Background
There has been significant progress over the last ten years in the fight against malaria, a major threat to public health. The World Health Organization (WHO) reported that of 102 countries in which transmission continues, 52 recorded a decrease of more than 50% in the number of malaria cases between 2000 and 2010. In parallel, malaria mortality has also decreased by more than 45% globally and, in particular, by 47% in Africa. These results have been obtained by a combination of interventions, including artemisinin-based combination therapy (ACT), extensive coverage of exposed populations with long-lasting insecticide-impregnated bed nets (LLITN), malaria diagnosis by rapid diagnostic tests (RDTs), and intermittent treatment strategies [
1]. The changes in malaria epidemiology needs to be monitored through time by control programmes to assess the effects of the strategies implemented and to anticipate the consequences of changing rates of malaria transmission [
2]. Additional and appropriate tools are required to evaluate transmission, because parasite prevalence and entomological measures can be insufficiently sensitive in areas of low transmission [
3].
Various methods have been used since the 1960s to measure humoral responses to malaria and thereby evaluate malaria seroprevalence and transmission intensity. However, the variability of sources of antigen and subjectivity of some of the detection methods led to these methods falling out of favour [
4]. Sensitive enzyme-linked immunosorbent assay (ELISA) remains the most attractive technique for assessment of malaria transmission and changes in prevalence following the establishment of control programmes [
5]. Recent studies using characterized antigens have concluded that the evaluation of antibody persistence through cross-sectional surveys could be a complementary or even an alternative approach to investigating markers of malaria [
6]. There are however constraints and limitations to serological tests with characterized antigens: immunological responses to a given antigen may be, in part, genetically determined; and some of these antigens have the potential to, or may, display substantial polymorphism [
7]. Previous studies have shown that responses to crude extracts of malaria parasites, which are mixtures of numerous antigens, are less sensitive to these constraints [
5,
8].
The objectives of this study were to assess the value and feasibility of testing immune responses to parasite crude extracts. This approach was used to evaluate changes in malaria transmission over a ten year period in Dielmo and Ndiop, two villages of Senegal where there has been longitudinal follow-up since 1990. Various malaria interventions were implemented between 2000 and 2010, in accordance with the recommendations of the National Malaria Control Programme [
9]. These interventions were associated with a large decrease in malaria transmission. The relationship between these changes and local populations antibody responses to crude parasite extracts was investigated.
Methods
A cross-sectional study was conducted for the years 2000 and 2010 using the sera collected at the end of the wet season, during the Dielmo and Ndiop cohort study described previously [
10,
11].
Setting
At the beginning of the cohort studies, (in 1990 for Dielmo and in 1993 for Ndiop), the features of malaria transmission risk differed between the two villages. At that time Dielmo was an area where malaria was holoendemic with substantial activity of the main mosquito vector
Anopheles gambiae sensu lato complex and, therefore, perennial parasite transmission. By contrast, Ndiop was a mesoendemic area with moderate and seasonal transmission: transmission was 10 times lower than that observed in Dielmo [
12,
13]. The estimated annual entomological inoculation rate (EIR) in these areas was established periodically during the course of the cohort study. Between 2000 and 2010, the EIR significantly changed in the two villages: 482 infected bites/person/year in 2000
vs 88.7 infected bites/person/year in 2010 for Dielmo and 79 infected bites/person/year in 2000
vs 4.6 infected bites/person/year in 2010 for Ndiop.
The parasitaemia prevalence in November in the three age groups defined in this study was respectively 48.1, 54.1 and 53.3% in 2000 versus 4.9, 8.1 and 8.7% in 2010 in Ndiop. This prevalence was 44.9, 25 and 25% in 2000 versus 0, 4.8 and 3.3% in 2010 in Dielmo.
Population
The study was conducted using sera collected from individuals less than 20 years of age. A total of 442 serum samples from inhabitants of Dielmo and Ndiop collected for the cohort study were included in this study: 220 for 2000 (141 from Dielmo and 79 from Ndiop) and 222 for 2010 (143 from Dielmo and 79 Ndiop). Socio-demographic characteristics for the donors of these sera were extracted from the database established for the cohort study [
10,
11]. This study was examined and approved by the Senegalese National Health Research Ethics Committee.
