INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The protozoan parasite Toxoplasma gondii is a ubiquitous human pathogen, but a robust cellular immune response ordinarily suppresses acute infection by the rapidly dividing tachyzoite stage of the parasite life cycle (9). As a consequence, primary T. gondii infections are generally self-limited and do not require therapeutic intervention, barring the special case of primary infection during pregnancy (50). Once infected, individuals carry a latent, encysted form of the parasite (the bradyzoite) for life. This condition poses a chronic health hazard for patients who are immunocompromised, immunosuppressed, or otherwise unable to respond effectively to the challenge posed by reactivation of latent infection (22). Long-term chemoprophylaxis is generally advised for such individuals (2), even in human immunodeficiency virus-infected patients receiving highly active antiretroviral therapy (although recent studies indicate that the duration of treatment for these patients can be shortened considerably) (34, 43).

Antifolates form the basis of chemotherapy against both primary and secondary (recurrent) infection by T. gondii (18, 19, 50; S. Y. Wong and J. S. Remington, Editorial, AIDS 7:299-316, 1993), either pyrimethamine combined with sulfadiazine or the somewhat less potent combination of pyrimethamine and clindamycin (5, 17). Both strategies are effective at disease remediation with short-duration use (10), but the frequency of treatment failures associated with long-term antifolate use is high, particularly among AIDS patients (2, 16, 42, 47; P. Caramello, T. Brancale, B. Forno, A. Lucchini, A. M. Pollono, A. Ullio, P. Gioannini, I. Viano, and E. Tonso, Letter, Antimicrob. Agents Chemother. 39:2371-2372, 1995). In one study, relapse rates over a 2-year period were 11% for pyrimethamine-sulfadiazine and 22% for pyrimethamine-clindamycin (17). Many determinants may contribute to treatment failures, including host factors, such as intolerance to either or both component drugs (22) or malabsorption into infected neural tissues (16, 49). Parasite-specific factors, such as preexisting differences in drug sensitivity between strains or the development of drug resistance mutations during chemotherapy, may also play an important role. Prolonged drug treatmentparticularly with drugs such as pyrimethamine that exhibit a long half-life (16, 26)can elicit a strong selection for drug-resistant pathogens, as has been seen with Mycobacterium tuberculosis (20, 21), human immunodeficiency virus (32, 33), and the malaria parasite Plasmodium falciparum (15). This report examines both the influence of parasite genotype on drug sensitivity and the potential for acquired resistance in disease relapse.

The target for pyrimethamine in T. gondii is the dihydrofolate reductase (DHFR) domain of a bifunctional fusion protein that also harbors thymidylate synthase (TS) activity (31). The primary sequence for the single-copy T. gondii dhfr-ts gene has been determined for strain RH (35). Restriction fragment length polymorphism and isozyme analysis indicate that most T. gondii isolates fall into one of three clonal lineages (designated type I, II, or III) (6, 12). While correlations among the parasite genotype, virulence, and disease presentation have been established (13, 44), less is known about the role of parasite genotypes in cases of failed drug therapy. The first part of this study considers intrinsic differences in pyrimethamine sensitivity between strains of T. gondii and evaluates whether the observed variations in drug susceptibility can be attributed to resident dhfr polymorphisms.

In P. falciparum (a member of the same phylum as T. gondii), resistance to pyrimethamine is commonly associated with dhfr point mutations (4, 29, 45; reviewed in reference 15). To better understand the potential role of dhfr polymorphisms in toxoplasmosis, we developed a method to generate and select pyrimethamine resistance mutations in T. gondii dhfr. Random mutagenesis of the dhfr-ts gene was followed by in vitro drug selection during intracellular replication of T. gondii parasites. Several pyrimethamine-sensitive T. gondii dhfr alleles were mutagenized and subsequently screened for drug resistance in either T. gondii parasites or yeast, yielding novel resistance mutations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Pyrimethamine sensitivities of natural T. gondii strains. Three strains representing the three major virulence groups of T. gondii (12) were assessed for pyrimethamine sensitivity: the type I strain RH (38), the type II strain P(LK) (23), and the type III strain Veg (28). Each strain was allowed to infect confluent host cell monolayers of primary human foreskin fibroblasts grown in modified essential medium (Gibco) containing 0.1% dialyzed fetal bovine serum, as previously described (37). Pyrimethamine was added at various concentrations, and parasite proliferation was measured by following [³H]uracil incorporation in a 4-h uptake assay carried out ~24 h postinfection (30).

