H-purin-6-one,2-amino-1,9dihydro-9-[(2(phosphonooxy)ethoxy]methyl]-P2-(9(R)-2(phosphonomethoxypropyl)adenine

AIDS

=

Acquired immunodeficiency syndrome

ANP

=

Acyclic nucleoside phosphonate

AZT

=

Azidothymidine

AZT-MP

=

Azidothymidine monophosphate

AZTpddA

=

[P1-(thymine-3’-azido-2’,3’-dideoxyβ-D-riboside-5’-P2-adenine-2’,3’dideoxy-β-D-riboside)phosphate]

AZTpddT

=

[P1-(thymine-3’-azido-2’,3’-dideoxyβ-D-riboside-5’-P2-thymine-2’,3’dideoxy-β-D-riboside)phosphate]

AZTp2ACV

=

[P1-(thymine-3’-azido-2’,3’-dideoxyβ-D-riboside-5’-P2-guanine-9-(2hydroxy-ethoxymethyl) pyrophosphate]

AZTp2AZT

=

P1,P2-bis[thymine-3’-azido-2’,3’dideoxy-D-riboside-5’]pyrophosphate

AZTp2EMB

=

[P1-thymine-3’-azido-2’,3’-dideoxyD-riboside-5’-P2(+)2,2’(ethylendiimino)di-1-butanol pyrophosphate]

AZTpPMPA

=

[P1-(thymine-3’-azido-2’,3’-dideoxyβ-D-riboside-5’-phosphate)-P2-(9-(R)2-(phosphonomethoxypropyl)adenine]

Bis-PMEA

=

P1,P2-bis[2-(adenin-9Hyl)ethoxymethyl]phosphonate

ddC

=

Dideoxycytidine

Targeting Nucleotide Dimers Containing Antiviral Nucleosides

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1 [13]

EMB

=

Etambuthol

EMB-MP

=

Etambuthol monophosphate

FIV

=

Feline immunodeficiency virus

HAART

=

Highly active antiretroviral therapy

HIV-1

=

Human immunodeficiency virus type 1

HSV-1

=

Herpes simplex virus type 1

[16] [17]

MAC

=

Mycobacterium avium complex

[18]

MAIDS

=

Murine acquired immunodeficiency syndrome

[19]

[14] [15]

N4-hxddC-AZT =

N4-hexadecyl-2’-deoxyribocytidylyl(3’à5’)-3’azido-2’,3’-deoxytimidine

[20]

N4-hxddC-ddC

N4-hexadecyl-2’-deoxyribocytidylyl(3’à5’)-2’,3’-dideoxycitidine

[21]

=

4

4

N -pamdC-AZT =

N -palmitoyl-2’-deoxyribocytidylyl(3’à5’)-3’azido-2’,3’-deoxytimidine

N4-pamdC-ddC =

N4-palmitoyl-2’-deoxyribocytidylyl(3’à5’)-2’,3’-dideoxycitidine

PMEA

=

9-[2-(phosphonomethoxy) ethyl]adenine

[22] [23] [24] [25] [26] [27] [28]

PMPA

=

9-[(R)-2-(phosphonomethoxy) propyl]adenine

[29]

RBC

=

Red blood cell

[30]

RT

=

Reverse transcriptase

UL RBC

=

Unloaded red blood cell

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9]

[10] [11] [12]

