INTRODUCTION
In Burkina Faso, as in many other sub-Saharan African countries, malaria is a leading health problem. In particular Plasmodium falciparum malaria represents the primary cause of child morbidity and mortality (Wurthwein et al. 2001).
Plasmodium species undergo sexual differentiation and develop gametocytes soon after the onset of asexual multiplication that elicits the acute phase of the disease. In P. falciparum, gametocytogenesis takes approximately 9–12 days. Once mature, the longevity of gametocytes and their infectivity to mosquitoes may vary considerably among different parasite isolates (Smalley and Sinden, 1977; Graves, Cater and McNeill, 1984; Trager and Gill, 1992).
Many factors influence gametocytogenesis and gametocyte prevalence in natural P. falciparum infection. Some environmental factors that influence asexual multiplication, for example anti-malarial drugs, host immune pressure, and presence of reticulocytes, have also been found to modulate gametocytogenesis (Smalley and Brown, 1981; Trager and Gill, 1992; Buckling et al. 1997; Buckling and Read, 2001; Targett et al. 2001). In addition, several studies have shown that a high parasitaemia and/or anaemia can also affect gametocyte production (Drakeley et al. 1999; Price et al. 1999; Nacher et al. 2002). Work on the rodent malaria parasite P. chabaudi has shown a positive association between high parasitaemia and gametocytaemia (Mackinnon and Read, 2003). In addition to the above environmental factors, it has also been demonstrated that P. falciparum clones can vary in their capacity to produce gametocytes (Graves et al. 1984).
In the present study we have examined P. falciparum gametocyte production in children of 2 sympatric ethnic groups, Mossi and Fulani, in villages in Burkina Faso. The two groups live in an area of hyperendemic and seasonal transmission in the dry savannah area in central Burkina Faso. Previous studies carried out among these groups have highlighted marked differences in parasitological and clinical susceptibility to malaria (Modiano et al. 1996; Luoni et al. 2001; Paganotti et al. 2004). Significantly lower prevalence of P. falciparum and incidence of clinical malaria have been reported among Fulani compared to Mossi. In addition, Fulani have stronger immune responses against some parasite surface antigens (Modiano et al. 1996, 1998), and more than 3-fold lower prevalence of malaria resistance genes than Mossi (Modiano et al. 2001).
Recently, we have noted a lower genetic complexity of P. falciparum among Fulani compared to Mossi in this area, despite similar rates of exposure to infected Anopheles mosquitoes (Paganotti et al. 2004). In the present study we have extended the above observations and examine the distribution of gametocytes and their genetic complexity among the two groups. The main aim was to examine whether the known stronger anti-malaria immunity in Fulani, that limits density and genetic complexity of P. falciparum, also affects gametocyte production.
MATERIALS AND METHODS
Study area, subjects and parasite identification
Two villages, Bassy and Zanga, approximately 1 km apart, are inhabited by Fulani and Mossi, respectively. The Geographical Position System (GPS) has been used to identify their position, which is 12 °19′N, 01 °09′W. A survey was carried out in each community at the end of the transmission season, between 21 October and 3 November 2002. A total of 92 and 58 children, aged <10 years, were recruited from Fulani and Mossi respectively. Formal consent was obtained from parents or guardians following explanation of the study design and objective. Finger-prick blood samples (200–300 μl) were collected from each child in EDTA, and stored at −20 °C in TRIZOL to prevent nucleic acid degradation. At the same time thick and thin blood smears were prepared for microscopical examination of parasite species and density. Stained thick smears were read over 200 high-power fields (hpf) and parasite density was estimated against 300 leucocytes. Children exhibiting fever (axillary temperature [ges ]37·5 °C) were treated with a standard chloroquine dose (25 mg/kg body weight).
The research was given ethical clearance from the Centre National de Recherche et Formation sur le Paludisme (CNRFP), Ouagadougou.