Mutagenesis of T. gondii dhfr alleles. Two constructs were used as templates for mutagenesis experiments (Fig. 1). In pC3 vectors, cDNA-derived (intronless) T. gondii dhfr-ts minigenes flanked by autologous genomic flanking sequences (8) were introduced into the pBluescript KS() plasmid (Stratagene). Two pyrimethamine-sensitive dhfr-ts alleles were employed in pC3 vectors: the wild type and allele T83S, which carries a ThrSer substitution at position 83 (31). In pTRP constructs, a 1-kb BamHI to EcoRI fragment encompassing the T. gondii dhfr domain was introduced into the nonintegrative, single-copy Saccharomyces cerevisiae shuttle vector pRS304 (40) under control of the yeast dhfr promoter (51). Both wild-type dhfr and the pyrimethamine-hypersensitive allele A10V (31) were introduced into the pTRP background.

View larger version (31K): [in this window] [in a new window]   FIG. 1.   Scheme for random mutagenesis of T. gondii dhfr and in vivo selection of pyrimethamine-resistant mutants. The plasmid maps for pC3 and pTRP are drawn to scale, indicating the coding sequence for T. gondii dhfr (shaded boxes) and ts (hatched box; pC3 only), flanking untranslated sequences (heavy black lines; the arrows indicate the promoters and the direction of transcription) derived from the T. gondii dhfr-ts genomic locus (7, 8) (pC3) or S. cerevisiae dhfr (51) (pTRP), and pBluescript KS() or pRS304 vector sequences (thin gray lines; miscellaneous features are as marked). YEPD, yeast extract-peptone-dextrose agar.

Selection for pyrimethamine-resistant T. gondii and S. cerevisiae transformants. For direct selection of pyrimethamine-resistant dhfr-ts alleles in transgenic T. gondii, 30 µg of mutated pC3 plasmid DNA (containing mutated wild-type or T83S alleles of dhfr-ts) were transfected into RH strain tachyzoites by electroporation (8). Stably resistant parasites were selected at 0.8 µM pyrimethamine in 25-cm² T flasks containing confluent monolayers of primary human foreskin fibroblasts and cloned by limiting dilution in microtiter plates, as previously described (37). Genomic DNA was obtained from resistant clones by phenol-chloroform extraction for analysis of the dhfr-ts transgene copy number and genomic organization by Southern hybridization (24).

Identification of DHFR polymorphisms and confirmation of resistance phenotypes. The dhfr domain of putative resistance genes was amplified by PCR and sequenced with Sequenase 2.0 (U.S. Biochemicals) to identify polymorphisms. Candidate resistance mutations were then introduced into pC3 plasmids containing various dhfr-ts backgrounds by site-directed mutagenesis (25) and transiently expressed in RH strain parasites under pyrimethamine selection. Parasite proliferation was measured 30 h posttransfection by following the uptake of [³H]uracil (30). IC₅₀s were calculated from percent growth inhibition at 0.3, 0.6, and 1.0 µM pyrimethamine relative to untreated controls, and mutants were assigned to resistance classes based on established categories (31).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Pyrimethamine sensitivity and dhfr sequences in different T. gondii strains. To evaluate whether patterns might be established between treatment failures and the parasite genotype, we examined the susceptibilities of parasites from different virulence lineages to pyrimethamine, as shown in Fig. 2. The three strains evaluated [RH, a type I strain; P(LK), type II; and Veg, type III] represent each of the major parasite lineages (12) and show a marked divergence in drug sensitivity. The type II and type III parasite strains were highly sensitive to pyrimethamine, exhibiting IC₅₀s of ~0.02 µM pyrimethamine. The type I isolate (RH) was significantly less sensitive to pyrimethamine (IC₅₀, ~0.9 µM). This strain is highly proliferative and well adapted to tissue culture (37, 38) but appears to be virtually identical to recent type I clinical isolates (12).

View larger version (55K): [in this window] [in a new window]   FIG. 2.   Pyrimethamine sensitivities of representative isolates from three parasite strains representing the major virulence lineages of T. gondii (12). Inhibition of parasite growth was assessed 24 h postinfection by incorporation of [3H]uracil and is presented as the average of two independent experiments (the error bars indicate the upward extent of the 95% confidence interval for the mean).