Velazquez, S.; Alvarez, R., San-Felix, A.; Jimeno, M.L.; De Clercq, E.; Balzarini, J., Camarasa, M.J. J Med. Chem., 1995, 38, 1641. Busso, M.; Mian, A.M.; Hahn, E.F.; Resnick, L. AIDS Res. Hum. Retrovir., 1988, 4, 449. Puech, F.; Gosselin, G.; Balzarini, J.; Good, S.S.; Rideout, J.L.; De Clercq, E.; Imbach, J.L. Antiviral Res., 1990, 14, 11. Schinazi, R.F.; Sommadossi, J.P.; Saalmann, V.; Cannon, D.L.; Xie, M.Y.; Hart, G.C.; Smith, G.A.; Hahn, E.F. Antimicrob. Agents Chemother., 1990, 34, 1061. Peghini, P.A.; Zahner, R.; Kuster, H.; Schott, H.; Schwendener, R.A. Antiviral Chem. Chemother., 1998, 9, 117. Schwendener, R.A.; Peghini, P.A.; Ludwing, P.S.; Schott, H. Nucleosides & Nucleotides, 1999, 18, 949. Renoud-Grappin, M.; Fossey, C.; Fontane, G.; Ladurée, D.; Aubertin, A.M.; Kirn, A. Antiviral Chem. Chemother., 1998, 9, 205. Magnani, M.; Casabianca, A.; Fraternale, A.; Brandi, G.; Gessani, S.; Williams, R.; Giovine, M.; Damonte, G.; De Flora, A.; Benatti, U. Proc. Natl. Acad. Sci. USA, 1996, 93, 4403. Rossi, L.; Brandi, G.; Schiavano, G.F.; Balestra, E.; Millo, E.; Scarfì, S.; Damonte, G.; Gasparini, A.; Magnani, M.; Perno, C.F.; Benatti, U.; De Flora, A. AIDS Res. Hum. Retroviruses, 1998, 14, 435. Rossi, L.; Brandi, G.; Schiavano, G.F.; Scarfì, S.; Millo, E.; Damonte, G.; Benatti, U.; De Flora, A.; Magnani, M. AIDS Res. Hum. Retroviruses, 1999, 15, 345. Franchetti, P.; Rossi, L.; Cappellacci, L.; Pasqualini, M.; Grifantini, M.; Balestra, E.; Forbici, F.; Perno, C.F.; Serafini, S.; Magnani, M. Antiviral Chem. Chemother., 2001, 12, 151. Franchetti, P.; Abu Sheikha, G.; Cappellacci, L.; Marchetti, S.; Grifantini, M.; Balestra, E.; Perno, C.; Benatti, U.; Brandi, G.; Rossi, L.; Magnani, M. Antiviral Res., 2000, 47, 149.

[31] [32] [33] [34] [35] [36]

[37] [38] [39] [40] [41] [42] [43] [44] [45] [46]

