Malaria Journal Concurrence of Plasmodium Falciparum Dhfr and Crt Mutations in Northern Ghana

Background: Both chloroquine (CQ) and sulfadoxine-pyrimethamine (SP) are failing drugs in much of sub-Saharan Africa. Previous findings suggest an association between resistance to CQ and to SP in vivo, in vitro, and on the molecular level.


Background
Chloroquine (CQ) and sulfadoxine-pyrimethamine (SP) are the most frequently used antimalarial drugs in sub-Saharan Africa. With the spread and intensification of drug resistance o f Plasmodium falciparum more effective combinatory drug regimes are now introduced in many African countries [1]. However, CQ and SP are cheap and widely available and will most likely be in use for some more years, particularly in home treatment [2][3][4].
In several East African countries where SP was introduced in response to intense CQ-resistance, the drug has gradu ally lost efficacy although the pace of this development is subject to controversy [5,6]. Previously, the core muta tions linked with resistance to CQ and SP were found to be associated with each other in isolates from southern Ghana [7]. Although CQ and SP are structurally unrelated and the mutations conferring resistance are located on separate chromosomes [8][9][10], this finding suggests that parasites resistant to CQ may acquire resistance to SP more easily than sensitive ones, or vice versa. Studies in murine models [11] and in vitro [12,13] support this hypothesis. If this was true, the spread o f CQ resistance could pave the way for an accelerated development of SP resistance, and thus, bear substantial importance to the health systems o f affected regions. In addition, transmis sion potential may be increased in both SP and CQ resist ant parasites [14,15].
Recently, mutations in the P. falciparum dihydrofolate reductase (dhfr) gene in northern Ghana were observed not only to be predictive for SP treatment failure but also to be associated with increased pre-treatment gametocyte car riage [16]. Here, associations between dhfr alleles, the core mutation in the P. falciparum chloroquine resistance trans porter gene (crt T76), residual antimalarials, and gametocyte carriage in children with uncomplicated malaria were re-examined.

M ethods
P. falciparum isolates were collected from children with uncomplicated malaria participating in a treatment trial in Tamale, Northern Region, Ghana, at the end of the rainy season 2002. The results o f this trial and on the dhfr and dihydropteroate synthetase gene (dhps) patterns are described elsewhere [16,17]. In the present report, data from all 126 children with complete follow-up after SP treatment for whom dhfr and dhps were genotyped are analysed. The study protocol was approved by the Ethics Committee, Ministry of Health, Northern Region, and by the Health Research Unit, Ministry o f Health, Accra, and parents' informed consent was obtained.
Venous blood was collected into EDTA. Asexual parasites and gametocytes were counted per > 200 and 500 white blood cells, respectively, on Giemsa-stained thick blood films, and densities were calculated based on a putative mean WBC count of 8,000/p.L. Gametocyte counts one week following treatment were available for 104 children. Pre-treatment levels o f CQ and pyrimethamine in blood were measured by ELISA assays [21] with limits of detec tion of 10 ng/mL and 25 ng/mL, respectively. P. falciparum dhfr, dhps, and crt alleles were assessed by restriction fragment length polymorphisms (RFLP) of amplicons generated by nested PCR assays applying hot start procedures (HotStart Taq, Qiagen, Germany). Prim ers and conditions are described elsewhere [22,23] as are restriction enzymes and RFLP conditions to characterize the codons dhfr, 16, 51, 59, 108, 164; dhps, 436, 437, 540, 581, 613; and crt, 76. Laboratory strains were used as con trols for PCR and RFLP assays. Electrophoresis was perfomed on 3% GTG agarose (FMC Bioproducts, USA) gels.
Frequencies were compared by %2-test or Fisher's exact test, and continuous variables by Student's t-test, Mann Whitney U-test, or Kruskal-Wallis test as applicable. Logis tic regression models were used to adjust for potential confounders o f the presence o f resistance mutations and to identify risk factors for gametocytaemia.
Dhfr, dhps, and crt alleles were examined with respect to pre-treatment gametocytaemia. As reported elsewhere [16], children were significantly more likely to harbour gametocytes in the presence o f dhfr 51, 59, or triple muta tions. This was also true for isolates exhibiting crt T76 (Tab. 2 ). Gametocytes were not observed in any (0/21) isolate with wildtype alleles for both dhfr and crt, but in 16% (8/50) of isolates exhibiting either dhfr or crt muta tions, and in 29% (16/55) of isolates revealing both, dhfr triple and crt T76 (%2trend = 8.7, P = 0.003). Adjusting for additional factors influencing gametocytaemia, i.e. high parasite density arbitrarily set as >50,000/|iL, the presence o f CQ in blood, and axillary temperature, both the dhfr triple mutation and crt T76 were associated with an increased likelihood of gametocytaemia (Tab. 2 ). Dhfr alleles or crt T76 were not associated with gametocyte den sity (data not shown).