Random mutagenesis of T. gondii dhfr. Previous studies have demonstrated that dhfr point mutations suspected to be responsible for pyrimethamine resistance in Plasmodium parasites can confer resistance to transgenic Toxoplasma tachyzoites when introduced in the context of the T. gondii dhfr-ts gene (7, 8, 31). On the other hand, direct selection for pyrimethamine-resistant T. gondii mutants in the laboratory has been difficult, and such parasites have thus far failed to yield dhfr-ts mutations (unpublished results). In order to assess the spectrum of T. gondii dhfr mutations with the potential to confer pyrimethamine resistance, we devised a method to generate and select such mutations, as outlined in Fig. 1.

In vivo selection for pyrimethamine resistance in transgenic T. gondii. Plasmid pC3 inserts randomly in the T. gondii genome, generating stable transformants at high frequency (7, 8). Mutagenized pC3 plasmids from both wild-type and T83S backgrounds were introduced into RH strain parasites by electroporation, and drug-resistant parasites were selected through multiple passages in 0.8 µM pyrimethamine. After three passages under pyrimethamine selection, stably resistant parasite clones were isolated by limiting dilution (37). Southern blot analysis of the population transfected with mutated wild-type dhfr-ts indicated the presence of four unique parasite clones, each carrying a single-copy transgene insertion. Only one unique pyrimethamine-resistant parasite clone was obtained from the population of resistant parasites transfected with the T83S dhfr-ts allele (although one variant of this clone contained additional tandem integrations, presumably of the same transgene).

Selection for pyrimethamine-resistant T. gondii dhfr alleles in yeast. In vitro assays conducted on protein expressed from the A10V variant of T. gondii DFHR have previously shown that this variant exhibits a pyrimethamine-hypersensitive phenotype (31), precluding analysis in parasites expressing wild-type dhfr. As there are no dhfr-deficient strains of T. gondii, a dhfr-deficient (dTMP-dependent) strain of S. cerevisiae (YH5) was employed for this purpose (14, 51). YH5 transformants harboring (nonmutagenized) pTRP plasmids including only the dhfr domain of wild-type T. gondii dhfr-ts were insensitive to pyrimethamine inhibition at drug concentrations as high as 10 µM (the highest concentration that could practically be tested). In contrast, transformants harboring the (nonmutagenized) A10V allele were inhibited by 2 µM pyrimethamine. Similar effects were observed in liquid culture: the IC₅₀s were 2.0 µM for transformants carrying the wild-type allele versus 0.75 µM for A10V transformants.

Expression of mutant dhfr genes in transgenic parasites. To determine the role of the observed mutations in generating resistance, and to assess the levels of resistance associated with each, we engineered mutant dhfr-ts genes for expression in wild-type T. gondii parasites. The mutations Trp 25Arg (W25R), Leu 98Ser (L98S), and Leu 134His (L134H) were reintroduced by site-directed mutagenesis into three T. gondii dhfr-ts backgrounds in the pC3 vector: (i) wild-type (pyrimethamine-sensitive), (ii) a moderately-resistant allele carrying the Thr 83Asn substitution (analogous to the S108N mutation in P. falciparum), and (iii) the pyrimethamine-hypersensitive A10V allele.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study provides a demonstration of intrinsic heterogeneity in pyrimethamine sensitivity among different stains of T. gondii parasites and outlines methods suitable for identifying new resistance mutations in the DHFR-TS target of pyrimethamine. We anticipate that the procedures reported here will also be useful in screening for mutations in other drug targets (e.g., DHPS, the target of sulfonamides). Preliminary observations from this study also provide the basis for considering the nature and evolution of pyrimethamine resistance mechanisms in T. gondii and related pathogens.

We find no evidence to suggest that differential strain sensitivity to pyrimethamine is governed by preexisting polymorphisms in the dhfr-ts gene, but it is interesting to note that significant differences in drug sensitivity were observed among the three parasite strains tested, representing the major T. gondii lineages (6, 12). A representative of type I parasites (which are characterized by extreme virulence in mice) was relatively insensitive to pyrimethamine (IC₅₀, ~0.9 µM). In contrast, type II parasites (which account for most human congenital and AIDS-related infections) and type III parasites (common in veterinary infections) displayed IC₅₀s of ~0.02 µM. These concentrations all fall below the ~2-µM serum levels for pyrimethamine typically achieved during low-dosage therapy (100 mg/week) (18), but drug accumulation in the cerebrospinal fluid can be as low as ~0.4 to 1.8 µM, even with doses of up to 175 mg/week (48). At the lower end of this range, treatment for toxoplasmic encephalitis might be problematic for parasites exhibiting drug sensitivity comparable to that of the RH strain (type I). A more thorough sampling of strains to assess the correlation between drug sensitivity and parasite genotype is clearly warranted.