Entomological parameters
The main malaria vectors in the area are Anopheles gambiae s.l. and A. funestus (Petrarca et al. 1986). An entomological survey was carried out between 24 and 25 September 2002 in 17 compounds in Bassy and 14 compounds in Zanga. Indoor pyrethrum spray catches were made in the houses in each compound where the children sampled were living, 21 houses in Bassy and 22 houses in Zanga. The collections were performed between 07.00 and 09.00 a.m. The mosquito species were identified morphologically, then separated into fed and unfed individuals, and stored at room temperature with silica gel. We adopted this procedure because of the marked endophilic behaviour of the vectors in the area (Costantini et al. 1999).
PCR and RT-PCR for detection of P. falciparum infection and gametocytes
Parasite DNA and RNA were extracted using TRIZOL reagents (Sigma, UK), according to the manufacturer's protocol. PCR was first carried out to amplify the pfg377 gene to detect presence of sub-patent P. falciparum infection (Menegon et al. 2000). Reverse transcription of pfg377 mRNA and subsequent amplification of cDNA (RT-PCR) was used to examine the presence of gametocytes, since this gene is expressed only in these stages (Alano et al. 1995). Gametocyte-infected blood obtained from an in vitro culture of the P. falciparum 3D7 clone was used as positive control. Conventional PCR was run using the same RT-PCR premix to exclude possible contamination of RNA with co-extracted DNA.
In addition, RT-PCR was carried out to amplify RNA of the gametocyte-specific protein genes pfs48/45 (Kochen et al. 1993), using primers outer-01 (5′-TCT CCA TTT AGT CCA AAA GAC-3′) and outer-02 (5′-CAC CAG GAC AAT TTA AAC CTA CC-3′). Nested PCR was carried out to increase sensitivity of detection using primers nested N1 (5′-GTT ACA TCC GTG TAT GAC TTT-3′) and nested N2 (5′-TTT TCA AGA AGG AAA AGA AAA AGC C-3′).
Detection of pfs48/45 alleles in gametocytes and asexual forms
Both PCR (genomic DNA) and RT-PCR (RNA) of pfs48/45 were examined for polymorphisms in amino acids 253 and 254 to detect alleles KK, KN, EK and EN (Drakeley et al. 1996). These alleles were detected using dot-blot and sequence-specific oligonucleotide probes; KK-type (TAT CAT AAA AAG TTA ACT), KN-type (TAT CAT AAA AAC TTA ACT), EK-type (TAT CAT GAA AAG TTA ACT) and EN-type (TAT CAT GAA AAC TTA ACT) (Conway et al. 2001). The details of the technique are described elsewhere (Abdel-Muhsin et al. 2002).
Statistical analysis
We used ANOVA to test for differences between ethnic groups in the mosquito-biting rate. The mean numbers of bites per person per night per compound values were log-transformed prior to analysis to normalize distribution of residuals. We used generalized linear models (GLM) to test for effects of age and ethnic group on parasite rate and density, and gametocyte rate and density. Analyses were performed with PROC GENMOD in SAS 8.2 (SAS Institute, 1999). Gametocyte rate, asexual parasite rate and multiplicity of infection were modelled with a binomial error distribution, using the DSCALE option to correct for over dispersion. In analyses of asexual parasite density and gametocyte density, we tried both a negative binomial and Poisson distribution (correcting for over dispersion with DSCALE option). Both types of analyses yielded the same conclusions; we report here results from Poisson regressions.
In the analyses of densities we used values obtained by microscopy. PCR positive but smear negative samples were set to zero. We also performed analyses where these cases were assigned a value of 5 parasites/μl blood (Moody, 2002). The detection limit by microscopy is 10 parasites/μl, so PCR positive but smear negative cases can be expected to have a mean density of around 5 parasites/μl. These analyses yielded the same conclusions. We used the type 3 option for all GLM analyses (analogous to type 3 sums-of-squares in ANOVA). Sample sizes differ slightly between tests because of missing values. We report least squares means ±1 S.E.
RESULTS
Entomological indices
Anopheles gambiae s.l. and A. funestus were the main mosquito vectors in the region. The two species were equally abundant both in the Mossi community (Zanga village) and in the Fulani community (Bassy village). The biting rate was measured based on the number of people sleeping in each hut and the number of resting fed Anopheles caught. The average biting rate of A. gambiae s.l. was slightly higher among Mossi than Fulani, being 20·1 and 17·5 bites/person/night respectively (Table 1). The overall Anopheles biting rate was not different in Fulani and Mossi compounds, in Fulani being 17·74±3·43 and in Mossi 20·33±3·78 (ANOVA: F(1,29)=1·18, P=0·29).