Discussion
In this study on P. falciparum isolates from northern Ghana, two m ajor findings are presented. First, SP and CQ resistance markers are strongly associated with each other, independent o f residual antimalarials. Second, both are associated with an increased prevalence of gametocytes.
This study has several limitations and particularly the finding of an association between unrelated m utations needs caution in interpretation. Because polyclonal infec tions predom inate in the area [24], the detection o f a m utant allele in an isolate does not necessarily m ean that all clones carry the m utation. Thus, it cannot be stated whether the linkage between the resistance markers is a true one, i.e. on the chromosom al level, or reflects co occurrence in individual isolates. The limited num ber of crt wildtypes also ham pered proper testing of a linkage disequilibrium. Although dhfr m utations were signifi cantly m ore com m on in the presence of crt T76 this obser vation needs to be construed with caution since 80% of the isolates in this study harboured the latter variant. Also, it is impossible to com m ent whether the presence o f crt T76 favours the presence of dhfr m utations or vice versa because this requires longitudinal observations. Likewise, the lack of tem poral inform ation impairs a clear state m ent on whether resistance m utations bring about increased gametocytogenesis. Due to m ethodology the resistance genotype of asexual parasites cannot be sepa rated from that o f gametocytes.
Despite these difficulties in drawing firm conclusions, sev eral previous findings support the hypothesis of a linkage disequilibrium between m utations associated with SP resistance and CQ resistance. Early reports from the 1950s and 1960s indicated that in areas or patients with estab lished pyrimethamine-resistance in Nigeria, Burkina Faso, and Venezuela, CQ exhibited a reduced activity (reviewed in [25]). In a m urine malaria m odel, CQ resistance could be induced in pyrimethamine-resistant parasites but not in sensitive ones [11]. In isolates from Cameroon, the in vitro activity of pyrim etham ine was ten times lower in CQ resistant than in sensitive P. falciparum [13]. Similar b u t less pronounced differences have also been observed in other studies [26,27]. In southern Ghana, we previously observed that the dhfr core m utation Asn-108 was three times more likely found in isolates exhibiting crt T76 than in isolates comprising crt wildtype parasites [7]. The rea son for this apparent association between resistance to CQ and pyrim etham ine or SP is obscure since b oth drugs have distinct m odes of action and resistance to these is determ ined by m utations on different chromosomes [8 10]. One alluringly simple explanation could be that par asites in the study area have merely become resistant to both drugs, possibly as a result o f drug pressure. In this regard, previous, simultaneous or sequential treatment with CQ and SP could have selected for resistant parasites which subsequently persisted for a longer period than the drugs can be detected in blood. However, pyrim etham ine was seen in only two children which indicates that it is rarely used in this com m unity and which argues against drug-induced selection for SP resistance. In addition, the results are corrected for the presence of CQ in blood which can be detected for approximately one m onth after intake [28]. This does no t exclude the possibility of selected and persisting parasites but renders it rather unlikely. Alternatively, the association between crt T76 and dhfr m utations could reflect a rapid m utator pheno type with the ability of accelerated resistance to m ultiple drugs. This hypothesis originates from in vitro studies observing that parasites resistant to com m on antimalarials acquire resistance to structurally unrelated drugs more rapidly than susceptible strains. The genetic basis of this phenom enon is unknow n b u t suggested to involve an increased frequency of m utations per se and consequently a higher probability of m odified proteins which could also include drug targets [12]. In fact, selection for high grade pyrim etham ine resistance in vitro has been shown to enhance the degree of overall genomic polym orphism [29]. In this regard, it is noteworthy that crt T76 was more frequently observed with increasing num ber of dhfr m uta tions, i.e. with increasing degree of SP resistance.
In the present study, the association between the resist ance markers meets with an increased presence of gametocytes in isolates comprising m utant dhfr or crt alleles. Gametocytaemia seemed to reflect a rather long duration of infection as can be deduced from its low prevalence in the presence of factors suggestive for acute disease, i.e. high body temperature and parasite density, and previous CQ treatm ent (Tab. 2). Again, an increased frequency of resistance m utations in gametocytaemic children could result from previous drug-related selection as outlined above. However, increased gametocytaemia preceding treatm ent has also been observed in infections subse quently found to be CQ resistant [30-32]. Hallett et al. [15] reported that in patients with crt T7 6 parasites, gametocyte density was highly increased one week following CQ treatment. In addition, gametocytes from patients car rying crt T76 parasites produced 38 times higher oocyst burdens in the m osquito as compared to crt wildtype par asites. In Tamale, both residual CQ levels and parasites with the crt T76 m utation are abundant [33], and a clear association of crt T76 and pre-treatm ent gametocyte prev alence is seen. Only one crt wildtype isolate contained gametocytes im pairing a sound analysis of the effect of crt T76 on gametocyte density. However, increased gametocyte production by crt T76 parasites [15] in the presence of residual CQ could partially explain the present finding. The synthesis of both, crt and dhfr m utations being asso ciated and increased gametocytaemia in their presence, gives rise to a grim scenario: Given that CQ resistant par asites have an im proved transmission potential [15] the association with SP resistance would contribute to an accelerating spread of resistance to both drugs, particu larly in areas where CQ resistance is frequent. In the present study, neither selection of crt T76 in SP treatm ent failure was observed nor a significantly elevated propor tion of post-treatment gametocytaemia am ong children initially harbouring crt T76 parasites. Although this was n o t expected and would reflect an extraordinary rapid process, both observations may be influenced by the small sample size.

Conclusion
The present data provide evidence supporting a hypothe sis on a connection between resistance to CQ and SP sug gesting that both, CQ and SP resistance favour transmission. This needs to be verified by carefully designed longitudinal studies in regions of differing levels o f drug resistance and endemicity. Per se, antimalarial treatment m ust be effective, and m ore effective than CQ and SP, not only to reduce treatment failures but also the transm ission of potential co-resistance. Eventually, as this has been shown to counterbalance enhanced transmis sion of resistant parasites [15] the results strongly support com binatory treatment including artemisinine-derivatives.

A u th o rs' contributions
FPM, JTB, and UB designed the study. RNO, SE and FPM were responsible for patient recruitment, clinical and par asitological examinations, and PCR assays. TAE measured drug concentrations. JTB and RWS did the gametocyte counts. FPM and JTB wrote the paper with maj or contribu tions of the other authors.