Inhibition of the in vitro growth of Plasmodium falciparum by acyclic nucleoside phosphonates

International Journal of Antimicrobial Agents 12 (1999) 53 – 61

Original article

Inhibition of the in vitro growth of Plasmodium falciparum by acyclic nucleoside phosphonates L.J.J.W. Smeijsters a, F.F.J. Franssen a, L. Naesens b, E. de Vries a,*, A. Holy´ c, J. Balzarini b, E. de Clercq b, J.P. Overdulve a a

Institute of Infectious Diseases and Immunology, Department of Parasitology and Tropical Veterinary Medicine, Faculty of Veterinary Medicine, Uni6ersity of Utrecht, P.O. Box 80.165, 3508 TD Utrecht, The Netherlands b Rega Institute for Medical Research, Katholieke Uni6ersiteit Leu6en, Leu6en, Belgium c Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Prague, Czech Republic Received 23 July 1998; revised manuscript accepted 17 December 1998

Abstract Fortyeight acyclic nucleoside phosphonates (putative prodrugs of acyclic nucleoside triphosphate inhibitors of DNA replication) have been evaluated for in vitro antiplasmodial activity. Only certain purine derivatives with a hydroxyl group attached to the acyclic sugar moiety displayed antiplasmodial activity. The two most active analogs were (S)-9-(3-hydroxy-2-phosphonylmethoxypropyl)adenine ((S)-HPMPA, IC50 =0.1890.07 mM) and (S)-3-deaza-HPMPA (IC50 = 0.2990.08 mM). Their cyclic derivatives, containing an ester bond between the phosphonate and the hydroxyl group, were slightly less active. All tested compounds that lacked the hydroxyl group, including potent antiretrovirus analogs such as 9-(2-phosphonylmethoxyethyl)adenine (PMEA) and the (S)-HPMPA derivatives (R)-PMPA and (S)-FPMPA, did not show any activity, even at very high concentrations ( \ 250 mM). Similarly, pyrimidine analogs of (S)-HPMPA, such as (S)-HPMPT, (S)-HPMPU and the anti-herpesvirus analog (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine ((S)-HPMPC), were devoid of any antiplasmodial activity. In addition, 11 acyclic nucleoside (non-phosphorylated) analogs — which in contrast to the acyclic nucleoside phosphonates require the presence of a monophosphorylating enzyme for the first activation step — were tested. None of them inhibited the growth of the parasite. In short three chemical entities seem to be imperative for antiplasmodial activity: a purine base, a hydroxyl group in the acyclic side chain and a phosphonate group terminating this chain. © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Plasmodium falciparum; DNA polymerase; Acyclic nucleoside phosphonates; Antimalarials

1. Introduction As long as no suitable vaccine against malaria is available and resistance of the parasite against existing drugs continues to emerge and spread, development of new antimalarials is urgently needed. DNA polymerases (enzymes that play a crucial role in eukaryotic * Corresponding author. Tel.: +31-30-2532582; fax: + 31-302540784. E-mail address: [email protected] (E. de Vries)

and prokaryotic DNA replication) have been shown to be suitable selective targets for chemotherapy, particularly for viruses. Previous research identified significant structural differences between the plasmodial and the human homologues of these enzymes [1,2]. Hence, DNA polymerases could be suitable targets for antimalarial drugs. Several (pro)drugs, mainly nucleoside analogs of which the triphosphate forms inhibit viral polymerases, are currently approved or are being tested for treatment

0924-8579/99/$ - see front matter © 1999 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 9 ) 0 0 0 0 3 - 5

54

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

of viral infections (for review see Refs. [3,4]). Among these are a few nucleoside phosphonates that have been shown to inhibit Plasmodium falciparum and Plasmodium berghei growth in vitro as well as extracted plasmodial DNA polymerases [5]. Inhibition of viral or plasmodial genome replication is dependent on uptake and activation of the nucleoside analogs. These pathways have been partially identified in viral model systems but are still largely unknown in Plasmodium infected red blood cells (RBC). Fig. 1 illustrates some of the revealed or proposed general principles. In virus infected cells uptake via fluid-phase endocytosis (e.g. (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine ((S)-HPMPC) [6], (S)-9-(3-hydroxy2-phosphonylmethoxypropyl)adenine ((S)-HPMPA) [7], 9-(2-phosphonylmethoxyethyl)adenine (PMEA) [7]) as well as by phosphate/phosphonate specific protein-mediated transport [8] has been described. In order to enter the cytosol of a Plasmodium parasite, extracellular compounds have to pass the red blood cell membrane (RCM), the parasite vacuole membrane (PVM) and the parasite plasma membrane. How nucleotides or nucleotide analogs enter the parasite is, however, yet unknown. Non-selective uptake via fluid-phase endocytosis does not occur in P. falciparum infected erythrocytes [9,10]. Different routes for the uptake of nutrients by the infected red cell have been proposed (see for review Refs. [11,12]). Lauer et al. [13] showed that transport through the tubovesicular membrane network (TVN; that extends from the PVM to the RCM) is a rate limiting step in the uptake of nucleosides such as adenosine and thymidine, and of the nucleobase orotate. The authors suggested that these compounds enter this network through an interaction between the TVM and the broad specific, non-saturable, anion selective channels [14] present in the RCM. An alternative possibility of aspecific transport via the proposed parasitophorous duct [15] is highly controversial [11,12,16]. Phosphorylation of nucleoside analogs in virus infected cells proceeds via the successive action of intracellular nucleoside and nucleotide kinases [17]. Part of the selective antiviral action of some of these prodrugs results from the presence of a viral nucleoside kinase [17,18]. Other enzymes such as 5%-nucleotidases [17] or a viral encoded protein kinase [19,20], however, may or have been proven to contribute to this activation. Nucleoside phosphonates, having a phosphate – carbon bond resistant to enzymatic cleavage unlike the normal phosphate–oxygen bond, have been described to be phosphorylated either by a two-step phosphorylation by adenylate kinase [21 – 23], or by direct activation by phosphoribosylpyrophosphate synthetase [24]. Malaria parasites lack pyrimidine deoxy- and ribonucleoside kinases [25] explaining their inability to salvage pyrimidine nucleosides (for review, see Refs. [26,27]). Instead, malaria parasites produce the monophospho-

