| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Article |
Malaria Section, Center for Vaccine Development, University of Maryland School of Medicine, Baltimore, Maryland
Abstract
Abstract I. Introduction A. Life Cycle of Plasmodium falciparum B. Folate Biosynthesis in Plasmodia II. Malaria Parasite Drug Resistance A. Historical Perspective 1. Chloroquine. 2. Synthetic Antimalarials. 3. Proguanil. 4. Pyrimethamine. 5. Sulfa Drugs. 6. Combined Dihydrofolate Reductase Inhibitors and Sulfonamide Drugs. 7. Aryl Amino Alcohols. 8. General Concepts Learned from Early Experiences. B. Antifolates and Nonfalciparum Malaria C. Drug Effects on Parasite Stages D. Parasite Clearance Following Antimalarial Drug Treatment III. Dihydrofolate Reductase-Thymidylate Synthase A. Dihydrofolate Reductase Inhibitors 1. Cross-Resistance between Dihydrofolate Reductase Inhibitors. B. Identification of Antifolate Drug Target C. Point Mutations within Dihydrofolate Reductase Are Responsible for in Vitro Resistance D. The Move to Field Isolates E. Gene Amplification F. Mutation Rates within the Dihydrofolate Reductase Gene G. Enzyme Kinetic Analysis of Dihydrofolate Reductase H. Relationship of Point Mutations to Dihydrofolate Reductase Structure—Crystallography IV. Pyrophosphokinase-Dihydropteroate Synthase A. Folate Effect 1. Folate Effect and Drug Resistance. 2. Folate Effect and in Vitro Sulfonamide Testing. B. Markers of in Vitro Resistance in Dihydropteroate Synthase C. Enzyme Kinetics Studies on Dihydropteroate Synthase D. Relationship of Point Mutations to Dihydropteroate Synthase Structure—Crystallography V. Parasitologic Resistance Does Not Equal Clinical Failure A. In Vivo Drug Failure, Additional Host Factors 1. In Vivo Folate Effect. B. Molecular Markers and Treatment Outcomes VI. Molecular Assays VII. Molecular Epidemiological Studies A. Drug Treatment Effect on Post-Treatment Parasite Genotype B. Molecular Markers and Treatment Outcomes 1. High Endemicity. 2. Low Endemicity. C. Worldwide Distribution of Dihydrofolate Reductase and Dihydropteroate Synthase Mutations D. Molecular Markers and Treatment Outcome—Summary VIII. Using Genotype to Predict Clinical Failure IX. Other Antifolates A. Trimethoprim-Sulfamethoxazole B. Chlorproguanil-Dapsone X. New Directions—Drug Development A. The ''Old'' Combinations B. New Directions—Combination Drug Therapy XI. Summary
Antifolate antimalarial drugs interfere with folate metabolism, a pathway essential to malaria parasite survival. This class of drugs includes effective causal prophylactic and therapeutic agents, some of which act synergistically when used in combination. Unfortunately, the antifolates have proven susceptible to resistance in the malaria parasite. Resistance is caused by point mutations in dihydrofolate reductase and dihydropteroate synthase, the two key enzymes in the folate biosynthetic pathway that are targeted by the antifolates. Resistance to these drugs arises relatively rapidly in response to drug pressure and is now common worldwide. Nevertheless, antifolate drugs remain first-line agents in several sub-Saharan African countries where chloroquine resistance is widespread, at least partially because they remain the only affordable, effective alternative. New antifolate combinations that are more effective against resistant parasites are being developed and in one case, recently introduced into use. Combining these antifolates with drugs that act on different targets in the parasite should greatly enhance their effectiveness as well as deter the development of resistance. Molecular epidemiological techniques for monitoring parasite drug resistance may contribute to development of strategies for prolonging the useful therapeutic life of this important class of drugs.
Malaria is a major burden for the most resource-poor nations of the world. The goal of eradicating malaria, once thought to be possible, was abandoned decades ago, and the present goal of malaria control is instead first to retard the accelerating rates of disease and death caused by the world's most important parasitic disease and then to "roll back malaria".
Between 200 and 500 million cases of malaria occur annually with an estimated 1.7 to 3 million deaths attributable to malaria, most among children of sub-Saharan Africa (Breman, 2001
). Many nations in sub-Saharan Africa have faced socio-economic instability and a dismantling of government sector malaria control programs (Garfield and Vermund, 1983; World Bank, 1993). The combination of such factors as the increased cost of insecticides, the vector's resistance to insecticides, and the lack of an effective vaccine, has resulted in reliance upon case management and effective curative chemotherapy as the primary approach to malaria control. Malaria parasite resistance to treatment with chloroquine has already complicated malaria management and has been associated with increased malaria morbidity and mortality (Greenberg et al., 1989; Trape et al., 1998), and increasing resistance to sulfadoxine-pyrimethamine will likely lead to similar results where it is the first-line antimalarial. Resistance of the malaria parasite, especially to chloroquine and to the antifolates, will continue to make progress in rolling back malaria a formidable challenge for the foreseeable future.
A. Life Cycle of Plasmodium falciparum
Four malaria species cause disease in humans: Plasmodium vivax, P. malariae and P. ovale, and the cause of most severe malaria disease and deaths, P. falciparum. Dozens of other Plasmodia species cause disease in other mammals, birds, and reptiles. Malaria parasites have a complex life cycle, involving both vertebrate (human) and invertebrate hosts (mosquitoes). Saliva from infected mosquitoes transmits the veriform malaria sporozoites to the subcutaneous tissues of the human host when the female mosquito takes a blood meal. The sporozoites travel rapidly to the liver and invade hepatocytes, where they develop into an exoerythrocytic stage called a tissue schizont. After 6 to 10 days, these exoerythrocytic schizonts undergo schizogony, multiplying via mitosis until they rupture the infected hepatocytes and discharge tens of thousands of merozoites from each infected hepatocyte into the bloodstream. This release of merozoites from the liver appears to be a continuous and asynchronous process in falciparum malaria (Murphy et al., 1990). The merozoites then invade erythrocytes where they again multiply and, after 48 h (72 h in the case of P. malariae), release 8 to 32 progeny merozoites. The progeny merozoites invade new erythrocytes to perpetuate the erythrocytic cycle, the stage of the parasite life cycle responsible for disease. A small percentage of the merozoites do not multiply after invading erythrocytes, but instead differentiate into sexual forms termed gametocytes. When gametocytes are ingested by a mosquito in a subsequent blood meal, male and female gametes mate, creating a zygote. This brief diploid stage in an otherwise haploid life cycle allows for sexual recombination of genetic material, including the chromosomal genes responsible for most drug resistance. Within the mosquito midgut the zygote matures into an oocyst, which in turn releases sporozoites that then migrate to the mosquito salivary glands, completing the life cycle.