Clinical malaria
There was a higher prevalence of fever among the Mossi compared to Fulani. Five out of 92 (5·6%) Fulani children had fever compared to 9 out of 58 (16·4%) of the Mossi group (P<0·05). The average temperatures among the examined children were 36·7 °C and 37 °C for Fulani and Mossi, respectively (P<0·001) (Table 1).
Parasitological indices
(i) Distribution of parasite species
P. falciparum, P. malariae and P. ovale occurred at a rate of 82·5%, 18·7% and 3·3% among Fulani, and 88·7%, 12·5% and 1·3% among Mossi, respectively, these differences being not significant. A total of 19·2% of mixed species infections were detected, 83% of which consisted of P. falciparum and P. malariae and 17% P. falciparum/P. ovale.
(ii) P. falciparum
P. falciparum infection was detected by microscopy and PCR. Microscopy parasite rate was not different among Mossi compared to Fulani (P>0·05 data) (Table 1). Similarly there was no significant difference in PCR parasite rate between the two groups (N=144, F=1·51, D.F.=1, 59, P=0·22). Further, there was no effect of age (F=3·00, D.F.=1, 59, P=0·083), and no interaction between age and ethnic group (F=0·89, D.F.=1, 58, P=0·35).
The density of asexual forms, as detected by microscopy among children who were positive by PCR (N=129), decreased with age (F=32·01, D.F.=1, 126, P<0·0001), and there was a significant difference between the 2 ethnic groups (F=6·38, D.F.=1, 126, P=0·013) (Fig. 1). However, there was no interaction between age and ethnic group (F=2·65, D.F.=1, 125, P=0·10), that is parasite density decreased with age at the same rate in Mossi and Fulani.
Fig. 1. Plasmodium falciparum asexual parasite density (LS means+/−S.E.) and gametocyte density (LS means+/−S.E.) in Mossi and Fulani.
(iii) P. falciparum gametocytes
Gametocyte rate, as detected by microscopy, was significantly higher among Mossi compared to Fulani (P<0·05) (Table 1). However, gametocyte rate as assessed by RT-PCR, did not differ between ethnic groups (F=0·50, D.F.=1, 59, P=0·48), but decreased with age (N=141, F=6·23, D.F.=1, 59, P=0·013). There was no age-by-ethnicity interaction (F=1·21, D.F.=1, 58, P=0·27).
Gametocyte density, among children who were positive by RT-PCR (N=81), was affected by ethnicity (F=24·11, D.F.=1, 78, P=<0·0001, Fig. 1) and age (F=16·9, D.F.=1, 78 P<0·001), their density decreasing with increasing age. There was no interaction between age and ethnic group (F=0·75, D.F.=1, 77, P=0·39).
There were no differences in gametocyte rates (BF and RT-PCR) between individuals with single species infections, predominantly P. falciparum, and those with mixed species infections P. falciparum/P. malariae and P. falciparum/P. ovale. In addition, there were no differences in gametocyte rates (BF and RT-PCR) seen among individuals with mixed species infections in either Fulani or Mossi.
(iv) Genetic complexity of gametocytes
Frequencies of gametocyte genotypes that carried KN, KK and EK alleles of pfs48/45 were 74·1%, 29·6% and 25·9% and 85·7%, 28·6% and 28·6% among Fulani and Mossi, respectively. No differences in allele frequency of pfs48/45 were seen between the two communities (P>0·9).
Multiple genotypes often co-existed in a single infection. The frequency of multiple gametocyte genotypes was slightly higher among Mossi children, with 24% of Fulani children having multiple gametocyte genotypes compared to 36% of Mossi (P>0.05). The mean number of gametocyte genotypes per infection was 1·42 and 1·46 among Fulani and Mossi, respectively (P>0·05). A generalized linear model showed no effect of ethnic group or age on multiplicity of gametocyte genotypes (N=49, F=0·17, D.F.=1, 29, P=0·68 and F=0·01, D.F.=1, 29, P=0·94), and no age-by-ethnicity interaction (F=1·54, D.F.=1, 28, P=0·22).