rylated nucleoside orotidylate (a key substrate for different pathways in pyrimidine synthesis) by phosphoribosylation of the de novo synthesized nucleobase orotate. In contrast, malaria parasites are dependent on host-derived preformed purines. It was shown that purine nucleoside kinases are undetectable [25]. Therefore, similarly to pyrimidine metabolism, the major route of purine salvage proceeds via phosphoribosylation of the nucleobase hypoxanthine, derived either directly from the host or indirectly through catabolic conversion of the nucleoside adenosine (for review, see Refs. [26,28]). The absence of substantial nucleoside kinase activity could hinder the applicability of nucleoside analogs as inhibitors of plasmodial polymerases as is indeed demonstrated below. However, acyclic nucleoside phosphonates do not require activation by nucleoside kinases [29,30] and we have recently demonstrated the uptake and phosphorylation of (S)-

Fig. 1. Schematic representation of steps leading to inhibition of viral or plasmodial replication within their host cells by nucleoside analogs. Abbreviations: AcPy, acyclic pyrimidine nucleoside; AcPu, acyclic purine nucleoside; AcPyCP, acyclic pyrimidine nucleoside phosphonate; AcPuCP, acyclic purine nucleoside phosphonate; AcPyOP, acyclic pyrimidine nucleoside monophosphate, OPpp, triphosphate etc.; PV, parasitic vacuole. Many details of the transport, activation and inhibition steps are discussed in the main text.

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

55

Table 1 Antiplasmodial activity of the different nucleotide/nucleoside sub classes Sub class

No. of compounds tested

Antimicrobial activity IC50 (mM)

I II III IV V

20 15 9 4 12

0.18–\100 \100 \100 0.57–\100 \100

Acyclic purine nucleoside phosphonates with hydroxyl group Acyclic purine nucleoside phosphonates without hydroxyl group Acyclic pyrimidine nucleoside phosphonates Acyclic nucleoside phosphonates with an altered charge Non-phosphonylated acyclic purine nucleosides

HPMPA into the cytosol of P. falciparum in infected RBCs (Smeijsters et al., unpublished). Although the exact mode of action of the acyclic nucleoside phosphonates is not yet fully elucidated (see below), the presence of mutations in the gene coding for the polymerase of different viruses [31 – 35] and polymerase d of P. falciparum [36] which acquired resistance against acyclic nucleoside phosphonates indicates that polymerases are a target of these compounds. Recombinant viruses with introduced mutations in the polymerase gene displayed resistance against these compounds [32,33,35], proving this enzyme indeed to be a selective target. The mode of action of the active metabolite of non-hydroxylated acyclic purine phosphonates such as PMEApp and PMEGpp most likely results from chaintermination [32,37,38]. Recent in vitro studies suggest that hydroxylated derivatives such as HPMPCpp, which can be elongated after incorporation, act in a similar way if two of these nucleotide analogs are incorporated in close proximity to each other [39]. Previous studies [5] demonstrated that acyclic nucleoside phosphonates like (S)-HPMPA and (S)-3deaza-HPMPA inhibited the growth of P. falciparum in vitro, with IC50s comparable to those reported for these analogs against different herpes viruses. In vivo efficacy of (S)-HPMPA against P. berghei in mice, however, has been limited by its stage-dependent inhibition [40], short plasma half-life, and dose and schedule-dependent nephrotoxicity [41]. This paper describes the parasite culture screening of a series of acyclic nucleoside phosphonates, from which we were able to identify several potential new antiplasmodial agents. Furthermore, we have identified the chemical moieties that are essential for antiplasmodial activity.

2. Materials and methods

2.1. Compounds Stock solutions of the test compounds were prepared in PBS, with dropwise addition of 0.1 mM NaOH if necessary for neutralization, and stored in aliquots at

−20°C. Bis(pivaloyloxymethyl)-(2-phosphonylmethoxyethyl)adenine (Bis(POM)PMEA; see Table 1 for abbreviations) and (3-phenyl-2-hydroxypropyl)adenine were dissolved in DMSO. All compounds were synthesized by Dr A. Holy´ (Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Prague, Czech Republic), except for the following: ganciclovir/DHPG (Syntex BV, Rijswijk, The Netherlands); SR3722, SR3772, SR3773, SR 3775 and SR3782 (methylene phosphonates of acyclic guanosine analogs ACV and DHPG, donated by Dr E. Reist, SRI International, Menlo Park, CA); and diphenyl-PMEA and bis(POM)PMEA (Gilead Sciences, Foster City, CA).