B. Folate Biosynthesis in Plasmodia
The empiric use of antifolates against malaria long predates definitive demonstration of the folate metabolism pathway in Plasmodium spp. De novo synthesis of folate by Plasmodium spp. was demonstrated over 25 years ago (Ferone, 1977), and although an exogenous folate salvage pathway has been found in isolates from around the world (Krungkrai et al., 1989), it does not appear to be Plasmodia's primary source of folate. Not all of the enzymes involved in folate metabolism have been identified in Plasmodium spp. (Fig. 1) (http://malaria.atcc.org/metabolic_pathways/maps/folatebiopath.html), but this will undoubtedly change with the sequencing of the malaria genome. The genes encoding the enzymes in the folate pathway targeted by existing antifolate drugs, dihydrofolate reductase (DHFR1) (Bzik et al., 1987) and dihydropteroate synthase (DHPS) (Brooks et al., 1994; Triglia and Cowman, 1994), have both been cloned and sequenced, and mutations in these genes have been determined to play a role in resistance to the antifolate drugs (Peterson et al., 1988). Disruption of folate synthesis by DHFR and DHPS inhibitors leads to decreased levels of fully reduced tetrahydrofolate, a necessary cofactor in important one-carbon transfer reactions in the purine, pyrimidine, and amino acid biosynthetic pathways (Ferone, 1977). The lower levels of tetrahydrofolate result in decreased conversion of glycine to serine, reduced methionine synthesis, and lower thymidylate levels with a subsequent arrest of DNA replication (Schellenberg and Coatney, 1961; Gutteridge and Trigg, 1971; Newbold et al., 1982; Gritzmacher and Reese, 1984; Triglia and Cowman, 1999).
|
II. Malaria Parasite Drug Resistance
To understand the important role that antifolate drugs currently play in malaria drug therapy, it is helpful to review the history of malaria drug development and the near parallel development of drug resistance in the malaria parasite. The antimalarial drug armamentarium in the Western world was limited to the cinchona bark extract quinine until the first World War, when declining stocks of quinine in Germany led to the development of the first synthetic antimalarials. Work with these synthetic dyes led to the development of the acridines and the 8-aminoquinolines, such as pamaquine (and subsequently primaquine), which were more toxic than quinine and therefore left quinine once again as the primary antimalarial. Quinine resistance was first documented in Brazil in 1908 and again demonstrated in 1938 in German railroad workers returning from the Madeira-Mamoré railroad on the Brazilian-Bolivian border (Clyde, 1972b). This resistance represented a true parasite drug tolerance because infections in these individuals were not cured by successive increases in quinine dose. Varying degrees of quinine resistance now can be found worldwide, but it is most common and severe in Southeast Asia.
1. Chloroquine.
Although chloroquine (Resochin) was synthesized in 1934, and amodiaquine shortly afterward, it initially was felt to be too toxic for use. Chloroquine, however, became the cornerstone of the malaria eradication campaign of the 1950s and 1960s. High-level chloroquine resistance (RIII, see Table 1) was first found in Thailand in 1962 (Harinasuta et al., 1965), then spread gradually and contiguously throughout Southeast Asia [Malaysia (Montgomery and Eyles, 1963), Cambodia (Eyles et al., 1963), Vietnam (Powell et al., 1964; Eppes et al., 1966), and Burma in 1969 (Clyde et al., 1972)]. At about the same time, chloroquine resistance also appeared in South America (Moore and Lanier, 1961). Chloroquine resistance appeared later in Africa, initially in nonimmune travelers returning from East Africa (Campbell et al., 1979; Fogh et al., 1979, 1984; Jepsen et al., 1983). Despite the current high levels of chloroquine resistance in most areas of malaria transmission, the cost of less than $0.20 U.S. per treatment continues to apply sufficient financial incentive to maintain chloroquine as a first-line treatment in much of West and Central Africa (Foster, 1994).
|
2. Synthetic Antimalarials. During World War II, the lack of access to the world's major supply of cinchona bark, and therefore quinine, again spurred development of synthetic antimalarials, such as mepacrine (quinacrine). Simultaneously, drug development focused on derivatives of pyrimidine, based upon its presence in nucleic acids and its metabolism by protein systems that were effectively blocked by antimalarial sulfonamides. Research in this direction resulted in the development of the antifolate biguanides, proguanil and chlorproguanil.
3. Proguanil.
Early reports on the use of proguanil (Paludrine) for both prophylaxis and treatment were very encouraging (Maegraith et al., 1945, 1946; Jones et al., 1948; Seaton and Lourie, 1949), despite a slower schizontocidal action (defined as activity against any asexual blood stage parasite, not only schizonts) compared with quinine or mepacrine (Covell et al., 1949). Thereafter, when proguanil was used in the late 1940s and the early 1950s as prophylaxis for plantation workers in Southeast Asia and elsewhere, it provided an opportunity for widespread drug selection pressure on the parasite and the subsequent development of resistance to this drug. The prophylactic experience with proguanil in Malay may have cast a shadow on the drug's further development. From a 100% clinical cure rate following a one-time 100-mg dose in late 1947, clinical failure rates rose to 25% following a 300-mg dose in early 1949, with some cases failing to clear after a 100-fold increase of this dose (Field and Edeson, 1949; Davey and Robertson, 1957). A similar example from Brazil (Walker and Lopez-Antunano, 1968) indicates that resistance was not limited to Southeast Asian isolates.
4. Pyrimethamine.
Developed shortly after proguanil, the DHFR inhibitor pyrimethamine (Daraprim) was a remarkably effective causal prophylactic and therapeutic agent (Archibald, 1951 first use in natural infection; Goodwin, 1952b; Vincke and Lips, 1952; Delannoy and Hugon, 1954; Miller, 1957), even against chloroquine-resistant parasites (Powell et al., 1963) (for excellent reviews of early experience with pyrimethamine, see Goodwin, 1952a and Hitchings, 1960). However, concerns about the rapid development of parasite resistance to pyrimethamine and its slow schizontocidal activity were raised shortly after its introduction (Coatney et al., 1952; Goodwin, 1952a; Wilson and Edeson, 1953; Petersen, 1987). Events following the continued use of pyrimethamine as a prophylactic supported this concern (Clyde and Shute, 1954; Jones, 1954; Rollo, 1955; Burgess and Young, 1959). When pyrimethamine was given as weekly prophylaxis to children for 1 year, pyrimethamine resistance increased throughout the course of the year, approaching 60% resistance at year's end (Clyde and Shute, 1957). Resistance, as measured by the dose of pyrimethamine required to clear asexual parasitemia (number of parasites per unit volume blood), increased 8- to 15-fold or possibly more in some cases, as the "resistance approached or exceeded the maximum therapeutic dosage". Transmission of resistant parasites to persons not on prophylactic treatment occurred within the central treatment areas and villages nearby the treatment villages. Resistance was found less commonly in villages two to four miles from the central treatment areas (0 to 7% resistance), and no resistance was found in villages more than five miles distant.
