Study Area and Collection of Clinical Data

The village of Dielmo (13°45′N, 16°25′W) is localized in one of the rare areas of Senegal, West Africa where malaria is holoendemic (experiencing perennially a high level of transmission by mosquitoes, due to the presence of a permanent stream), with an average of 5.16 infective bites per week during the first 2 y of the survey [16]. The 247 inhabitants of Dielmo village were enrolled in a prospective study using to our knowledge one of the most stringent protocols of clinical follow-up ever applied in the field and consisting of daily surveillance by medical staff (present 24 h/d, 7 d/wk) in order to identify and to analyse all episodes of morbidity [16]. The field set-up was designed and tested over 1 y before the actual study was conducted (e.g., questionnaires used for daily surveillance were written in three languages, and the reliability of responses were systematically addressed).

Each villager was visited daily at home and had the ability to consult at any time one of the two medical doctors permanently on-site. In the event of a report or complaint of fever, headache, or vomiting, a medical examination and three thick blood films were made. One of the thick blood smears was Giemsa-stained and examined immediately on-site for the purpose of deciding on treatment. The other two slides were dehaemoglobinized, stained, and examined in our central laboratory in Dakar, using more rigorous and standardized conditions with quality control assessment [16]. The results of the latter slides were used for the present study.

The criteria leading to a given episode of morbidity being attributed to malaria have been studied in detail and defined previously [7]: a malaria attack was defined as an episode of fever (temperature >38.5 °C) associated with a parasite density exceeding an age-dependent pyrogenic threshold described for this village (the parasite density threshold for each age group was determined to be 24,500 parasites/μl at ages < 12 mo; 27,000 at 12–23 mo; 24,000 at 2 y; 20,000 at 5 y; 15,500 at 10 y; 10,000 at 20 y; 7,500 at 30 y; and 5,000 in adults older than 40 y) [7]. The improved ability of this criterion to distinguish malaria attacks from other causes of fever has been documented [17] and has been further confirmed independently in different African settings [18–20].

Antimalarial treatment was initiated in all participants with confirmed malaria attack according to previously established criteria [7,21]. During the study period, treatment relied on the administration of quinine chlorhydrate (Quinimax), chosen in view of its efficacy in this area of chloroquine resistance and its fast effect, given orally at a dose of 8 mg/kg under medical supervision (with assessment of proper ingestion) at 8 h intervals (i.e., 25 mg/kg/d) for 7 d. It is of note that the very short half-life of quinine has the advantage that it avoids a buildup of a prophylactic concentration of the drug in the receivers' blood that could provide artificial protection in the following weeks or months, i.e., the influence of this confounding factor is avoided. Since free medical care was available 24 h/d, self-treatment was most uncommon, as was ascertained by systematic detection of antimalarial drugs previously [16].

All villagers were farmers with equivalent economic and social status. Entomological studies were conducted 15 d/mo, year-round, by 12 investigators rotating each night from one household to another. These studies were conducted for several years simultaneously with clinical data recording. They did not show any substantial differences in exposure to Anopheles from one house to the other, and clinical cases were at similar prevalence in all households. The average entomological inoculation rates reported in this paper correspond to the actual entomological inoculation rates measured during the same years of the clinical survey.

The Dielmo Project was initiated in 1990, and the present study initially focussed on the period from July 1990 to July 1992; over that two-year period we calculated that each individual received an average of 520 P. falciparum-infected bites. The serum samples were collected in a cross-sectional study in May and June 1991. The clinical data collected from July 1990 to July 1991 were used for retrospective statistical analysis, and the clinical data collected between July 1991 and July 1992 were used for prospective statistical studies. Since the village continued to be followed up clinically according to the same criteria, and in view of results obtained from July 1990 to July 1992, clinical data obtained over the ensuing 5 y (up to July 1997) were employed for the long-term analysis.

Ethical Approval

The informed consent of each villager (or that of the parents in the case of children) was obtained at the beginning of the study after a thorough explanation of its purpose and was renewed at the beginning of each year of the survey. The study design received clearance from the Senegal National Ethics Committee (Dakar, Senegal).

Antigens

The peptide MSP3-b used in this study is part of the C-terminal domain of the MSP3 DG210 protein and has been described previously [22,23]. It is located in the highly conserved C-terminal region of the molecule (amino acids 185–254) [24,25] (Oeuvray et al., unpublished data). It defines a B cell epitope targeted by naturally occurring antibodies. Human affinity-purified antibodies on peptide MSP3-b react with the parasite native protein in Western blots of P. falciparum schizont extracts, and in immunofluorescence on infected RBCs, thus showing the relevance of the peptide epitope to the original protein, and inhibit parasite growth in cooperation with blood monocytes, in vitro or in vivo, in passive transfer experiments [24,25]. The RESA peptide was purchased from Bachem (Bubendorf, Switzerland). The same sera were tested under the same experimental conditions using the MSP1–19 [26,27] and two MSP2 (fcr3 and 3d7 alleles) polypeptides expressed in E. coli [28], and under the supervision of David Narum (Biomedical Primate Research Center, The Netherlands), using the AMA-1 recombinant protein, derived from clone 7G8 sequence, expressed in baculovirus, and hence properly conformed [27] (a gift of D. Narum and A. Thomas (Biomedical Primate Research Center, The Netherlands), ). Control peptides and recombinant proteins were chosen among pre-erythrocyte stage antigens, namely the peptides LSA1-R derived from liver stage antigen 1 [29], and LSA3-RE derived from liver stage antigen 3 [30], and the recombinant R32 LR derived from the circumsporozoite (CS) protein [31] spanning the NANP repeats.