2.2. Parasite culture P. falciparum K1 parasites were cultured in suspension with 2% packed cell volume human O + erythrocytes at 37°C under 2% O2/5% CO2/93% N2. Culture medium consisted of modified [42] RPMI 1640 supplemented with 10% human AB + serum, 35 mM hypoxanthine and 50 mg/ml gentamicin. Parasites were synchronized with sorbitol as described [43].

2.3. Drug susceptibility assessment Determination of the IC50 of the compounds listed in Table 1 was performed as described [44]. Briefly, synchronized midterm ringform cultures (infection level9 0.5%) were exposed to different drug concentrations over 72 h. Parasites from 0.8 ml culture samples were then liberated by addition of 1 ml 0.08% saponin/PBS, and free parasites were collected (4 min at 15 800×g). After removal of supernatant by aspiration, the parasite pellet was dissolved in 25 ml guanidinium/Na acetate pH 5.5, and DNA was extracted after addition of 2 ml 2 M NaCl/50 mM Tris–HCl pH 7.8/0.33 mg/ml Hoechst 33258 and 50 ml chloroform/iso-amylalcohol (24:1). Interfering turbidity was removed (1 min at 15 800× g) and the amount of DNA was quantified by measuring DNA-specific Hoechst 33258 fluorescence (model TKO 100, Hoefer Scientific Instruments, San Francisco, CA; emission mercury lamp: 3659 50 nm, detection fluorescence at 4609 5 nm). Measurements

56

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

Fig. 2. Structure formulas of nucleotide analogs of subclasses I – V. One example representative for each class is shown. Class I, acyclic purine nucleoside phosphonates with hydroxyl group. Class II, acyclic purine nucleoside phosphonates without hydroxyl group. Class III, acyclic pyrimidine nucleoside phosphonates. Class IV, acyclic nucleoside phosphonates with altered charge. Class V, acyclic nucleosides.

were corrected for background fluorescence by subtraction of the amount of fluorescence from chloroquinetreated cultures (100% inhibitory concentration of 2 mM) or from samples of non-treated control cultures at t = 0 h. The (net) 100% growth value (corrected for background fluorescence) was determined by sampling of two non-treated control cultures. Control cultures with DMSO were included when this solvent was applied for dissolution of the compound. Dose-response curves were constructed by the statistical computer program; Number Cruncher Statistical Systems (NCSS, Kaysville, UT) after plotting of drug concentrations versus (net) fluorescence values. The IC50 was defined as the drug concentration which corresponds with 50% of the net amount of fluorescence of non-treated control cultures. All cultures were also examined by microscopic evaluation of Giemsa-stained smears.

3. Results and discussion The antiplasmodial activity of 48 acyclic nucleoside phosphonates was determined. These compounds were classified as five groups according to potentially relevant structural features (Fig. 2). Group I contains acyclic purine nucleoside phosphonates with a hydroxyl group in the acyclic sugar moiety that may somehow act as a substitute for the 3%-hydroxyl group of a normal nucleotide (e.g. (S)-HPMPA). Compounds in group II are similar to group I but lack the hydroxyl group (e.g. PMEA). Group III contains acyclic nucleoside phosphonates containing a pyrimidine instead of purine base (e.g. (S)-HPMPC). Compounds in group IV carry changes at the phosponate group (internal ring closure with the hydroxyl group (e.g. cyclic HPMPA) or esterification with lipophilic groups), thus reducing charge and aiming at increased membrane permeability and oral uptake [45] or to prevent nephrotoxicity [46]. Group V contains acyclic purine nucleosides (e.g. gancyclovir).

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

57

Table 2 Antiplasmodial activity of acyclic purine nucleoside phosphonates with hydroxyl groupa Compoundb

Abbreviation

IC50 (mM)

(S)-(3-hydroxy-2-phosphonylmethoxypropyl)adenine (R)-(3-hydroxy-2-phosphonylmethoxypropyl)adenine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)guanine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-chloroadenine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-aminopurine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2,6-diaminopurine (R)-(3-hydroxy-2-phosphonylmethoxypropyl)-2,6-diaminopurine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-aminomethyladenine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-aminomethylhypoxantine (RS)-(3-hydroxy-2-phosphonylmethoxypropyl-)2-methyladenine (RS)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-methyl-thioadenine

(S)-HPMPA (R)-HPMPA (S)-HPMPG

0.18 90.07 6c 4.3 9 0.9 4.4 9 1.0 18.3 9 4.0 2.1 9 0.3 1.5 90.3 \100 \100 \100 \100

(S)-(3-hydroxy-2-phosphonylmethoxypropyl)-3-deazaadenine (R)-(3-hydroxy-2-phosphonylmethoxypropyl)-3-deazaadenine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-7-deazaadenine (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-2-azaadenine 7-(S)-(3-hydroxy-2-phosphonylmethoxypropyl)adenine 7-(S)-(3-hydroxy-2-phosphonylmethoxypropyl)-8-azaadenine (R)-[(1-hydroxymethyl-3-phosphonyl)propyloxymethyl]guanine (S)-[(1-hydroxymethyl-3-phosphonyl)propyloxymethyl]guanine

(S)-HPMPMAP (S)-HPMPDAP (R)-HPMPDAP

(S)-3-deaza-HPMPA (R)-3-deaza-HPMPA (S)-7-deaza-HPMPA 7-(S)-HPMPA SR3773 SR3772

0.29 90.08d 9100c 4.35 9 0.65d 4.4 90.7 15.0 91.7 \100 9.8 9 1.2 \100

a The indicated IC50 represents the mean ( 9 S.D.) of three independent experiments. As a rule, inactivity of compounds (IC50\100 mM) was confirmed in a second independent experiment. b If not indicated, the acyclic sugar moiety is at position 9 and 1 of the purine and pyrimidine base, respectively. c Only one test could be performed due to insufficient amount of compound. d Deviation of the mean IC50 from two independent tests.