Using in vitro mutagenesis, we identified three T. gondii dhfr-ts mutations capable of producing pyrimethamine resistance (W25R, L98S, and L134H). These mutations are distinct from the canonical mutations found in drug-resistant malaria, perhaps due to the small number of mutants isolated (there is no reason to suspect that these experiments saturated the range of possible mutations). It is also possible that differences in T. gondii and P. falciparum codon usage (27) or dhfr-ts enzyme structure may play a role. All three mutations affect conserved or semiconserved amino acids in the dhfr domain that are likely to affect substrate and/or cofactor binding (Table 2), and at least one of these mutations (W25R) has previously been observed to cause antifolate resistance in mice (46).

All three mutations produce relatively low-level resistance to pyrimethamine, comparable to the levels observed for the mutations T83N and F245S (Table 3)T. gondii homologs of two single-point mutations known to produce pyrimethamine resistance in P. falciparum (4, 15, 29, 45). It is possible that no single-point mutation is capable of yielding a highly pyrimethamine-resistant parasite (31). In P. falciparum, high-level resistance to pyrimethamine is associated with multiply mutated dhfr-ts alleles, presumably derived by either successive addition of epistatically favorable mutations (31, 41) or genetic recombination between two low-level resistance alleles. The recombination scenario may be less likely, as mutations such as S36R (which confers high-level resistance in combination with T83N, equivalent to the Arg 59 and Asn 108 mutations in P. falciparum) typically confer no resistance on their own (31) and combining two low-level drug resistance mutations (such as L98S or L134H and T83N) does not necessarily enhance resistance (Table 3).

P. falciparum and T. gondii share many aspects of biology and pathogenesis (obligate intracellular lifestyle in vertebrate cells, common mechanisms of host cell invasion, similar ultrastructure and biosynthetic pathways, etc.) (36), but these parasites also differ in ways which are likely to influence the evolution of drug resistance. Because P. falciparum spreads from one human host to another via mosquito vectors, drug pressure can be maintained for many parasite generations, increasing the probability of acquiring multiple resistance mutations. The obligatory sexual cycle within the mosquito may also enhance the potential for recombination. In contrast, toxoplasmosis in humans is usually acquired by ingestion of water or soil contaminated with T. gondii oocysts (shed in cat feces) or by eating undercooked meat containing T. gondii tissue cysts (bradyzoites). Because humans are effectively dead-end hosts for T. gondii, it seems unlikely that high-level resistance will evolve through a succession of epistatically favorable mutations (the probability of obtaining multiple fortuitous resistance mutations within a single host is low). Since antifolates are not commonly used for treatment or chemoprophylaxis in domestic animals, there is little opportunity for drug resistance to develop in the field or for sexual recombination between individual drug-resistant mutants. Overall, the possibilities for mutation and selection beyond an initial host are considerably diminished for T. gondii relative to P. falciparum. Should clinical resistance to antifolates arise in T. gondii, it is most likely to develop via single, moderately potent mutations (such as those identified in this study), enhanced dhfr-ts expression (cf. the yeast experiments noted above and low-level resistance produced by transient transfection of wild-type DHFR-TS [Table 3]), or other attributes (such as the unidentified factors responsible for observed differences between T. gondii strain RH versus P(LK) and Veg [Fig. 2]).

In conclusion, little or no clinical data are presently available to address the extent to which drug resistance or strain sensitivity contributes to treatment failures for toxoplasmosis. However, long-term use of antibiotics is expected to impose strong selection for resistant organisms, and this study demonstrates that single-point mutations in T. gondii dhfr-ts (e.g., W25R, L98S, and L134H) can produce moderate levels of drug resistance in RH strain parasites. We also show that pyrimethamine sensitivity can vary markedly among different strains of T. gondii. These results do not answer the question of whether resistance is a clinical problem in toxoplasmosis, but they do provide new insights into potential sources of treatment failures and possible drug resistance mechanisms.