Resistance to both proguanil and pyrimethamine was encountered during the American occupation of Vietnam (Peters, 1970). At this same time, the first report of a multidrug-resistant isolate from Thailand was made (Young et al., 1963; Powell et al., 1964). This isolate and others from nearby locales were resistant to chloroquine, mepacrine, proguanil, pyrimethamine, and partially resistant to quinine, whereas the Malay strain mentioned earlier was resistant to amodiaquine, hydroxychloroquine, quinacrine, chlorguanide, pyrimethamine, sulfadiazine, and chloroquine (DeGowin and Powell, 1964).
5. Sulfa Drugs.
The increased exposure of nonimmune persons to these resistant strains rekindled interest in the sulfonamides and sulfones as antimalarials. Prontosil, the active component of which is sulfanilamide, was developed in 1932. A trial using Prontosil for treatment of falciparum malaria in 1937 cured 100% of 93 individuals following four, 12 hourly injections (Hill and Goodwin, 1937; Niven, 1938; Coggeshall et al., 1941). Interest in sulfonamides then waned, partly because of the introduction of synthetic antimalarials and the continued effectiveness of quinine, until sulfa drugs with longer half-lives and improved toxicity profiles were developed in the late 1950s and 1960s (Hill, 1963) (for a detailed review of early sulfonamide use, see Curd, 1943). Sulfadoxine in particular (now used in combination with pyrimethamine in Fansidar), demonstrated considerable promise as a prophylactic and curative drug against P. falciparum in Tanzania (Laing, 1965a). Several early trials confirmed that a single 1-g dose of sulfadoxine was an effective, although slow, schizontocide. Although in vivo resistance to sulfonamides and sulfones could be demonstrated with relative ease in animal models (Scholer et al., 1984), in vivo human field studies continued to offer encouragement. For example, weekly sulfadoxine (500 mg) prophylaxis resulted in no positive blood smears at the end of 6 weeks compared with a 26% positive rate with pyrimethamine prophylaxis (Laing, 1964).
6. Combined Dihydrofolate Reductase Inhibitors and Sulfonamide Drugs.
A 1959 study found that sulfadoxine potentiated pyrimethamine in human falciparum infections, demonstrating that combined pyrimethamine and sulfadoxine was more effective than either drug alone (Greenberg and Richeson, 1950; Hurly, 1959). In other field studies of pyrimethamine combined with a sulfonamide, in pyrimethamine-resistant or multidrug-resistant infections, the combination was superior to either drug or chloroquine alone (McGregor et al., 1963; DeGowin and Powell, 1964; Chin et al., 1966; Harinasuta et al., 1967; Laing, 1968b, 1970b; Martin and Arnold, 1968b). Sulfadoxine schizontocidal activity remained slower than that of chloroquine (DeGowin and Powell, 1964; Chin et al., 1966) or of quinine, even in the presence of low-level quinine resistance (Peters, 1970). However, faster schizontocidal activity and improved clinical response was seen when sulfadoxine was combined with pyrimethamine (Richards, 1966; Harinasuta et al., 1967; Laing, 1968a, 1970a) or cycloguanil (in mice Thompson et al., 1965). When multidrug-resistant infections were noted with increased frequency in Southeast Asia, in the mid to late 1960s, sulfadoxine-pyrimethamine was a logical first-line drug replacement for chloroquine in Thailand.
Despite these largely successful field trials with sulfadoxine-pyrimethamine, evidence that the combination might prove to be an ineffective long-term solution was building in reports of clinical failure from Southeast Asia and South America in semi-immune persons (Bunnag et al., 1980) and from Africa and the United States in nonimmune persons (Chin et al., 1967; Spencer, 1985; Miller et al., 1986). An early paper on the efficacy of sulfadoxine-pyrimethamine wisely cautioned that the combination of pyrimethamine and sulfadoxine, whose dose activity regression lines are nearly flat, might lead to rapid development of parasite drug resistance (see Jacobs et al., 1963; Harinasuta et al., 1967).
7. Aryl Amino Alcohols.
Mefloquine and halofantrine, both aryl amino alcohol derivatives of quinine, were developed by the U.S. Army soon after the introduction into clinical practice of sulfa drug-DHFR inhibitor combinations. Because the aryl amino alcohols were introduced into areas where quinine resistance already existed, such as Southeast Asia, and cross-resistance between quinine and the aryl amino alcohols may exist (Peters, 1984), it is perhaps not surprising that resistance to these compounds developed quickly (Nosten et al., 1987, 1991; ter Kuile et al., 1992; Smithuis et al., 1993). The cost of these drugs has prohibited them from being used widely in sub-Saharan Africa.
8. General Concepts Learned from Early Experiences.
Important concepts of drug resistance in Plasmodium spp. followed from these early observations. Covell et al. (1949) described a proguanil-resistant isolate from Nigeria and suggested that resistance, as seen in the Lagos and Malayan strains, was an acquired trait. Clyde and Shute (1957) found incomplete cross-resistance between pyrimethamine and proguanil, and Peters (1975) found the same between the sulfonamides and sulfones. Multidrug resistance was described by Earle et al. (1948) in a Central American strain resistant to proguanil, mepacrine, and quinine. These findings largely have been proved accurate. Interestingly, Peters (1987) opined that the early success with proguanil and pyrimethamine followed by the rapid development of resistance to these agents, along with sharply increasing drug development costs, were primary reasons that drug companies and international agencies failed to continue a concerted effort of antimalarial drug development.