Table 1 summarizes the results obtained with all subclasses showing that only subclass I and IV contain inhibitory compounds. Results of the individual compounds of subclass I and IV are listed in Tables 2 and 3. Names of all inactive compounds of subclasses II, III and V are listed in the addendum (Section 4).

3.1. Acyclic purine phosphonates with a hydroxyl-group: a class with potential antiplasmodial acti6ity (S)-HPMPA and (S)-3-deaza-HPMPA were the most potent of a group of P. falciparum inhibitors from subclasses I and IV. The observed IC50 values (0.18 and 0.29 mM) were higher than previously reported values of 47 and 8 nM [5] which can be explained by the different methods used for IC50 determination. Here we measured the increase in DNA content whereas previously the increase in parasitaemia after 72 h was measured. Even at relatively high (S)-HPMPA concentrations, DNA replication proceeded quite well to the eight-nucleated schizont stage [40]. However, low concentrations of (S)-HPMPA already prohibited further nuclear division to mature schizonts and thus even at low concentrations no increase in parasitaemia was observed. The presence of a hydroxyl moiety seems essential for antiplasmodial activity, as all compounds tested with-

out this group (subclasses II, III and V) lacked activity. For instance, while (S)-HPMPA displayed marked antiplasmodial activity, replacement of the 3%-hydroxyl group it by a hydrogen or fluorine atom (yielding (R)-PMPA or (S)-FPMPA) resulted in a complete loss of activity. In several compounds of subclass IV (Table 3) the hydroxyl group was blocked by cyclic ester formation between the phosphonate group and the hydroxyl group ((S)-cHPMPA, (S)-3-deaza-cHPMPA, SR3775). This resulted in a minor (three-fold) decrease in antiplasmodial activity. Mendel et al. [47] reported that the ester bond in (S)-cHPMPC is stable in plasma and is hydrolyzed efficiently by a broad substrate-specific intracellular cCMP phosphodiesterase. Whether this or another related phosphodiesterase was responsible for the cleavage of the antiplasmodial cyclic derivatives and whether their reduced activity resulted from a less efficient uptake or incomplete intracellular conversion to the non-cyclic form requires further investigation. The presence of an hydroxyl group in the acyclic phosphonates was shown to be closely related to their anti-DNA-virus activity [48,49]. While (S)-HPMPA was a potent anti-DNA-virus compound, with no activity against retroviruses (i.e. human immunodeficiency virus), the inverse antiviral spectrum was being displayed by (S)-FPMPA [48].

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

58

Table 3 Antiplasmodial activity of purine nucleoside phosphonates with an altered chargea Compoundb

Abbreviation

IC50 (mM)

(S)-(3-hydroxy-2-phosphonylmethoxypropyl)adenine (S)-[(2-hydroxy-2-oxo-1,4,2,-dioxaphosphorinan-5-yl)methyl]adenine 3%-O-pivaloyl-(RS)-HPMPA-1-oxid diethyl ester 3%-O-pivaloyl-(RS)-HPMPA-1-oxid (S)-HPMPA di(2-propyl)ester (S)-(3-hydroxy-2-phosphonylmethoxypropyl)-3-deazaadenine (S)-[(2-hydroxy-2-oxo-1,4,2,-dioxaphosphorinan-5-yl)methyl]-3-deazaadenin (R)-1-[(2-hydroxy-2-oxo-1,2,-oxaphosphorinan-5-yl)methoxy]guanine (S)-1-[(2-hydroxy-2-oxo-1,2,-oxaphosphorinan-5-yl)methoxy]guanine (2-phosphonylmethoxyethyl)adenine bis(pivaloyloxymethyl)-(2-phosphonylmethoxyethyl)adenine diphenyl-(2-phosphonylmethoxyethyl)adenine

(S)-HPMPA (S)-cHPMPA

0.18 90.07 0.57 90.13 \40 mg/ml \40 mg/ml \100 0.29 90.08c 0.86 90.17 26.4 9 0.8c \100 \250 37.3 9 6.0c \100d

(S)-3-deaza-HPMPA (S)-3-deaza-cHPMPA SR3775 SR3782 PMEA bis(POM)PMEA diphenyl-PMEA

The indicated IC50 represents the mean ( 9 S.D.) of three independent experiments. As a rule, inactivity of compounds (IC50\100 mM) was confirmed in a second independent experiment. b If not indicated, the acyclic sugar moiety is at position 9 and 1 of the purine and pyrimidine base, respectively. c Deviation of the mean IC50 from two independent tests. d Only one test could be performed due to insufficient amount of compound. a

