Naturally acquired immune responses to malaria vaccine candidate antigens MSP3 and GLURP in Guahibo and Piaroa indigenous communities of the Venezuelan Amazon

Background Malaria transmission in most of Latin America can be considered as controlled. In such a scenario, parameters of baseline immunity to malaria antigens are of specific interest with respect to future malaria eradication efforts. Methods A cross-sectional study was carried out in two indigenous population groups in Amazonas/Venezuela. Data from the regional malaria documentation system were extracted and participants from the ethnic groups of the Guahibo (n = 180) and Piaroa (n = 295) were investigated for the presence of Plasmodium parasites and naturally acquired antibodies to Plasmodium falciparum antigens in serum. The GMZ2 vaccine candidate proteins MSP3 and GLURP were chosen as serological markers. Results The incidence of P. falciparum in both communities was found to be less than 2%, and none of the participants harboured P. falciparum at the time of the cross-sectional. Nearly a quarter of the participants (111/475; 23,4%) had positive antibody titres to at least one of the antigens. 53/475 participants (11.2%) were positive for MSP3, and 93/475 participants (19.6%) were positive for GLURP. High positive responses were detected in 36/475 participants (7.6%) and 61/475 participants (12.8%) for MSP3 and GLURP, respectively. Guahibo participants had significantly higher antibody titres than Piaroa participants. Conclusions Considering the low incidence of P. falciparum, submicroscopical infections may explain the comparatively high anti-P. falciparum antibody concentrations.


Background
To date, malaria is among the top ten causes of death in low-income countries [1]. In Venezuela, particularly the southern states of Bolivar and Amazonas are affected by malaria transmission. However, mortality is insignificant and malaria can be considered as controlled [2]. Amazonas (180,145 km²) covers an area nearly twice as big as Portugal and is inhabited by approximately 150,000 persons. Roughly half of the population is of indigenous origin and -although the absolute number of cases is low -Amazonas has the highest malaria incidence per capita in Venezuela (2007: 68.4 cases/1,000 inhabitants). Overall, the main malaria species is Plasmodium vivax (ca. 80%) followed by Plasmodium falciparum (ca 20%, with declining tendency). Plasmodium malariae is only occasionally detected [3,4].
As drug resistance continues to be a major problem, an effective malaria vaccine against P. falciparum would be a powerful tool in the control of malaria [5,6]. The vaccine candidate GMZ2 is a fusion protein of P. falciparum merozoite surface protein 3 (MSP3) and glutamate rich protein (GLURP), which has been evaluated during phase I trials as a promising vaccine candidate in Germany and Africa [7,8]. Antibodies against both antigens have been shown to provide partial protection in Saimiri sciureus monkeys [9] and were associated with protection from human clinical malaria [10]. However, no data are available in respect to naturally acquired immunity to MSP3 and GLURP from populations of Latin America [11]. This is of interest since a malaria vaccine such as GMZ2 might be used in other continents than Africa where exposure to P. falciparum is low.
In 2008, a pilot programme for the evaluation of blisterpacked treatment in distinct ethnic groups was initiated. An initial cross-sectional study was carried out in two indigenous population groups in order to gain knowledge about the micro-epidemiology of malaria in the project area. A second cross-sectional study is planned to close the project. As serological markers are useful indicators to measure transmission variations especially in low endemicity areas [12], immune responses to MSP3 and GLURP antigens were investigated as indicators of naturally acquired immunity to P. falciparum antigens.

Study population
The study took place in June 2009 in the municipality of Atures, Amazonas, Venezuela. Three indigenous communities were visited. They were inhabited by two ethnic indigenous population groups, the Guahibo (also: Guajibo, Wahibo, Hiwi, Jivi) and the Piaroa: 1) Platanillal (475 inhabitants, Guahibo); 2) Cerro de Oro (60 inhabitants, Guahibo); 3) Paria Grande (463 inhabitants, Piaroa). The distance between the communities is small (less than 20 km linear distance) and they are comparable with respect to ecological, geographical and malariological conditions such as distance to Anopheles breeding sites. The majority of residences in the communities are non-traditional governmental housing projects. The communities of Platanillal and Paria Grande have a health post, each with a malaria microscopist.