53

Rossi, L.; Dominici, S.; Serafini, S.; Casabianca, A.; Cerasi, A.; Chiarantini, L.; Celeste, A.G.; Cappellacci, L.; Franchetti, P.; Grifantini, M.; Magnani, M. J. Antimicrob. Chemother., 2002, 50, 365. De Clercq, E. Biochem. Pharmacol., 1991, 42, 963. Barditch-Crovo, P.; Toole, J.; Hendrix, C.W.; Cundy, K.C.; Ebeling, D.; Jaffe, H.S.; Lietman, P.S. J. Infect. Dis., 1997, 176, 406. Jolicoeur, P. FASEB J., 1991, 5, 2398. Hartley, J.W.; Frederickson, T.N.; Yetter, R.A.; Makino, M.; Morse, H.C. III. J. Virol., 1989, 63, 1223. Morse, H.C. III; Chattopadhyay, S.K.; Makino, M.; Frederickson, T.N.; Hügin, A.W.; Hartley, J.W. AIDS, 1992, 6, 607. Tanji, N.; Tanji, K.; Kambham, N.; Markowitz, G.S.; Bell A.; D’Agati, V.D. Human Pathol., 2001, 32, 734. Fisher, E.J.; Chaloner, K.; Cohn, D.L.; Grant, L.B.; Alston, B.; Brosgart, C.L.; Schmetter, B.; El-Sadr, W.M.; Sampson, J. AIDS, 2001, 15, 1695. Ho, E.S.; Lin, D.C.; Mendel, D.B.; Cihlar, T. J. Am. Soc. Nephrol. , 2000, 11, 383. No authors listed. FDA Consum., 2002, 36, 5. Robbins, B.L.; Fridland, A. Int. Antiviral News, 1996, 4, 57. Mulato, A.S.; Cherrington, J.M. Antiviral Res., 1997, 36, 91. Rossi, L.; Serafini, S.; Franchetti, P.; Casabianca, A.; Orlandi, C.; Schiavano, G.F.; Carnevali, A.; Magnani, M. J. Antimicrob. Chemother., 2002, 50, 639. De Loach, J.R. J. Appl. Biochem., 1983, 5, 149. De Loach, J.R.; Sprandel, U. Red Blood Cells as Carriers For Drugs; De Loach & Sprandel, Eds.; Karger: Basel, 1985. De Flora, A.; Benatti, U.; Guida, L.; Zocchi, E. Proc. Natl. Acad. Sci. USA, 1986, 86, 7029. Ropars, C.; Chassaigne, M.; Nicolau, C. Red Blood Cells as Carriers for Drugs. Potential Therapeutic Applications; Ropars C., Chassaigne M. & Nicolau, C., Eds.; Pergamon Press: Oxford, 1987. Rossi, L.; Serafini, S.; Cenerini, L.; Picardi, F.; Bigi, L.; Panzani, I.; Magnani, M. Biotechnol. Appl. Biochem., 2001, 33, 85. Rossi, L.; Serafini, S.; Magnani, M. In Erythrocyte Engineering for Drug Delivery and Targeting; Magnani, Ed.; Landes Bioscence: New York, 2003; pp.1-18. Kovacs, J.A.; Masur, H. N. Engl. J. Med., 2000, 342, 1416. Pallela, F.J.; Delancy, K.M.; Moorman A.C.; Loveless, M.O.; Fuhrer, J.; Satten, G.A.; Aschman, D.J.; Holmberg, S.D. N. Engl. J. Med., 1998, 338, 853. Race, E.M.; Adelson-Mitty, J.; Kriegel, G.R.; Barlam, T.F.; Reimann, K.A.; Letvin, N.L.; Japour, A.J. Lancet, 1998, 351, 252. Phillips, P.; Kwiatkowski, M.B.; Copland, M.; Craib, K.; Montaner, J. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol., 1999, 20, 122. Furman, P.A.; Fyfe, J.A.; St Clair, M.H.; Weinhold, K.; Rideout, J.L.; Freeman, G.A.; Lehrman, S.N.; Bolognesi, D.P.; Broder, S.; Mitsuya, H.; Barry, D.W. Proc. Natl. Acad. Sci. USA, 1986, 83, 8333. Benatti, U.; Giovine, M.; Damonte, G.; Gasparini, A.; Scarfì, S.; De Flora, A.; Fraternale, A.; Rossi, L.; Magnani, M. Biochem. Biophys. Res. Commun., 1996, 220, 20. Magnani, M.; Giovine, M.; Fraternale, A.; Damonte, G.; Rossi, L.; Scarfì, S.; Benatti, U.; De Flora. A. Drug Deliv., 1995, 2, 57. Zocchi, E.; Guida, L.; Franco, L.; Silvestro, L.; Guerrini, M.; Benatti, U.; De Flora, A. Biochem. J., 1993, 295, 121. Robbins, B.L.; Rodman, J.; McDonald, C.; Srinivas, R.V.; Flynn, P.M.; Fridland, A. Antimicrob. Agents Chemother., 1994, 38, 115. Perno, C.F.; Yarchoan, R.; Cooney, D.A.; Hartman, N.R.; Gartner, S.; Popovic, M.; Hao, Z.; Gerrard, T.L.; Wilson, Y.A.; Johns, D.G. J. Exp. Med., 1988, 168, 1111. Richman, D.D.; Kornbluth, R.S.; Carson, D.A. J. Exp. Med., 1987, 166, 1144. Gao, W.; Cara, A.; Gallo, R.; Lori, F. Proc. Natl. Acad. Sci. USA, 1993, 90, 8925. Gao, W.; Agraria, R.; Driscoll, J.S.; Mitsuja, H. J. Biol. Chem., 1994, 269, 12633. Perno, C.F; Yarchoan, R.; Cooney, D.A.; Hartman, N.R.; Webb, D.S.; Hao, Z.; Mitsuya, H.; Johns, D.G.; Broder, S. J. Exp. Med., 1989, 169, 933. Magnani, M.; Rossi, L.; Brandi, G.; Schiavano, G.F.; Montroni, M.; Piedimonte, G.; Proc. Natl. Acad. Sci. USA, 1992, 89, 6477.