The antiplasmodial activity of several analogs of this active subclass of acyclic nucleoside phosphonates also depended on their stereospecificity. In particular, (R)-HPMPA and (R)-3-deaza-HPMPA displayed much lower activity in parasite cultures than their (S)-forms (Table 2). Merta et al. [21] reported that adenylate kinase phosphorylates different adenine phosphonylmethoxyalkyl derivatives like PMEA, (S)HPMPA and (S)-FPMPA. As this enzyme was incapable of activating the corresponding (R)-enantiomers [37], lack of phosphorylation could explain the lower antiplasmodial activity of the above mentioned (R)forms. Stereospecificity has also been observed in respect of the anti-herpes and anti-retrovirus effects of (S)-HPMPA and (S)-FPMPA, respectively [48]. On the other hand, the (R)- and (S)-form of (3-hydroxy2 - phosphonylmethoxypropyl) - 2,6 - diaminopurine (HPMPDAP) proved to be almost equally active against P. falciparum. A similar lack of stereospecificity has been observed for the antiviral activity of acyclic guanine nucleoside phosphonates such as (3 - hydroxy - 2 - phosphonylmethoxypropyl) - guanine (HPMPG) [48]. Besides adenine derivatives, guanine derivatives such as (S)-HPMPG and SR3773 and its cyclic derivative SR3775 also displayed antiplasmodial activity. The only documented phosphorylation of (S)HPMPG proceeded via the successive action of guanylate kinase and nucleoside diphosphate kinase [50] and also PMEG was not activated by adenylate kinase [22]. Plasmodial guanylate kinases, capable of activating acyclic nucleotides that do not contain an adenine base, have not been identified. All steps illustrated in Fig. 1 (transport, activation

and target inhibition) could be involved in the requirement for a hydroxyl group to make a potent antiplasmodial nucleoside phosphonate. Unfortunately, the near complete lack of radioactive labeled acyclic nucleoside phosphonates hinders the investigation to discriminate between these possibilities. The ineffectiveness of the analogs lacking a hydroxyl group could be related to selective uptake if these acyclic nucleotide analogs enter the infected red cell similarly as described for PMEA via a phosphonate/phosphate specific protein-mediated transport, which shows selectivity towards substitution in the acyclic moiety of these compounds [8]. In this context, the observed moderate activity of bis(POM)PMEA (Table 3) is remarkable since this is a non-charged, lipophilic, membrane permeable ester derivative of the inactive analog PMEA. On the other hand lipophilic esters of the most active analog (S)HPMPA even lacked activity completely. Extended research is needed to elucidate whether this results from the inability of plasmodial (or host cell) esterases to remove the lipophilic aliphatic group. Lack of phosphorylation could also explain the absence of antiplasmodial activity, not only for the nonhydroxyl containing analogs, but also for acyclic nucleoside phosphonates in general. It was postulated that especially the differential metabolism of acyclic nucleoside phosphonates determines whether these drugs ultimately inhibit viral replication. This suggestion was based on the observation that although the diphosphorylated metabolite of (S)-HPMPA (designated (S)-HPMPApp) inhibits reverse transcriptase in vitro, the non-activated compound does not inhibit retrovirus replication in cell cultures [51]. However,

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

information on the activation of acyclic nucleosides phosphonates is limited and ambiguous. Finally, inactivity of non-hydroxylated analogs like the HPMPA derivatives (R)-PMPA and (S)-FPMPA, if taken up and activated by the parasite, could result from inability of the diphosphoryl form to inhibit the final target. Kramata et al. [37] indeed showed that diphosphorylated acyclic nucleoside phosphonates displayed different inhibitory potencies towards DNA polymerases a, d and o, in vitro.

3.2. Acyclic pyrimidine phosphonates are inacti6e and (S)-HPMPC, (S)-HPMPU, (S) -HPMPT pyrimidine derivatives of the potent antiplasmodial analog (S)-HPMPA, did not at all inhibit the growth of P. falciparum, even at concentrations up to 250 mM (subclass 3). Although (S)-HPMPU does not inhibit herpesvirus replication and (S)-HPMPT shows only moderate activity [49], (S)-HPMPC can be considered as one of the most potent anti-herpesvirus compounds among the acyclic nucleoside phosphonates. Again, several possibilities can be reiterated (see above) to explain the lack of antimalarial activity of (S)-HPMPC. Ineffectiveness of (S)HPMPC against Plasmodium might be due to lack of uptake since to date only uptake of (S)-HPMPC via endocytosis has been reported [6], which does not occur in P. falciparum infected erythrocytes [9,10].

59

Several alterations or substitutions of the purine base reduce antiplasmodial activity (Table 2) whereas other modifications result in a complete loss of activity (e.g. addition of a methyl or aminomethyl group at position 2 of the purine base). An alternative linkage between the sugar component and the nucleobase at position 7 is tolerated as 7-(S)HPMPA retains antiplasmodial activity. A considerable change in the acyclic sugar moiety, as in the guanine phosphonate analog SR3773, does also not completely abolish inhibition of the in vitro growth of the parasite. In contrast, removal of either the hydroxyl group or the phosphonate group invariably yielded compounds without any antiplasmodial activity. The identification of the three chemical entities which seem to be essential for antiplasmodial activity (a purine base, a hydroxyl group in the acyclic side chain and a phosphonate terminating this chain) will enable a more rational approach in the search for nucleotide analogs with selective antimalarial activity.

4. Addendum The following compounds from classes II, III and V (see Fig. 1 for general structure) displayed no antiplasmodial activity at the indicated concentration (mM).