B. Antifolates and Nonfalciparum Malaria
The DHFR and DHPS inhibitors are inherently less active against P. vivax, P. malariae, and P. ovale than against P. falciparum (Coggeshall et al., 1941; Earle et al., 1948; Laing, 1968b). For example, the Chesson strain of P. vivax was not inhibited by 1 g of dapsone daily for 10 days and both sulfadoxine and sulfalene were incapable of effecting a radical cure of the same strain (Martin and Arnold, 1969). A field trial in Malaysia also found poor activity of both sulfonamides alone and in combination with pyrimethamine against P. vivax (Laing, 1968a). Likewise, DHFR/DHPS inhibitors are less active against both P. malariae and P. ovale than is chloroquine. The antifolates have had mixed clinical success in these latter species (Archibald, 1951; Young, 1957; Hurly, 1959; Clyde, 1967b; Laing, 1968b; Michel, 1968; quoted by Scholer et al., 1984). P. falciparum is inherently more sensitive to the effects of DHFR/DHPS inhibitors, has well described molecular markers for drug resistance, and will therefore be the focus of the rest of this review.
C. Drug Effects on Parasite Stages
Antimalarials have varying effects on the different stages of the malaria parasite's life cycle (Terzian, 1970). The antifolates, quinine and mefloquine, all exert little or no effect on the parasites during the first 24 h of their life cycle (Dieckmann and Jung, 1986a; Rieckmann et al., 1987; Watkins et al., 1993) and appear to affect only the actively dividing forms of Plasmodium spp. (schizonts) (Jones et al., 1948; McGregor and Smith, 1952; Gutteridge and Trigg, 1971). The DHFR and DHPS inhibitors inhibit DNA synthesis, and their toxic effect on the parasite reaches a peak in the late erythrocytic schizont stage, precisely when DNA synthesis peaks (Hyde, 1990). Parasites treated with antifolates will continue to mature, cytoadhere (attach to vascular endothelium and/or other red blood cells), and develop into gametocytes following treatment. Any early decline in peripheral parasitemia following administration of these drugs is that which would have occurred in the absence of drug treatment (McGregor and Smith, 1952; White, 1997) and is due to cytoadherence and "sequestration". In contrast, chloroquine, artemesinin, and other drugs act on early ring stages (Rieckmann et al., 1987; Geary et al., 1989; Landau et al., 1992; ter Kuile et al., 1992, 1993) and will enhance clearance of parasites shortly after administration, potentially preventing further development of susceptible parasites and worsening of clinical illness (White, 1994; Enosse et al., 2000).
D. Parasite Clearance Following Antimalarial Drug Treatment
Several factors affect the rate of parasite clearance from the peripheral blood after drug treatment, including parasite biomass at initiation of treatment, the degree of parasite life-stage synchronization, the proportion of parasites in a life-stage susceptible to the drug's effects, host immunity, and micronutrient levels, as well as any innate parasite drug resistance. As the parasites mature, they adhere to vascular endothelium and "disappear" from peripheral blood before re-entering the peripheral circulation as merozoites. This re-entry of parasites into the peripheral circulation may be misinterpreted as drug failure, but an early increase in the peripheral parasitemia (<12 h following initiation of therapy) is normal following administration of sulfa drugs. Recovery from an acute falciparum malaria episode following antimalarial treatment is assessed in vivo by parasitologic and clinical parameters (see Tables 1 and 2) (World Health Organization, 1973, 1996, 2002). Parasitologic outcomes are concerned only with the presence or absence of parasites and do not consider clinical signs such as fever. Parasitologic recovery is defined as the clearance of parasites from peripheral blood smears. In infections with highly drug-resistant parasites, parasite density in the peripheral blood may not decline to undetectable levels and may continue to increase following treatment. Parasites with lower levels of resistance are generally cleared completely from the peripheral blood (that is, parasite density levels drop below those detectable by microscopy, about 10–50 parasites/µl), but reappear or recrudesce, at a later time, with or without a return of symptoms. In vivo parasitologic response to treatment traditionally has been measured with a threetiered grading scheme, RI–RIII, as outlined in Table 1. Treatment efficacy can also be assessed using definitions that consider clinical as well as parasitologic responses to treatment (see Table 2) (Rieckmann, 1990; in vivo validation, see Plowe et al., 2001). Therapeutic efficacy, or adequate clinical and parasitological response, is characterized by an early reduction of parasite density and lack of fever or signs of severe malaria and lack of recurrent parasitemia. These parasitologic and clinical in vivo methods of measuring parasite resistance tend to overestimate high-level parasite resistance and early treatment failures (Plowe et al., 2001) and are influenced by host and pharmacokinetic factors independent of parasite drug resistance.
|
Clyde wrote "a clear distinction must be made between drug resistance and drug failure because the latter may imitate resistance, thus complicating unnecessarily the choice of eradication or control operations" (Clyde, 1972b). He further suggested a difference between parasitologic resistance and drug failure, the latter depending upon pharmacodynamic and other factors in addition to parasite resistance (Clyde, 1972a). Parasitologic resistance in this context was defined as the ability of a parasite strain to multiply or to survive in the presence of drug concentrations that normally destroy parasites of the same species or prevent their multiplication (World Health Organization, 1963). Clyde also discussed the term "relative" resistance, wherein he defined two kinds of relative parasitologic resistance: "real" parasite factors (see also McNamara et al., 1967) and "apparent" host factors, such as immunity, with treatment of nonimmunes versus semi-immunes providing an example (Clyde, 1972b). Even recent discussions of parasitologic resistance are complicated by interpretation. A recent review noted that "the term `chloroquine resistance' can lead to some misunderstandings when it is taken by some to refer to in vitro phenotypes, by others to refer to the ability of malaria parasites to survive chloroquine at therapeutic serum concentrations in vivo, and yet by others to refer to the outcome of a clinical episode after chloroquine therapy" (Wellems and Plowe, 2001).
III. Dihydrofolate Reductase-Thymidylate Synthase
There is significant homology between P. falciparum DHFR and other species' DHFR, despite only 24 to 42% sequence homology (Rastelli et al., 2000; Yuthavong, 2002). P. falciparum dihydrofolate reductase-thymidylate synthase (DHFR-TS) is encoded by a single-copy gene on P. falciparum chromosome four, with the two enzymes forming a bifunctional protein (Bzik et al., 1987) similar to other protozoans but distinct from bacteria and higher order eukaryotes. The DHFR-TS of P. falciparum contains 608 amino acids, the first 231 comprising the DHFR domain, the next 89 residues forming the junction region, which joins the remaining 288 residues of the thymidylate synthase domain. Dihydrofolate reductase is comprised of eight central 
-strands between four 
-helices, with an additional three short -helices (Yuvaniyama et al., 2003). As noted earlier, DHFR-TS provides reduced folate for use in the thymidylate cycle, and inhibition of DHFR-TS results in arrested DNA synthesis secondary to reduced levels of dTMP (Ferone, 1977).