Ethical clearance, informed consent, and treatment
Ethical clearance was obtained from the institutional ethical committee of the Amazon Center for the Investigation and Control of Tropical Diseases 'Simón Bolívar', Autonomous Service, Puerto Ayacucho, Amazon State, Venezuela (SACAICET). Residents were informed about the ideas and procedures of the study, when necessary with the help of translators. All residents were invited to take part. Those consenting orally were interviewed and examined. Individuals with a blood smear positive for malaria were treated according to the national guidelines (P. falciparum: artesunate, mefloquine, and primaquine. P. vivax: chloroquine, primaquine) [13].
Data extraction of the regional malaria documentation system Data from the regional malaria documentation system were extracted. Malaria cases detected in the health posts of Platanillal and Paria Grande before the study took place, from January 2003 to May 2009, were analysed.

Interviews and basic physical examination
Every participant, or the legal guardian, was interviewed using semi-structured interviews. Translators assisted if it was necessary, and standardized questions were phrased with the support of anthropologists. The structured part of the interviews concerned malaria antecedents, possession and use of mosquito nets, general condition, unspecific symptoms during the last seven days, and pregnancy. Every participant was examined physically including a check for splenomegaly and body temperature.

Smears and serum
Thick blood smears and thin films were taken from every participant, air-dried and stained for 45 min with 3% Giemsa using standard procedures1 [14]. Slides were read and re-read by trained professionals of the SACAI-CET and of the Regional Health Directory in Puerto Ayacucho. Ten ml of blood was taken from every participant into a vacutainer and centrifuged. Serum was stored at −20°C until further processing.

Recombinant antigens
Recombinant proteins were produced in Escherichia coli as described elsewhere [15]. The 48 kDa merozoite surface protein 3 (MSP3) is divided into two major allele types by sequence polymorphism. The C-terminal region of the molecule is highly conserved [16]. The 220 kDa glutamate-rich protein (GLURP) contains one non-repeat and two repeat regions. Both antigens are involved in antibody dependent cellular inhibition (ADCI) and are used as subunits of the GMZ2 fusion protein [17]. For all experiments the conserved fragments MSP3 212-318 and GLURP 27-500 were used.

Antibody measurements
Serum antibodies to MSP3 and GLURP were measured applying a modified version of a quantitative ELISA [18]. In brief, microtitre plates were coated with antigens in a dilution of 0.5 mg/ml in PBS overnight at 4°C. After a first washing cycle (3 times) with PBS-T plus 0,5 mol/litre NaCl wells were blocked for one hour with a dilution of 30 g of defatted milk powder per litre PBS-T. As standard, after another washing cycle, duplicate rows were coated with samples, controls and a standard sample (with quantified high antibody concentration) in blocking buffer for two more hours. The standard sample was added in an X-fold serial dilution from 1/20 to 1/327,680. Samples and controls were added in duplicates in dilutions of 1/50 and 1/200. Samples with optical densities higher than the positive control were repeated at 1/800. After washing again, plates were incubated with peroxidase conjugated goat anti-human IgG (Caltag, 1/5,000) for one hour, washed, treated with TMB-One for 15 min in gloomieness until the reaction was stopped with sulphuric acid. Then absorbance was read at 450 nm. The standard sample was used to calculate IgG concentration of the other samples in mg/dl. Pooled serum from 20 individuals, never having lived in malaria-endemic areas, served as negative control. The arithmetic mean of negative controls was 0.0746 mg/dl (SD 0.0396 mg/dl) and 0.092 mg/dl (SD 0.0441 mg/dl) for MSP3 and GLURP, respectively.