4.1. Class II 3.3. Non-phosphorylated acyclic nucleosides are inacti6e In contrast to the acyclic nucleoside phosphonates, all test compounds that lacked a phosphonate group are also devoid of antiplasmodial activity (subclass 5). Lack of (substantial) nucleoside kinase activity in Plasmodium (Fig. 1) can certainly be considered as one of the factors that contribute to the inactivity of these analogs. The differential activities of DHPG (ganciclovir), (IC50 \100 mM) and its closely related methylene phosphonate derivative SR3773 (IC50 of 9.8 mM) supports this hypothesis. Furthermore, it is known that DHPG permeates erythrocytes via both the nucleoside and purine nucleobase carrier [52] whose transport capacity are increased upon infection with P. falciparum [26,50]. Therefore, lack of antiplasmodial activity of DHPG, which selective antiviral effect in part results from activation via a viral encoded protein kinase [19,20], most likely result from lack of phosphorylation in Plasmodium. In conclusion, the S-enantiomers of HPMPA and 3-deaza-HPMPA still display the most potent antiplasmodial activity of all compounds tested here.

(2-phosphonylmethoxyethyl)diaminopurine (250), (S)-(2-phosphonylmethoxypropyl)adenine (250), (R)-(2phosphonylmethoxypropyl)adenine (250), (R)-(2-phosphonylmethoxypropyl) - 2,6 - diaminopurine (250), (S)(3 - fluoro - 2 - phosphonylmethoxypropyl)adenine (250), (RS)-(3-methoxy-2-phosphonylmethoxypropyl)adenine (100), (RS)-(2-phosphonylmethoxybutyl)adenine (100), (RS)-(3-phosphonylmethoxybutyl)adenine (100), (RS)(3-methyl-2-phosphonylmethoxybutyl)adenine (100), (5phosphonylpentyl)adenine (100), (5 - phosphonylmethoxypentyl)adenine (100), (RS)-(2-phosphonylmethoxypentyl)adenine (100), (RS)-(4-methyl-2-phosphonylmethoxypentyl)adenine (100), (R)-(3-phosphonylpropyloxymethyl)guanine (250).

4.2. Class III (S)-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine (250), (2-phosphonylmethoxyethyl)cytosine (250), (S)(3-hydroxy-2-phosphonylmethoxypropyl)thymine (150), (S) - (3 - hydroxy - 2 - phosphonylmethoxypropyl)uracil (250).

60

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61

4.3. Class V (3-phenyl-2-hydroxypropyl)adenine (100), (RS)-(3azido-2-hydroxypropyl)-1-deazaadenine (100), (RS)-(3azido-2-hydroxypropyl)-3-deazaadenine (100), (RS)-(3amino-2-hydroxypropyl)-1-deazaadenine (100), (RS)-(3amino-2-hydroxypropyl)-3-deazaadenine (100), 3-(2-azaadenin-9-yl)-2-hydroxypropanoic acid (100), 9-(1,3-dihydroxy-2-propoxymethyl)guanine (ganciclovir) (100), (RS)-(2-fluoro-3-hydroxypropyl)adenine (100), (LS)(2,3 dihydroxypropyl)adenine (100), (2,3 dihydroxypropyl)-2-aminomethyladenine (100), (S)-(2,3 dihydroxypropyl)-7-deazaadenine (100).

Acknowledgements We thank Dr A.W.C.A. Cornelissen for valuable suggestions and advice. This work was supported by the Netherlands Minister of Development Co-operation; grant NL/92/851. A. Holy´ and L. Naesens were supported by EC grant TS*-CT94-0297.

References [1] De Vries E, Stam JG, Franssen FJJ, van der Vliet PC, Overdulve JP. Purification and characterization of DNA polymerases from Plasmodium berghei. Mol Biochem Parasitol 1991;45:223– 32. [2] Makioka A, Ellis J. DNA polymerases of parasitic protozoa. Int J Parasitol 1994;24:463–76. [3] De Clercq E. Antiviral therapy for human immunodeficiency virus infections. Clin Microb Rev 1995;8:200–39. [4] Field AK, Biron KK. ‘‘The end of innocence’’ revisited: resistance of herpesviruses to antiviral drugs. Clin Microb Rev 1994;7:1 – 13. [5] De Vries E, Stam JG, Franssen FJJ, et al. Inhibition of the growth of Plasmodium falciparum and Plasmodium berghei by the DNA polymerase inhibitor HPMPA. Mol Biochem Parasitol 1991;47:43 – 50. [6] Connelly MC, Robbins BL, Fridland A. Mechanism of uptake of the phosphonate analog (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine (HPMPC) in Vero cells. Biochem Pharmacol 1993;46:1053 –7. [7] Palu G, Stefanelli S, Rassu M, Parolin C, Balzarini J, de Clercq E. Cellular uptake of phosphonylmethoxyalkylpurine derivatives. Antiviral Res 1991;16:115–9. [8] Cihlar T, Rosenberg I, Votruba I, Holy´ A. Transport of 9-(2phosphono methoxyethyl)adenine across plasma membrane of HeLa S3 cells is protein mediated. Antimicrob Agents Chemother 1995;39:117–24. [9] Haldar K, Uyetake L. The movement of fluorescent endocytic tracers in Plasmodium falciparum infected erythrocytes. Mol Biochem Parasitol 1992;50:161–78. [10] Haldar K. Ducts, channels and transporters in Plasmodium-infected erythrocytes. Parasitol Today 1994;10:393–5. [11] Elford BC, Cowan GM, Ferguson DJP. Parasite-regulated membrane transport processes and metabolic control in malaria-infected erythrocytes. J Biochem 1995;308:361–74. [12] Ginsburg H. Transport pathways in the malaria-infected erythrocyte. Biochem Pharmacol 1994;48:1847–56.