A. Dihydrofolate Reductase Inhibitors
The potency of the different DHFR inhibitors varies widely. WR99210 (the active metabolite of PS-15) is the most potent plasmodial DHFR inhibitor identified thus far, whereas cycloguanil (DHFR-inhibiting active metabolite of proguanil) and chlorcycloguanil (DHFR-inhibiting active metabolite of chlorproguanil) (Hawking, 1947; Carrington et al., 1954) are more potent than pyrimethamine (Ferone et al., 1969; Milhous et al., 1985; Winstanley et al., 1995; Sirawaraporn et al., 1997a; Nzila-Mounda et al., 1998). Trimethoprim is the least potent of the antimalarial DHFR inhibitors (Ferone et al., 1969; Iyer et al., 2001).
1. Cross-Resistance between Dihydrofolate Reductase Inhibitors.
Several studies have found a lack of or incomplete in vitro cross-resistance between pyrimethamine and cycloguanil in P. falciparum (Milhous et al., 1985; Winstanley et al., 1995), P. gallinaceum (Rollo, 1952a), and P. berghei (Thompson and Bayles, 1968), and in vivo findings were suggestive of the same (Clyde, 1967a; Vestergaard Olsen, 1983). Other studies showed that the degree of cross-resistance in resistant clones varied (Robertson et al., 1952; Jones, 1953) and that cycloguanil-induced resistance was "broader" than that induced by pyrimethamine (Thompson and Bayles, 1968), although a significant, but incomplete, in vitro (Petersen, 1987) and in vivo (Martin and Arnold, 1968b) cross-resistance between pyrimethamine and trimethoprim existed. These studies suggested a related but incompletely shared mechanism of resistance.
B. Identification of Antifolate Drug Target
In line with the earlier in vivo prophylaxis studies by Clyde and Shute in the 1950s, selection for an in vitro drug-resistant phenotype was shown to occur following administration of pyrimethamine for treatment of acute malaria. For example, 4 days after pyrimethamine treatment of a mixture of sensitive and resistant parasites, only parasites with a drug-resistant phenotype remained (Nguyen-Dinh et al., 1982). Separate investigations revealed that pyrimethamine-resistant and -sensitive isolates had identical uptake of pyrimethamine, but that the DHFR activity of resistant strains was 300 times less sensitive to the inhibitory effects of pyrimethamine (Dieckmann and Jung, 1986b). These early in vivo and later in vitro studies showing rapid development of resistance in response to drug pressure suggested a relatively simple mechanism of resistance, such as individual point mutations in a single gene. Evidence of a genetic basis for antifolate resistance first arose from a genetic crossing study in which pyrimethamine-sensitive and -resistant parasite lines were crossed in the mosquito vector (Walliker et al., 1975). The drug resistance phenotype segregated independently of the other enzymatic markers, but in a similar manner, demonstrating that recombination had occurred between the original parental lines.
C. Point Mutations within Dihydrofolate Reductase Are Responsible for in Vitro Resistance
Continued investigation of the dhfr gene with the eventual sequencing of both phenotypically sensitive and phenotypically resistant P. falciparum dhfr provided the first direct evidence that point mutations within the dhfr gene were responsible for the resistant phenotype. Dihydrofolate reductase derived from the pyrimethamine-sensitive clone, 3D7, and from isolates with varying degrees of resistance to pyrimethamine were sequenced (Cowman et al., 1988). A serine resided at position 108 in the sensitive 3D7 clone, but there was a change to asparagine (S108N) in the resistant isolates. Other, successively more resistant isolates demonstrated additional mutations at codons 51 (N51I), 59 (C59R), and 164 (I164L). It was believed that the conservative change of isoleucine to leucine at codon 164 would not have a profound effect on pyrimethamine binding. The Palo Alto "clone" carried a unique set of mutations, A16V plus S108T. A study published at the same time by another group similarly found that the addition of DHFR N51I and C59R mutations confer greater levels of pyrimethamine resistance than does S108N alone (Peterson et al., 1988), strengthening the evidence that point mutations in dhfr were the cause of pyrimethamine resistance. Once again, the allele containing A16V and S108T was found in only one isolate (FCR3). This second group proposed that these mutations arose independently, i.e., that pyrimethamine resistance was an acquired trait, a supposition first put forth by Covell in 1949 (Covell et al., 1949).
The 50% inhibitory concentrations (IC50) of pyrimethamine and cycloguanil of 10 isolates from around the world, each with differing degrees of resistance to DHFR inhibitors, were compared, and in general, resistance to these drugs rose in parallel, although there were exceptions (Foote et al., 1990). In particular, the A16V/S108T allele yielded much higher IC50 values for cycloguanil than for pyrimethamine and was the only allele to do so (Foote et al., 1990). All other mutations contained at a minimum S108N and yielded higher IC50 for pyrimethamine than cycloguanil with the exception of S108N/C59R/I164L, which showed similarly high levels of resistance to both drugs (Peterson et al., 1990). These findings suggested that S108N was an essential first mutation in DHFR and that additional mutations at codons 51 and/or 59 and 164 increased the IC50 for pyrimethamine. This finding was supported by further in vitro resistance tests of field isolates and sequencing of their respective dhfr, which confirmed that greater in vitro pyrimethamine resistance correlated with a greater number of DHFR mutations (Basco et al., 1995; Nzila-Mounda et al., 1998). For example, the single mutation S108N caused a 25-fold greater IC50 of pyrimethamine than wild-type DHFR. Double mutations did not cause significant further increases in IC50, but the triple DHFR mutant (S108N/N51I/C59R) was 225- fold more resistant to pyrimethamine and 48-fold more resistant to cycloguanil than wild-type DHFR. These and subsequent genetic transformation studies (van Dijk et al., 1995; Wu et al., 1995, 1996; Crabb and Cowman, 1996;) added further strength to the concepts that 1) point mutations in DHFR are responsible for in vitro resistance, 2) greater numbers of mutations in DHFR leads to greater drug resistance, 3) cross-resistance between pyrimethamine, cycloguanil, and proguanil is incomplete, and 4) S108N is a necessary first mutation in DHFR (except the case of the rare A16V/S108T allele).