DISCUSSION
The most important new finding resulting from this work was that there were significant differences between Fulani and Mossi children in their density of P. falciparum gametocytes. In view of the known higher levels of immunity in Fulani, and previous findings of studies on P. falciparum gametocyte carriage in nature and experimental work with rodent parasites (Graves et al. 1988; Baird et al. 1991; Taylor and Read, 1997; Buckling and Read, 2001), it might be expected that Fulani would have higher gametocyte production than Mossi. However, we found no differences in gametocyte production per se (rate) and their genotype multiplicity between the two groups. Instead, Mossi had significantly higher gametocyte densities than Fulani.
Knowledge of environmental and genetic factors that promote parasite transmission is critical for better understanding of malaria epidemiology and novel control strategies that reduce parasite transmission. In human and laboratory models acquired immunity against malaria has been found to reduce asexual parasitaemia, gametocyte density and transmission rate (Graves et al. 1988; Boudin et al. 1991; Buckling and Read, 2001). The observed lower P. falciparum asexual parasite and gametocyte density among Fulani, agrees with the above observations, and can be attributed to their high anti-malarial immune responses. However, previous studies have indicated that environmental conditions that are unfavourable for parasite growth result in increased gametocyte production. One would, therefore, have expected a higher gametocyte rate among Fulani compared to Mossi.
Work on P. chabaudi, has demonstrated reduced asexual parasitaemia and parasite transmission in immunized mice compared to controls. However, gametocyte production per se (prevalence) increased in immunized mice (Buckling and Read, 2001), which agrees with the idea of stress-induced gametocytogenesis. In vitro studies in P. falciparum have demonstrated that addition of lymphocytes and serum obtained from infected people result in increased gametocyte production (Smalley and Brown, 1981). However, Baird et al. (1991) have noted reduced prevalence of gametocytes detectable by microscopy among natives of a hyperendemic area for malaria in Irian Jaya, compared to immigrants from an area of little malaria in Java. The authors concluded that these differences could not be explained by level of asexual parasitaemia, and were attributed to gametocyte-specific immune suppression (Baird et al. 1991). Similarly, Graves et al. (1988) have suggested that gametocyte-specific immunity was the cause of decreased gametocyte prevalence, in the absence of reduction in asexual forms, among older children in Papua New Guinea (PNG). In the present study we used RT-PCR to detect subpatent gametocytes and demonstrated no differences in gametocyte rate between the two groups. However, using microscopy alone, our results are similar to the above findings and showed significantly lower gametocyte rate among the naturally immune Fulani. Therefore, the higher gametocyte density seen in Mossi could be a reflection of their high asexual parasitaemia rather than a difference in the rate of gametocytogenesis between the two groups. The hypothesis of gametocyte-specific immune suppression (transmission blocking immunity) could be investigated using RT-PCR and newly developed quantitative RT-PCR (Schneider et al. 2004) to examine age and ethnic differences of P. falciparum gametocytes.
In view of the uniformity in parasite genotypes among the two groups and the absence of possible parasite population substructuring, the differences in parasite density can only be attributed to host factors. Our findings accord well with the earlier observations that the Fulani suffer less patent P. falciparum infections and clinical episodes of malaria than Mossi, and they were found to have better immune responses to malaria (Modiano et al. 1996). Thus, some density regulatory mechanism that maintains the parasite below clinical threshold appears to be operating more efficiently among Fulani. This is in agreement with our recent observation of lower genetic complexity of P. falciparum among Fulani compared to Mossi (Paganotti et al. 2004). The Fulani children are thus able to control parasite infection due to non-specific immunomodulatory factors which could be more frequent in this population compared to other ethnic groups in the area (Modiano et al. 1998).
There was no difference in gametocyte rate among people with mixed species infection compared to single species carriers among the two groups. Similarly, no particular combination of any mixed species infection was associated with high gametocyte carriage. The dynamics of individual clones in mixed species infections is complex (Bruce et al. 2000) and the transmission consequences of these dynamics very poorly understood. The use of RT-PCR and possible development of species-specific RT-PCR for gametocyte detection should shed light on this subject.