[13] Lauer SA, Rathod PK, Ghori N, Haldar K. A membrane network for nutrient import in red cells infected with the malaria parasite. Science 1997;276:1122– 5. [14] Kirk K, Horner HA, Elford BC, Ellory JC, Newbold CI. Transport of diverse substrates into malaria-infected erythrocytes via a pathway showing functional characteristics of a chloride channel. J Biol Chem 1994;269:3339 – 47. [15] Pouvelle B, Spiegel R, Hsiao L, et al. Direct access to serum macromolecules by intraerythrocytic malaria parasites. Nature 1991;353:73 – 5. [16] Gero AM, Kirk K. Nutrient transport pathways in Plasmodiuminfected erythrocytes: what and where are they? Parasitol Today 1994;10:395 – 9. [17] Shugar D. Phosphorylating enzymes involved in activation of chemotherapeutic nucleosides and nucleotides. In: Shugar D, Rode W, Borowski E, editors. Molecular Aspects of Chemotherapy. Springer – Verlag/Polish Scientific Publishers PWN, 1991:239 – 270. [18] De Clercq E. Antiviral agents: characteristic activity spectrum depending on the molecular target with which they interact. Adv Virus Res 1993;42:1 – 55. [19] Littler E, Stuart AD, Chee MS. Human cytomegalovirus UL97 open reading frame encodes a protein that phosphorylates the antiviral nucleoside analogue ganciclovir. Nature 1992;358:160– 2. [20] Sullivan V, Talarico CL, Stanat SC, Davis M, Coen DM, Biron KK. A protein kinase homologue controls phosphorylation of ganciclovir in human cytomegalovirus infected cells. Nature 1992;358:162 – 4. [21] Merta A, Votruba I, Jindrich J, et al. Phosphorylation of 9-(2-phosphono methoxyethyl)adenine and 9-(S)-(3-hydroxy-2phosphonomethoxypropyl)adenine by AMP(DAMP) kinase from L1210 cells. Biochem Pharmacol 1992;44:2067 – 77. [22] Robbins BL, Connelly MC, Marshall DR, Srinivas RV, Fridland A. A human T lymphoid cell variant resistant to the acyclic nucleoside phosphonate 9-(2-phosphonylmethoxyethyl)adenine shows a unique combination of a phosphorylation defect and increased efflux of the agent. Mol Pharmacol 1995;47:391–7. [23] Robbins BL, Greenhaw J, Connelly MC, Fridland A. Metabolic pathways for activation of the antiviral agent 9-(2-phosphonylmethoxyethyl)adenine in human lymphoid cells. Antimicrob Agents Chemother 1995;39:2304 – 8. [24] Balzarini J, Hao Z, Herdewijn P, Johns DG, de Clercq E. Intracellular metabolism and mechanism of anti-retroviral action of 9-(2-phosphonyl methoxyethyl)adenine, a potent anti-human immunodeficiency virus compound. Proc Natl Acad Sci USA 1991;88:1499 – 503. [25] Reyes P, Rathod PK, Sanchez DJ, Mrema JEK, Rieckmann KH, Heidrich HG. Enzymes of purine and pyrimidine metabolism from the human parasite, Plasmodium falciparum. Mol Biochem Parasitol 1982;5:275 – 90. [26] Gero AM, O’Sullivan WJ. Purines and pyrimidines in malaria parasites. Blood Cells 1990;16:467 – 84. [27] Sherman IW. Biochemistry of Plasmodium (Malaria parasites). Microbiol Rev 1979;43:453 – 95. [28] Van Dyke K. Purines and pyrimidines in malaria parasites (Commentary). Blood Cells 1990;16:485 – 95. [29] Balzarini J. Metabolism and mechanism of antiretroviral action of purine and pyrimidine derivatives. Pharm Wld Sci 1994;16:113 – 26. [30] Neyts J, de Clercq E. Mechanism of action of acyclic nucleoside phosphonates against herpes virus replication. Biochem Pharmacol 1994;47:39 – 41. [31] Cherrington JM, Mulato AS, Fuller MD, Chen MS. Novel mutation (K70E) in human immunodeficiency virus type 1 reverse transcriptase confers decrease susceptibility to 9-[2-(phosphonomethoxy)ethyl]adenine in vitro. Antimicrob Agents Chemother 1996;40:2212 – 6.