Hitherto, unknown mutations at DHFR codon 50 (C50R) and a 15-base pair repeat inserted between codons 30 and 31, termed the Bolivia repeat, were described in samples from Latin America (Plowe et al., 1997). Genetic transformation studies in yeast suggested that the C50R mutation plays a role similar to the African C59R mutation (Cortese and Plowe, 1998), and these two mutations appear to be mutually exclusive (Plowe et al., 1997; Cortese et al., 2002). The Bolivia repeat was subsequently not found to play a role in resistance (Cortese and Plowe, 1998) and may instead compensate for the decreased DHFR enzyme function that accompanies the I164L mutation in South American isolates.
Only one study has reported isolates with wild-type serine-108 in association with other DHFR mutations. This study from Uganda found the serine-108/N51I/C59R mutation in all "resistant" infections following treatment of uncomplicated falciparum malaria with trimethoprim-sulfamethoxazole (Jelinek et al., 1999), suggesting that serine-108 is important for trimethoprim resistance and that trimethoprim and pyrimethamine have different molecular mechanisms of resistance. Homology modeling predicts that trimethoprim would be less affected than pyrimethamine by mutation at DHFR codon 108 (Warhurst, 2002), but is unclear on the effect on trimethoprim of further mutations at 51 and 59 in combination with S108N or S108. The unusual finding by Jelinek, which could have implications for malaria control and for trimethoprim-sulfamethoxazole prophylaxis against opportunistic infections in human immunodeficiency virus-infected persons in Africa where sulfadoxine-pyrimethamine is used, was not supported by later in vitro experiments using recombinant yeast expressing the novel serine-108/N51I/C59R allele (Iyer et al., 2001) and P. falciparum isolates with known dhfr mutations (Khalil et al., 2003). In these studies, as had been previously reported (Winstanley et al., 1995), there was significant in vitro cross-resistance between pyrimethamine and trimethoprim. Alleles that included the S108N mutation were more resistant to both trimethoprim and pyrimethamine. Specifically, the triple mutant S108N/N51I/C59R was more resistant to trimethoprim than the genetically engineered serine-108/N51I/C59R mutant reported to have been selected by drug treatment with trimethoprim-sulfamethoxazole in the Uganda study (Iyer et al., 2001). Again, in contrast to the Uganda study, Khalil reported that this triple mutant allele predominated in recrudescent infections after treatment with trimethoprim-sulfamethoxazole (Khalil et al., 2003). The serine-108/N51I/C59R genotype has not been found in nature by other groups nor has it been confirmed by DNA sequencing. At this time, the weight of the evidence supports the idea that trimethoprim does induce the DHFR S108N mutation and that this mutation confers resistance to trimethoprim as it does to pyrimethamine.
Gene amplification as a source of folate resistance has not been demonstrated in nature and appears to play no role in resistance. A single in vitro study did find gene amplification to be the only method of pyrimethamine resistance to develop after 22 to 46 weeks of cultivation with increasing concentrations of pyrimethamine (Thaithong et al., 2001). Two other studies showed increased amounts of DHFR enzyme in resistant isolates, which suggested gene amplification (Kan and Siddiqui, 1979; Inselburg et al., 1987). Yet, analysis of pyrimethamine-resistant field isolates from around the world has consistently failed to demonstrate any evidence of DNA fragment or extrachromosomal amplification (McCutchan et al., 1984; Chen et al., 1987; Cowman et al., 1988; Peterson et al., 1988; Foote et al., 1990). Human thymidylate synthase expression can be regulated via binding to its own mRNA (Lin et al., 2000). Translational regulation of DHFR-TS occurs similarly in Plasmodium spp., although unlike regulation in humans, the binding of mRNA does not occur at the active site of the enzyme. Therefore, binding of inhibitor or substrate does not free mRNA for translation (Zhang and Rathod, 2002).
F. Mutation Rates within the Dihydrofolate Reductase Gene
The determination of mutation rates of genes whose products are targets of antimalarial drugs is important in the testing of new antimalarial compounds or combinations. Mathematical modeling may be applied to anticipate the time course of the rise in drug resistance and to predict the spread of drug resistance within a population. Numerous studies have demonstrated the development of antifolate drug resistance in both rodent malarias (Bishop and Birkett, 1947; Williamson et al., 1947; Ramakrishnan et al., 1961; Bishop, 1962; Martin and Arnold, 1968a) and P. falciparum (Gassis and Rathod, 1996; Paget-McNicol and Saul, 2001), and together they suggest a spontaneous mutational rate of nuclear genes, such as dhfr, on the order of 10-9/parasite/cycle. Support for such low mutational rates was indirectly obtained in experiments using fewer than 109 parasites, which failed to detect any drug-resistant parasites or mutations (Bishop, 1958; Rathod et al., 1997). Furthermore, other studies achieved more rapid development of resistance when strong drug pressure was applied to animals with heavy parasitemias (Rollo, 1952b; Clyde and Shute, 1954; Diggens et al., 1970) compared with suppressive dosing of animals with lowgrade parasitemias. Another in vitro study induced the DHFR S108N mutation in a pyrimethamine-sensitive parasite line by applying drug pressure (Paget-McNicol and Saul, 2001) and estimated the mutation rate of dhfr mutation to be <2.5 x 10-9 mutations/dhfr gene/replication.
There is also evidence to suggest that different parasite isolates have differing mutational capabilities. For example, the most "mutagenic" parasite line of five tested, W2 from Southeast Asia, developed mutations at a rate of 10-6, which was at least 100 times greater than any of the other clones tested (Rathod et al., 1997), and older in vivo data with P. gallinaceum suggests variation of mutational capabilities between different parasite isolates (Bishop, 1962). Studies of the cytochrome b gene of P. falciparum also demonstrated a low mutational rate of 2 x 10-9 mutations/parasite/cycle (Avise, 1991), and field studies of msp1 and ama1 genes of P. vivax (Figtree et al., 2000) are all in accordance with the mutational rates found for the dhfr gene of P. falciparum.
Many more than 109 parasites would be present in symptomatic infections in semi-immune individuals (108–1012; White et al., 1999), but even so, the rapid induction of resistance in natural populations (Clyde and Shute, 1957; Nguyen-Dinh et al., 1982) may be due, in part, to selection of small numbers of parasites with pre-existing mutations (Bishop, 1962; Martin and Arnold, 1968a; Nguyen-Dinh et al., 1982; Spencer et al., 1984; Kun et al., 1999; Wootton et al., 2002; Roper et al., 2003, 2004). Mathematical models validated with data from field studies may help to determine whether selection for pre-existing mutations or spontaneous point mutations within the parasite population of the host play a more important role in antimalarial drug resistance.