It was of interest that differences in parasitaemia and gametocytaemia among Fulani and Mossi were associated with age, and most noticeable among older children. This is consistent with our previous observation of age structure pattern of parasite multiplicity within each group (Paganotti et al. 2004). It is therefore likely that the densities of asexual forms and gametocytes are influenced by the immune status of the examined children, and the observed difference between the two groups is a reflection of the superior immune status of Fulani compared to Mossi. We did not examine immunological indices in the present study. The lower marked differences in parasitological and clinical P. falciparum indices have been attributed to stronger humoral and cellular immune responses against P. falciparum among Fulani compared to Mossi (Modiano et al. 1996, 1998).
Knowledge of the genetic diversity of gametocytes within an individual infection is important for predicting the parasites' ability to outcross and recombine to produce new strains. There has been a gap in our knowledge of the extent of gametocyte multiplicity in infected individuals. A large number of studies have revealed a varied extent of multiplicity of clones in the blood stages of P. falciparum in natural infections (e.g. Beck et al. 1997). It has been assumed that all parasite clones are committed to gametocyte production, but only a few studies have been made on complexity of gametocyte genotypes in natural infections by using RT-PCR to examine diversity of RNA of gametocyte-specific proteins (Niederwieser, Felger and Beck, 2001; Abdel-Wahab et al. 2002; Nassir et al. 2005). Using this method here, we have found extensive diversity of gametocyte genotypes among the two groups studied, with a higher multiplicity of gametocyte genotypes among Fulani compared to Mossi, although this difference was not significant. This contrasts with our previous observation of lower multiplicity of asexual forms and mean number of clones per person among Fulani compared to Mossi (Paganotti et al. 2004). The lack of differences in gametocyte multiplicity compared to asexual forms could perhaps be attributed to the lower polymorphism of the pfs48/45 used to examine gametocytes compared to the highly polymorphic genes MSP-1 and MSP-2 and GLURP used to examine the asexual forms in our previous study (Paganotti et al. 2004). Limited polymorphism of gametocyte-specific protein genes has been reported among parasites in other endemic countries in Africa (Niederwieser et al. 2001; Abdel-Wahab et al. 2002; Nassir et al. 2005).
The transmission consequence of high gametocytaemia in Mossi is significant. Generally, a positive correlation has been found between gametocyte density and infectiousness to mosquitoes (Graves et al. 1988), although subpatent gametocytes are known to be infectious (Muirhead-Thompson, 1954). We did not estimate entomological inoculation rates (EIR) amongst the two groups in the present study, although, in a previous study we noted a 2-fold higher EIR in the Mossi community compared to Fulani (Paganotti et al. 2004). The fact that the parasite populations in the two villages are not isolated suggests that Anopheles mosquitoes do move frequently between the two villages. The higher EIR as well as parasite and gametocyte densities among Mossi suggest a bigger source and force of infection in this community. As Fulani are equally exposed to mosquitoes, acquiring their infection from Mossi, control measures targeting gametocytes amongst Mossi could be beneficial to Fulani as well.
In summary, we have seen significantly lower gametocyte densities among the Fulani community compared to Mossi. This cannot be attributed to differences in exposure to Anopheles mosquitoes between the two groups. In this area of relatively high EIR, frequent crossing and recombination can readily generate new parasite strains; however, it is possible that the competent immune response of Fulani could protect them from super-infection with novel strains. When an effective vaccine against P. falciparum is eventually developed which could elicit and sustain such immunity, it could limit the acquisition, development and transmission of P. falciparum. Some studies have demonstrated lower multiplicity of asexual forms following administration of anti-malarial vaccines (e.g. Beck et al. 1997). Monitoring gametocyte multiplicity following vaccination could provide information on whether a vaccine might limit the parasite's ability to generate novel strains by crossing and recombination.
We would like to thank the inhabitants of Bassy and Zanga villages for their co-operation and help. We also thank the Humanitarian Organization ‘Bambini nel Deserto’ from Modena, Italy, that built a school for the two villages and that provides the GPS position of the two villages (www.bambinineldeserto.org). Financial support was received from the European Commission, the Medical Research Council, UK, and the Wellcome Trust, UK.