Plasmodium falciparum culture and crude extract preparations
The
P. falciparum strain 07/03 (
Pf 07/03) was isolated from a patient in Dielmo and adapted to culture in the Immunology Unit of Pasteur Institute in Dakar [
14]. Parasites were cultured continuously on O
+ erythrocytes in RPMI containing 0.5% Albumax in candle jars according to the method described by Trager and Jensen [
15]. Crude extracts of schizonts of this strain (Sch07/03) were prepared by the method published by Wahlgren
et al.[
16]. After centrifugation of the culture, one volume of the pellet was lysed in three volumes of sterile distilled water and vortexed for homogenization. The extract was frozen without centrifugation at −80°C in working aliquots.
ELISA assay
ELISA Maxisorp plates (Nunc, Roskilde, Denmark) were used to optimize the dilution by dose effect. The optimal dilution was 1/300 and this was used for coating. Plates were coated with 100 μl of water-soluble crude extracts of Pf 07/03 schizonts (Sch07/03). Non-infected red blood cell were diluted in PBS and distributed into some wells on each plate as a control. The coated plates were incubated overnight at 4°C and washed. Then, 200 μl of blocking buffer (2% BSA in PBS with 0.05% Tween) was added to each well and the plates incubated for 1 hour at 37°C. Plasma samples were diluted at 1/200 in a dilution buffer (1% BSA in PBS with 0.05% Tween) and 100 μl of diluted serum was distributed into each well. Negative and positive controls were included on each plate: positive controls were from Dielmo/Ndiop hyperimmune individuals and negative controls were from European individuals. Polyclonal goat anti-human IgG conjugated to peroxidase at a dilution of 1/6,000 in the dilution buffer (1% BSA in PBS-Tween 0.05%) was then added. Bound peroxidase was detected with ortho-toluidine/H2O2 (100 μl) and the reaction was stopped by addition of 4 N H2SO4 (50 μl/well). Between each incubation phase, the ELISA plates were washed extensively with PBS-0.05% Tween. The optical density (OD) at 450 nm was read in a BIO-RAD Microplate Reader (iMark). The threshold for positivity was defined as an OD ratio >2 (OD sample/OD naive serum). Inter-assay variations for positive controls did not exceed 20%.
Statistical analysis
R software was used for statistical analysis [
17]. Three age groups were defined: [5-9] yrs, [10-14] yrs and [15-19] yrs. The seroprevalence of anti-Sch07/03 antibodies (Abs) was compared between 2000 and 2010 after standardization. Fisher’s exact tests were used to compare categorical variables and antibody levels between the villages and between the years 2000 and 2010. Non parametric tests (Kruskall Wallis, Wilcoxon) and the Spearman test were used to study non-parametric quantitative variables. Differences were considered statistically significant for
p < 0.05.
Discussion
The objectives of this study were to investigate the prevalence of anti-malarial antibodies and the levels of antibodies to crude extract of schizonts of a local P. falciparum strain. The ideal test for detecting malarial antibodies in epidemiological studies would be specific, give reproducible results and could be used at large scale. Therefore, the ELISA method was used as a technique suitable for high throughput and reproducible testing. Another advantage of this technique is its low cost, facilitating generalized and standardized use in the areas where it is needed.
Many seroepidemiological analyses have used a single recombinant antigen to evaluate serological responses. This strategy does not favour sensitivity and may lead to substantial underestimation of the immune response to
P. falciparum infection. Antigenic polymorphism and the diversity of individual responsiveness can have large effects on the serological responses observed, and thus on the results obtained with recombinant antigens [
18,
19]. By contrast, crude extracts containing numerous antigens may allow greater sensitivity and thus detection of low-level residual transmission. ELISA based on crude extracts is, therefore, more informative and a better tool for following endemicity [
5,
20].
In this study a schizonts crude extract of Pf Sch07/03 was used in the ELISA as a tool to measure and compare immune responses and seroprevalence. Antibody responses against this crude extract of Pf Sch07/03 of subjects recruited in the villages of Dielmo and Ndiop for antibodies were analysed.