G. Enzyme Kinetic Analysis of Dihydrofolate Reductase
Enzyme kinetic studies can help explain why certain mutations or allelic patterns are more prevalent than others. The first kinetic studies used clones of pyrimethamine-resistant isolates rather than recombinant DHFR. Such resistant parasite lines had similar substrate affinity (Km), as did sensitive clones, but significantly increased Ki for pyrimethamine, in one case 300- fold greater than a sensitive isolate (McCutchan et al., 1984; Dieckmann and Jung, 1986b; Chen et al., 1987). Purified DHFR enzyme from highly resistant (7G8) and moderately resistant (HB3) parasite lines had 500- and 15-fold less affinity, respectively, for pyrimethamine than did the enzyme from the sensitive parasite line (3D7) (Zolg et al., 1989). Recombinant P. falciparum DHFR enzyme has since been expressed in Escherichia coli for enzyme kinetic studies (Sirawaraporn et al., 1990, 1993, 1997a; Sano et al., 1994; Hekmat-Nejad et al., 1997). These elegant studies provide a rationale for the specific sequence of mutation accumulation and commonly occurring alleles, based on the combined effects on drug resistance and enzyme kinetic function. Serine-108-Asn (S108N) alone confers moderate pyrimethamine and cycloguanil resistance (approximate 10-fold increase in Ki compared with wild type) at minimal catalytic/kinetic cost to the enzyme as measured by substrate turnover (Kcat/Km) and substrate affinity (Km). Subsequent mutation at codon 51 (N51I) results in similar Ki for pyrimethamine and cycloguanil as the single-mutated (S108N) DHFR, but an improved kcat/km on par with wild-type enzyme (Sirawaraporn et al., 1997a). Cysteine-59-Arg (C59R), in combination with the aforementioned double mutation, forms the so-called "triple mutant" (S108N/N51I/C59R). The triple mutant confers a 100-fold increase in the Ki of pyrimethamine and cycloguanil, but incurs a significant cost to kinetic activity of the enzyme, demonstrating kcat/km values 50-fold lower than the wild-type enzyme. The addition of I164L to the triple mutant, forming the "quadruple mutant", increases pyrimethamine and cycloguanil Ki 500-fold over wild-type DHFR, without further significant decline in enzymatic function compared with the triple mutant.
Interestingly, the mutation combination associated with cycloguanil rather than pyrimethamine use, A16V/S108T, had near wild-type pyrimethamine Ki, but an 800-fold increase in cycloguanil Ki. The kinetic activity of the enzyme was severely impaired and unlike all other mutation combinations studied, the free energies for pyrimethamine and cycloguanil binding were significantly higher than the sum of the corresponding single mutations. The A16V mutation alone entirely explains cycloguanil resistance, but the addition of S108T markedly improves kinetic parameters (Sirawaraporn et al., 1997a). An allele not seen in nature, A16V/S108N, was found to lack any catalytic activity, explaining its natural absence.
Each mutation, with the exception just noted, is more than additive in its effect on resistance, with over 1000-fold synergism in the quadruple mutant, as measured by interaction energy (Sirawaraporn et al., 1997a). One can conclude that mutations evolve not solely for their ability to cause resistance, but for their synergistic contribution to enzyme function in the context of existing mutations, suggesting a necessary sequential progression of mutations. The mutations seen only in combination with S108N: N51I, C50R/C59R, and I164L do not confer resistance to pyrimethamine or cycloguanil on their own; to do so, S108N must be present. The cost of greater reductions in kcat/km values for dihydrofolate and NADPH suggests that the more highly mutated forms of DHFR might be selected against in the absence of sufficient drug pressure.
In an attempt to determine why it is that S108N is the preferred primary mutation, 10 different amino acid substitutions were used in place of asparagine at position 108 (Sirawaraporn et al., 1997b). Only asparagine provided an enzyme with near wild-type kinetics (<2- fold lower Kcat/Km). Despite other substitutions providing a higher level of antifolate resistance, the asparagine substitution provides the enzyme kinetics necessary for parasite survival with the least number of base-pair substitutions.
Further kinetic studies show that S108N/C59R is favored in terms of pyrimethamine resistance with a 5-fold increase over wild type, but it has significantly impaired kinetic function. The double mutation S108N/N51I has a 2- to 3-fold increase in pyrimethamine resistance and kinetic parameters closer to wild type. Kinetic studies alone cannot explain why certain alleles are more common in nature (Sirawaraporn et al., 2002). For example, the quadruple mutant is not found in Africa, but it demonstrates markedly increased resistance with only minimal loss in kcat over the triple mutant, which is found in Africa. However, it must be noted that these enzyme kinetic studies in recombinant DHFR monomers that are not linked to TS may not reflect exactly the kinetics of the bifunctional DHFR-TS dimeric protein of parasites in nature. This leaves open the possibility that the DHFR I164L mutation is more deleterious to parasite fitness in nature than these experiments predict, which could explain its absence in Africa, where the large reservoir of asymptomatic, untreated infections favors survival of more fit, less resistant parasites.
In summary, it appears that individual point mutations are favored for the best Ki to Kcat/Km benefit. The ideal combination of DHFR mutations in any given population is influenced by both the local dynamics of drug pressure particularities and any competition among circulating parasite "clones". It might be difficult for parasites harboring newer drug-resistant alleles to successfully invade and establish themselves in a parasite population, particularly if previously established, circulating drug-resistant alleles offer adequate protection to circulating parasites from the antifolate drugs in use, as may be the case with the triple mutant in Africa.
H. Relationship of Point Mutations to Dihydrofolate Reductase Structure—Crystallography
Prior to the crystallization of DHFR-TS, several attempts were made to model the effect of known DHFR mutations on drug binding (McKie et al., 1998; Santos-Filho et al., 2001). With the recent crystallization of the DHFR-TS enzyme of P. falciparum, it is now apparent that many of the predictions based upon the models were correct. The rigid length of the pteridine ring of many folate inhibitors fits between residues 108 and 54 within the active site of the enzyme (Warhurst, 1998; Yuvaniyama et al., 2003) (Fig. 2). Pyrimethamine is relatively rigid, notably more so than trimethoprim, which therefore allows less torsional freedom within the active site. The longer distances between residues 108 and 54 and trimethoprim reduce the inhibitory activity of trimethoprim against P. falciparum DHFR (Warhurst, 1998, 2002). Of all the known DHFR inhibitors, WR99210 most closely resembles the flexibility seen in the natural substrate dihydrofolate (DHF), perhaps explaining its greater potency and reduced susceptibility to point mutations in DHFR. Crystallization of WR99210 within the binding site of DHFR has confirmed this prediction (Yuvaniyama et al., 2003) (Fig. 3). Point mutations disrupting WR99210 binding may also disrupt substrate binding significantly enough to be detrimental to parasite survival.