Data capture and analysis
All data were subjected to plausibility, double-entered and checked for transcription errors in Epidata v3.0 ® . Optical densities (ODs) were averaged and normalized. The cutoff value for the difference between duplicates was set at 25%. If the difference was higher, the measurement was repeated. All results were converted to values in mg/dl using the positive standard serum. Positivity and high positivity of samples was defined as any OD higher than the mean of negative controls plus three and five standard deviations (SD), respectively. Statistical analysis was performed using SPSS 19 ® software. T-Tests were run to screen for significant differences between the variables subgroups concerning antibody titres, Chi-Square-Tests to learn about disparities in respect to antibody positivity. Age in years, sex, self-reported incidence of malaria episodes during the last 12 months and lifetime, possession and frequent utilization of mosquito nets and ethnic affiliation were included in a binary logistic model to predict anti-MSP3 and anti-GLURP antibody positivity. This quite conservative model has the advantage to show an Odds ratio (OR) for each expression of a variable. Furthermore, a linear regression model was performed to expose the influence of the aforementioned independent variables on the direct IgG-titre against the GMZ2-subunits.To do so, antibody levels had to be transformed to their decadic logarithm to gain Gaussian distribution. The resulting model was then de-logarithmized to show comparable and comprehensible results.

Study population
Four hundred and seventy-five persons participated in the study. Demographic data of participants and questionnaire results are summarized in Table 1.

Malaria infections in the two indigenous communities from 2003 to 2008 Platanillal
Within the six years preceding the cross-sectional study (2003 -2008)

Paria grande
Within the five months preceding the cross-sectional study, 148 slides were taken in the health post of Paria Grande; 15/148 slides (10.1%) were positive for P. vivax, and none (0%) was positive for P. falciparum.

Stratification for age and sex
Antibody titres were significantly increasing with age (Spearman Rank Test, R SpGLURP = 0.130, p = 0.005. R SpMSP3 = 0.259, p < 0.001). When the study population was grouped by age -each group with a span of 10 yearsa significantly higher number of positives and high-positives was seen in the older age groups (ANOVA: positives: p MSP3 < 0.001, p GLURP = 0.001, high-positives: p MSP3 < 0.001, p GLURP = 0.028). No significant differences in antibody titres or positivity rates were found regarding the sex of the persons (T-test: p GLURP = 0.101, p MSP3 = 0.169, Chisquare: p GLURP = 0.101, p MSP3 = 0.184).

Stratification for ethnic origin
Participants from the ethnic Guahibo group displayed significantly higher antibody titres and positivity rates for both antigens when compared to the ethnic Piaroa group (Table 2).

Binary logistic and linear regression model
The above-mentioned parameters were included in a binary logistic and a linear regression model to predict positivity and titres of antibodies against MSP3 and GLURP.
In the binary logistic model, the parameters age, possession of mosquito nets, and ethnic origin had a significant outcome:   Table 3.