L.J.J.W. Smeijsters et al. / International Journal of Antimicrobial Agents 12 (1999) 53–61 [32] Gu Z, Salomon H, Cherrington JM, et al. K65R mutation of human immunodeficiency virus type 1 reverse transcriptase encodes cross-resistance to 9-(2-phosphonylmethoxyethyl)adenine. Antimicrob Agents Chemother 1995;39:1888–91. [33] Lurain NS, Thompson KD, Holmes EW, Read GS. Point mutations in the DNA polymerase gene of human cytomegalovirus that result in resistance to antiviral agents. J Virol 1992;66:7146 – 52. [34] Mendel DB, Barkhimer DB, Kern ER, Chen MS. Isolation and initial characterization of a herpes simplex virus type 2 (HSV-2) strain with decreased sensitivity to cidovir. 8th International conference of antiviral research, Santa Fe, New Mexico. 1995; abstract 162. [35] Sullivan V, Biron KK, Talarico C, et al. A point mutation in the human cytomegalovirus DNA polymerase gene confers resistance to Ganciclovir and phosphonylmethoxyalkyl derivatives. Antimicrob Agents Chemother 1993;37:19–25. [36] Smeijsters LJJW, Zijlstra NM, de Vries E, et al. Identification of two conserved amino acid substitutions in a residue encoded by the polymerase ( block C of (S)-9-(3-hydroxy-2-phosphonylmethoxypropyl)adenine [(S)-HPMPA]-resistant Plasmodium falciparum clones. In: Smeijsters LJJW, editor. Ph.D Thesis. Utrecht: Utrecht University, 1996:67–86. [37] Kramata P, Votruba I, Otova B, Holy´ A. Different inhibitory potencies of acyclic phosphonomethoxyalkyl nucleotide analogs toward DNA polymerase a, d and o. Mol Pharmacol 1996;49:1005 – 11. [38] Pisarev VM, Lee S, Connelly MC, Fridland A. Intracellular metabolism and action of acyclic nucleoside phosphonates on DNA replication. Mol Pharmacol 1997;52:63–8. [39] Xiong X, Smith JL, Chen MS. Effect of incorporation of cidovir into DNA by human cytomegalovirus DNA polymerase on DNA elongation. Antimicrob Agents Chemother 1997;41:594 – 9. [40] Smeijsters LJJW, Zijlstra NM, de Vries E, Franssen FJJ, Janse CJ, Overdulve JP. The effect of (S)-9-(3-hydroxy-2-phosphonylmethoxypropyl) adenine on nuclear and organellar DNA synthesis in erythrocytic schizogony in malaria. Mol Biochem Parasitol 1994;67:115 – 24. [41] Smeijsters LJJW, Nieuwenhuijs H, Hermsen RC, Dorrestein G, Franssen FJJ, Overdulve JP. Antimalarial and toxic effects of the acyclic nucleoside phosphonate (S)-9-(3-hydroxy-2-phosphonylmethoxypropyl)adenine in Plasmodium berghei-infected mice. Antimicrob Agents Chemother 1996;40:1584–8.

.

61

[42] Fairlamb AH, Warhurst DC, Peters W. An improved technique for the cultivation of Plasmodium falciparum in vitro without daily medium change. Ann Trop Med Parasitol 1984;79:379–84. [43] Lambros C, Vandenberg JP. Synchronization of Plasmodium falciparum erythrocytic stages in culture. J Parasitol 1979;65: 418 – 20. [44] Smeijsters LJJW, Zijlstra NM, Franssen FJJ, Overdulve JP. Simple, fast and accurate fluorometric method to determine drug susceptibility of Plasmodium falciparum in 24-well suspension cultures. Antimicrob Agents Chemother 1996;40:835 – 8. [45] Srinivas RV, Robbins BL, Connelly MC, Gong Y, Bischofberger N, Fridland A. Metabolism and in vitro antiretroviral activities of Bis(pivaloyloxymethyl) prodrugs of acyclic nucleoside phosphonates. Antimicrob Agents Chemother 1993;37:2247–50. [46] Bischofberger N, Hitchcock MJM, Chen MS, et al. 1-[((S)-2-hydroxy-2-oxo-1,4,2-dioxaphosphorinan-5-yl)methyl]cytosine, an intracellular prodrug for (S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine with improved therapeutic index in vivo. Antimicrob Agents Chemother 1995;38:2387 – 91. [47] Mendel DB, Cihlar T, Moon K, Chen MS. Conversion of 1-[((S)-2-hydroxy-2-oxo-1,4,2-dioxaphosphorinan-5-yl)methyl)

[48]

[49]

[50]

[51]

[52]

cytosine to cidovir by an intracellular cyclic CMP phosphodiesterase. Antimicrob Agents Chemother 1997;41:641 – 6. Balzarini J, Holy A, Jindrich J, et al. Differential antiherpesvirus and antiretrovirus effects of the (S) and (R) enantiomers of acyclic nucleoside phosphonates: potent and selective in vitro and in vivo antiretrovirus activities of (R)-9-(2-phosphonomethoxypropyl)2,6-diamonopurine. Antimicrob Agents Chemother 1993;37:332 – 8. De Clercq E, Sakuma T, Baba M, et al. Antiviral activity of phosphonyl methoxyalkyl derivatives of purine and pyrimidines. Antiviral Res 1987;8:261 – 72. Terry BJ, Mazina KE, Tuomari AV, et al. Broad-spectrum antiviral activity of the acyclic guanosine phosphonate (R,S)HPMPG. Antiviral Res 1988;10:235 – 52. Aduma P, Connelly MC, Srinivas RV, Fridland A. Metabolic diversity and antiviral activities of acyclic nucleoside phosphonates. Mol Pharmacol 1995;47:816 – 22. Mahony WB, Domin BA, Zimmerman TP. Ganciclovir permeation of human erythrocyte membrane. Biochem Pharmacol 1991;41:263 – 71.