|
|
As predicted in the model presented by Delfino et al. (2002), the primary DHFR mutation, S108N, affects drug accommodation at the active site of the enzyme, rather than affecting drug entry into this site. The NH2 of asparagine 108 interferes with the chlorine atom of pyrimethamine, causing the later to be displaced within the active site (Warhurst, 1998, 2002). Further compromise of the planar aspect of the pyrimethamine pteridine ring is caused by the carboxylate group of asparagine 54 and the pyrimethamine ethyl group. This loss of planarity is theorized to be responsible for the pyrimethamine resistance caused by the S108N mutation (Delfino et al., 2002). There is less hindrance of cycloguanil planarity, being slightly less rigid than pyrimethamine, although the chloride and NH2 group of asparagine 108 do interfere to some degree. Modeling and crystallization images of WR99210 in DHFR active site show that there is no hindrance of this molecule in the enzyme's active site because the flexible side chain is oriented away from the S108N side chain (compare Figs. 2 and 3). This correlates well with in vitro data that demonstrates a higher level of resistance to pyrimethamine than to cycloguanil and essentially no effect on WR99210 in the presence of the S108N mutation alone or in combination with N51I (S108N/N51I) (Ferlan et al., 2001).
Likewise, A16V mutation does not displace pyrimethamine from the active site, but causes significant loss of cycloguanil planarity due to interactions between methyl groups on both valine 16 and cycloguanil (Rastelli et al., 2000; Delfino et al., 2002) and may interfere with NADPH binding (Yuvaniyama et al., 2003). The additional mutations N51I and C59R are distant from the enzyme active site, but on the same helix as residue 54, which is critical to substrate binding (Yuvaniyama et al., 2003). The side chain of C59R extends away from the inhibitor binding site and may be involved in DHF binding, improving substrate binding affinity in the presence of S108N and N51I. Mutations at residues 51 and 59 may also impede admission of the inhibitor binding site (Santos-Filho et al., 2001; Warhurst, 2002). This could occur via interactions of polar or charged residues on isoleucine and arginine and the protonated pyrimethamine and cycloguanil (Delfino et al., 2002). The effect of the additional mutations at 51 and 59 must be essentially silent in the absence of the S108N mutation, but additive when it is present (Sirawaraporn et al., 1997a), and this is supported by the crystal structure, which demonstrates little change in the orientation of D54 in the presence of N51I and C59R. In agreement with kinetic data (Sirawaraporn et al., 1997a), I164L incorporated into the model in the absence of other mutations predicts a decrease in pyrimethamine resistance, a mild increase in cycloguanil resistance, and a marked increase in resistance to both drugs in the presence of S108N, N51I, and C59R. Crystal data suggests that this is due to an increase in the active site gap between the -carbon of C50 and I164L.
IV. Pyrophosphokinase-Dihydropteroate Synthase
The role of DHPS in sulfonamide resistance was elucidated following the same pattern as for DHFR, with the cloning of the gene, the identification of point mutations associated with in vitro drug resistance, and detailed characterization of the heterologously expressed wild-type and mutant enzymes. Like DHFR, DHPS is a bifunctional polypeptide (in contrast to bacterial DHPS) (Hyde and Sims, 2001) encoded by a gene also encoding 7,8-dihydro-6-hydroxymethylpterin pyrophosphokinase, the enzyme immediately proceeding DHPS in the folate biosynthesis pathway (Brooks et al., 1994; Triglia and Cowman, 1994). DHPS catalyzes the condensation of p-aminobenzoic acid with 6-hydroxymethyldihydropterin pyrophosphate yielding 7,8-dihydropteroate. As with antagonism of DHFR, antagonism of DHPS in Plasmodium spp. results in depletion of the dTTP precursor and subsequent decrease in DNA production (Schellenberg and Coatney, 1961). Unlike DHFR, there is no known human counterpart to DHPS.
Sulfonamides and sulfones, such as sulfadoxine and dapsone, can inhibit DHPS activity, but acceptance of this paradigm was delayed because of the ability to utilize exogenous folate by many, but not all, P. falciparum isolates (Trager, 1958; Ferone, 1977; Krungkrai et al., 1989; Wang et al., 1997b,c, 1999). This salvage pathway is believed to provide only a minority of folate production in P. falciparum, the majority being procured via de novo synthesis. The capability of Plasmodium spp. to utilize exogenous folate was demonstrated in experiments in which the addition of physiologic concentrations of folate to culture medium caused a 1000-fold decrease in sulfadoxine inhibitory activity, and higher, nonphysiologic concentrations of folate were able to partially inhibit the activity of pyrimethamine (Chulay et al., 1984). Similarly, three of four P. falciparum strains tested were capable of growing in culture media devoid of folic acid or p-ABA (Milhous et al., 1985). The fact that folic acid was a more potent antagonist of sulfadoxine activity than p-ABA provided further support for exogenous folate utilization via a pathway that obviated the need for DHPS (i.e., bypassed the enzyme) (Wang et al., 1999). Other groups have subsequently generated ample evidence for utilization of exogenous folate in P. falciparum (Schapira et al., 1986; Krungkrai et al., 1989; Wang et al., 1997b). The folate effect is abolished by the addition of pyrimethamine to assay cultures in concentrations substantially higher than those needed to inhibit wild-type DHFR (Wang et al., 1999). Moreover, even parasites with highly mutated, pyrimethamine-resistant DHFR are unable to metabolize exogenous folate in the presence of pyrimethamine, suggesting that the folate effect is a DHFR-independent pathway, but one which may be inhibited by high concentrations of pyrimethamine. The exact mechanism or genetic basis of the "folate effect" is not known at this time, nor is the prevalence of this capacity in natural parasite populations, although it has been found in parasite lines from both Africa and Southeast Asia. The folate effect did not segregate with dhps genotypes in a genetic cross (Wang et al., 1997b), and although it was linked to the dhfr gene in the cross progeny, it was not linked with dhfr sequence in other unrelated parasite lines (Wang et al., 1997b), suggesting that a gene responsible for the folate effect is located near, but not at the dhfr locus on chromosome four.
1. Folate Effect and Drug Resistance.
The folate effect is thought to be of potential importance in resistance to sulfonamide drugs and sulfa-DHFR inhibitor combinations. The concentration of pyrimethamine necessary to inhibit wild-type DHFR is <20 nM, which is far below that needed to completely abolish the folate effect (
100 nM). When e