Discussion
Results show that the two GMZ2 antigens are recognized by antibodies acquired through natural infections in the Venezuelan Amazon. Overall, the results correspondexcept high responders -to rather low responses which are roughly two-fold higher than the cut-off value for positives. As antibody subclasses are of great importance [19] further studies on IG-subclasses have to show whether functional antibodies are acquired through natural exposure. The question if a GMZ2 vaccine would be effective in this area remains open, because antibodies stimulated by vaccines do not automatically protect as do naturally acquired antibodies [20]. In respecto to vaccine efficacy, only clinical trials with disease as endpoint can prove the impact of a vaccine. The two main findings of this study were 1) the high prevalence of antibodies against conserved parts of MSP3 and GLURP antigens in the two population groups with low incidence of P. falciparum malaria, and 2) the significantly higher prevalence of MSP3 and GLURP antibodies in the Guahibo group when compared to the Piaroa group.
The total population of the Piaroa is about 12,000 persons. Anthropologists describe them as peaceful and egalitarian. Traditionally, the Piaroa's agroforestal subsistence strategy implicated small, dispersed and mobile ways of settlement in groups of about 50 individuals. Nowadays they are living in bigger communities mostly built by the Venezuelan government in the periphery of their ancestral territories, which they still use as cropland and hunting grounds [21].
The Guahibo consist of about 30,000 persons settling mainly around Puerto Ayacucho and neighbouring Colombian areas. As a patriarchal society, great importance is placed on the nuclear family. Traditionally, they inhabited vast flatlands in Colombia and Venezuela as nomadic foragers. Nowadays, they inhabit communities constructed by the Venezuelan government. They are mainly assimilated according to language and clothing style, but still are threatened by pauperisation [22].
The serological results of this study confirm previous investigations hinting to an elevated risk for malaria in the Guahibo population [2,23]. Living conditions, socioeconomic parameters, adherence to medicaments, confidence in the health system, and cultural habits may explain the different transmission levels in the two ethnic groups under study. Different immune responses with different protection patterns in genetically different populations are also a possible factor [24,25]. Further research involving multidisplinary expertise as well as members of the communities should focus on this "hot spot" of malaria transmission in the Venezuelan Amazon.
The positive correlation of antibody titeres and prevalences with age has been known for a long time [26]. When people live in an area endemic for malaria the chances of contact with the antigen raise with the years, and so do antibody titres and prevalences. Similarly, the chances of contact decrease with the use of mosquito nets. Such, as shown in other studies [27], a difference in GLURP antibody titres and prevalences could be observed between mosquito net users and non-users.
This cross-sectional study evaluated immune responses against two blood stage antigens considered as promising vaccine candidates on the basis of studies carried out in completely different settings. Samples were taken in the period of low transmission when the direct contact with P. falciparum parasites is expected to be minimal. Indeed, only two persons in nearly half a thousand participants were found harbouring parasites (2/475; 0.4%). Both patients were positive for P. vivax, none was positive for P. falciparum. Nevertheless, nearly a quarter of participants had antibodies to at least one of both antigens. As adding up immune responses to individual antigens might not be an appropriate procedure for results, which show -not surprisingly for outbred human populationsan extensive variation between individuals, it is important to note that 7.6% and 12,8% of participants had high antibody titres against MSP3 and GLURP, respectively. Possibly only high-positive antibody responses are relevant and/or specific results in this endemic area. Moreover, it should be taken into account that higher antibody responses might persist longer than lower ones.
Using the records of the health information system of Amazonas the level of malaria transmission in the communities during the years and months before the study was approximated. As the health posts of Platanillal and Paria Grande were comparable in the volume of slides taken per year (e.g. 2009: Platanillal, 280 slides; Paria Grande, 269 slides) and data of ethnic origin, origin of infection, and population census were available, the prevalence in both communities could be estimated for malaria in general, and for P. falciparum in particular. The average pan-malarial incidence was higher in the Guahibo population of Platanillal (ca 18%) than in the Piaroa population of Paria Grande (ca 8%). However, this was due to P. vivax infections, whereas the transmission of P. falciparum was surprisingly low in both population groups. Within a time period of 6.5 years, only 50 and 36 cases of P. falciparum were detected in Guahibo and Piaroa people of the health posts of Platanillal and Paria Grande, respectively, and the average incidence of P. falciparum was estimated to be less than two percent in both population groups, without a significant difference.
MSP3 and GLURP antibodies last six-12 months if no boost infection takes place [8]. In order to explain the phenomenon that antibodies were detected in more individuals than expected, submicroscopical infections must be considered as an influencing variable. The use of other antigens (for example, pre-erythrocytic antigens) might have led to other results, yet it can be assumed that conclusions would have differed slightly only. This is in line with previous investigations demonstrating asymptomatic carriers in this area [28]; interestingly, more than 20% of those who denied they ever suffered from any type of malaria had positive IgG titres.
Another -yet to be proven -hypothesis would be the polyclonal activation of effector T killer cells and memory B-cells against P. falciparum antigens by P. vivax [29].
Controlled malaria is the result of an interplay of host defense systems, public health measures, ambiental conditions, behavioural features, and epidemiology and infection biology of the parasite. Interventions should be weighed carefully in order not to destabilize an established equilibrium. The epidemiology and immunology of controlled malaria might be subject of growing interest for malariologists and decision makers, as lessons learnt in low endemic areas could be instructive after eradication and control efforts have been successful in high endemic areas.