54 [47] [48] [49] [50] [51] [52] [53] [54]

Curr. Med. Chem. – Anti-Infective Agents, 2005, Vol. 4, No. 1 Lutz, H.U. In Blood Cell Biochemistry; Harris, Ed.; Plenum: New York, 1990; Vol. 1, pp. 81-120. Magnani, M.; Casabianca, A.; Fraternale, A.; Brandi, G.; Chiarantini, L.; Benatti, U.; Scarfì, S.; Millo, E.; De Flora, A. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol., 1998, 17, 189. Golden, M.P.; Kim, S.; Hammer, S.M.; Ladd, E.A.; Schaffer, P.A.; DeLuca, N.; Albrecht, M.A. J. Infect. Dis., 1992, 166, 494. Vlach, J.; Pitha, P.N. J. Virol., 1992, 66, 3616. Heng, M.C.; Heng, S. Y.; Allen, S.G. Lancet, 1994, 343, 255. O’Brien, W.A.; Namazi, A.; Kalhor, H.; Mao, S.H.; Zack, J.A.; Chen, I.S.Y. J. Virol., 1994, 68, 1258. Rossi, L.; Serafini, S.; Cappellacci, L.; Balestra, E.; Brandi, G.; Schiavano, G.F.; Franchetti, P.; Grifantini, M.; Perno, C.F.; Magnani, M. J. Antimicrob. Chemother., 2001, 47, 819. Perno, C.F.; Balestra, E.; Aquaro, S.; Panti, S.; Cenci, A.; Lazzarino, G.; Tavazzi, B.; Di Pierro, D.; Balzarini, J.; Calio, R. Mol. Pharmacol., 1996, 50, 359.

Received: January 06, 2004

Accepted: March 19, 2004

Magnani et al. [55] [56] [57] [58] [59] [60] [61]

Magnani, M.; Rossi, L.; D’Ascenzo, M.; Panzani, I.; Bigi, L.; Zanella, A. Biotechnol. Appl. Biochem., 1998, 28, 1. Rimura, T.; Matsumoto, H.; Matsuda, T.; Hamawaki, T.; Akaji, K.; Kiso, Y. Bioorg. Med. Chem. Lett., 1999, 9, 803. Matsumoto, H.; Hamawaki, T.; Ota, H.; Rimura, T.; Goto, T.; Sano, K.; Hayashi, Y.; Kiso, Y. Bioorg. Med. Chem. Lett., 2000, 10, 1227. Heinonen, P.; Winqvist, A.; Sanghvi, Y.; Stromberg, R. Nucleosides, Nucleotides & Nucleic. Acids, 2003, 22, 1387. Sluis-Cremer, N.; Dmitrienko, G.I.; Balzarini, J.; Camarasa, M.J.; Parniak, M.A. Biochemistry, 2000, 39, 1427. Kers, I.; Girardet, J.L.; Gosselin, G.; Perigaud, C.; Imbach, J.L.; Stawinski, J. Nucleosides & Nucleotides, 1999, 18, 993. Salunke, D.B.; Hazra, B.G.; Pore, V.S.; Bhat, M.K.; Nahar, P.B.; Deshpande, M.V. J. Med. Chem., 2004, 47, 1591.