This study was conducted in the villages of Dielmo and Ndiop. These villages were selected because the epidemiological contexts were different. Dielmo was holoendemic with a perennial and high transmission, with 200 to 300 infected bites/person/year [
13] whereas Ndiop was a mesoendemic area with moderate and seasonal transmission (during the rainy season between September and December), with 20 to 30 infected bites/person/year (10 fold less than in Dielmo). However, by 2010, the EIR had decreased to 79 infected bites/person/year in Dielmo and to 4.6 infected bites/person/year in Ndiop.
Seropositivity to
PfSch07/03 crude extracts was tested among the inhabitants of Dielmo and Ndiop for the years 2000 (220 subjects) and 2010 (222 subjects). The prevalence of anti-
P. falciparum Abs differed between Dielmo and Ndiop in 2000. The seroprevalence to
Pf Sch07/03 was high for all age groups studied in Dielmo, and highest in young adults (around 100%) in 2000. This high seroprevalence to
Pf Sch07/03 is consistent with the high levels of exposure in this region of perennial transmission and holoendemicity [
21]. The seroprevalence in Ndiop increased with age in 2000 and was at 83.6% for 15–19 yrs older group. This reflects the seasonal transmission and the delay in acquiring an immune response as described by Perraut
et al.[
22]. These findings are in agreement with seminal epidemiological reports [
23‐
25]. Studies conducted before the introduction of intensive malaria control programmes reported a higher prevalence and protective role of antibodies in the endemic Dielmo and Ndiop areas [
21] and West African countries more generally [
26]. The seroprevalence had declined substantially after 10 years in both Dielmo (44.4%) and Ndiop (34.6%), such that they were similar in the two villages in 2010; this is presumably a consequence of the anti-malarial strategies implemented since 2008 in Senegal [
27]. There was no significant difference in seroprevalence between males and females in Dielmo or Ndiop, in either year (
p > 0.05). These findings are in agreement with previous epidemiological studies [
28].
Differences in the increases and decreases of antibody prevalence between age groups were also considered. Seroprevalence decreased further and more significantly in the youngest group (by more than 40%) than the oldest group (less than 19%). There is probably a delay before acquiring an immune response in younger individuals, whereas adults display a ‘memory effect’ and their immune response can persist as a consequence of a history of contact with the parasite [
29,
30]. This could be due to the change of endemicity in Dielmo where transmission has declined following implementation of control programmes [
27,
31]. Similar findings have also been reported in other western African countries in 2013 [
32]. The analyses of the anti-
Pf Sch07/03 IgG response in the youngest age group in 2000 (corresponding to the 15–19 year-old group in 2010) revealed another interesting feature of malaria transmission and seroprevalence: the percentage of positive subjects in Dielmo was significantly lower 10 years later, suggesting that adults had, by 2010 lost the immune response to malaria antigens that they had had when they were ten years younger. By contrast, in Ndiop the seroprevalence was unchanged after 10 years in the same group. This observation might reflect the low intensity of transmission, especially among adults, in Ndiop.
The mean levels of anti-
Pf Sch07/03 Abs decreased substantially between 2000 and 2010 in both Dielmo and Ndiop. These declines were such that the magnitude of antibody responses did not differ between the two villages in 2010. Although seroprevalence appears to have remained constant over 10 years in Ndiop, the mean levels of Abs responses were highly significantly different between the younger group in 2000 and the corresponding older group in 2010. The decrease of the antibodies levels was comparable in Dielmo and Ndiop suggesting that a similar mechanism was operating to reduce malaria prevalence in the two villages. Trape
et al. reported a longitudinal follow up of inhabitants of Dielmo, showing that the incidence density decreased after LLITN distribution [
31]. Recent reports in Gambia demonstrate that these policies substantially reduce malaria morbidity, mortality and prevalence [
26]. The present study argues that sero-epidemiology could be a valuable additional monitoring tool particularly in the context of pre-elimination of malaria. The standardization by different laboratories of this method based on using crude extracts of
P. falciparum in assessments of the overall anti-blood-stage immune responses would make this approach even more useful for monitoring exposure to malaria.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AT conceived and coordinated this study. FD and AT conceived the manuscript. FD VR, GD and AT designed the methodology, conducted the analyses and wrote the manuscript and BD, MF, CS, JT, ND and AdT contributed to developing the concepts presented in this paper and provided valuable insights during the revision and editing stages. All authors read and approved the final manuscript.