ELISAs

The total and isotype-specific ELISAs were performed as described previously [22,32,33] using secondary monoclonal antibodies originally selected as reacting faithfully with subclass allotypes of European, African, and Asian origins [22,32,33]. For IgG1 to IgG4 (IgG1–4), monoclonal mouse anti-human subclasses (clones NL16 = IgG1 [Boehringer], HP6002 = IgG2 [Sigma], Zg4 = IgG3, and RJ4 = IgG4 [both from Immunotech]) were used, at a final dilution of 1:2,000; 1:10,000; 1:10,000; and 1:1,000 respectively. For IgM (clone M11) determinations, the final dilution was 1:15,000. Each dilution of each monoclonal antibody had been previously determined as specifically reacting with the corresponding human isotype without showing significant cross-reaction with other isotypes [22,32,33].

All ELISA determinations were performed on coded serum samples, in a blinded manner—i.e., the investigators were not aware of the morbidity data of the corresponding patients.

The results were expressed first as ratios of OD values (or arbitrary units) as previously described [22,32,33] which were then used to estimate immunoglobulin concentrations. To calculate the OD ratios, each ELISA plate included sera from seven healthy French non-malaria-exposed blood donors, chosen as being representative of results obtained previously using sera from 200 French blood donors without exposure to malaria, i.e., yielding the same mean OD ± standard deviation (SD) value. The OD values corresponding to background responses from blood donors never exposed to malaria calculated in this manner were usually low, eg for MSP3: 0.035 (for IgG1); 0.050 (for IgG2); 0.027 (for IgG3); 0.031 (for IgG4); and 0.112 (for IgM). However, the precise cut-off value was calculated for each ELISA plate using the OD values recorded using the seven negative control sera included in each plate. The same positive and negative sera were systematically included in each 96 wells ELISA plate throughout the entire study. For the evaluation of immunoglobulin concentrations, the CLB reference serum with known amounts of each IgG subclass, IgM, and IgA (Central Laboratory of the Netherlands Red Cross blood transfusion service, Amsterdam, The Netherlands), and human reagents (either IgG myelomas or purified human IgM from Sigma) were used. The OD values obtained with our secondary monoclonal antibodies over a range of dilutions of the different reagents were used to determine the actual amounts of each immunoglobulin specific for each antigen present, and were converted to μg/ml equivalents.

Analysis of Sequence Polymorphism in the msp3 Gene

MSP3 PCR and sequencing focussed on the original DG 210 clone sequence corresponding to the C-terminal domain containing the three B cell epitope regions a, b, and c, as well as three T helper cell epitopes [23]. Forty-five P. falciparum isolates and strains were studied, obtained as follows: (i) ten from Kanbauk, Myanmar, Southeast Asia, where transmission is seasonal with between two and ten infective bites per year; (ii) 15 from Dielmo, Senegal, where transmission is holoendemic with 200 infective bites per year [16]; iii) ten from Ariquemes, State of Rondônia, Brazil, where transmission is perennial with 6.6 infective bites per month, and (iv) 13 laboratory strains—four from Tanzania, one from either Uganda, Liberia, Senegal, Thailand, Myanmar, Papua New Guinea, India, Brazil or Honduras and two from South East Asia. Parasite DNA from P. falciparum laboratory strains was extracted as described earlier [34]. DNA was extracted from blood samples using the QIAamp Blood Kit (Qiagen) in accordance with the manufacturer's instructions.

PCR and sequencing on parasite DNA was performed with primers DA151 (5′-G CCG GAA TTC CAT GAA AGG GCA AAA AAT GCT TA-3′) (nucleotides 562 to 585) (EcoRI cleavage site underlined) and DA152 (5′-G CCG GGA TCC ATT TTC CTT AGA TAT ATT TTC C-3′) (nucleotides 770 to 748) (BamHI cleavage site underlined). The program for PCR was a denaturation step at 94 °C for 1 min, an annealing step at 65 °C for 1 min followed by an extension step at 72 °C for 2 min and repeated for 35 cycles. PCR was performed using the GeneAmp PCR Kit (Perkin Elmer). The amplified products were purified by Spin-X (Costar) before sequencing on the ABI PRISM 373A DNA Sequencer using the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer). Nucleotide sequences were aligned by the CLUSTAL V program.

Statistical Analysis

Data from 217 of the 247 Dielmo inhabitants were used, corresponding to those from whom serum samples had been collected. Thirty individuals, mostly adults, were excluded from the 247 because they spent less than half of their time in Dielmo during the follow-up, and were therefore deemed unsuitable for inclusion given that the degree of exposure plays a crucial role in the development and persistence of protective immunity. Of note, women who were pregnant for more than 10% of the study period were not included in the analysis and will be the subject of a separate report. A period of two years—one year before and one year after the serum sampling—was chosen for the initial clinical analysis, during which the number of malaria attacks was recorded. The villagers who were tested during a longer period of time, i.e., up to 6 y, corresponded to a subset of children and adults who stayed for more than 75% of their life in the village. The time spent in the village was determined on a continuous, active, and daily survey carried out by the medical team.

Information pertaining to the IgG1, IgG2, IgG3, IgG4, IgM, and IgA antibodies to MSP3, MSP1, MSP2 (FCR3), MSP2 (3D7), AMA1, and RESA; and IgG and IgM levels to LSA1-R, LSA3-RE, CS protein (R32LR recombinant protein), were thus available, as well as age, sex, number of days spent in the village, G6PD profile, and haemoglobin AS phenotype.

An inspection of the number of malaria attacks (identified over 2 y) plotted against the age of the individual affected, showed that they were not linearly distributed. A maximum-likelihood method was developed for adjustment of the data corresponding to malaria attacks. A nonlinear model was fitted to using a negative binomial loss function to calculate the breakpoint of threshold of age-dependent acquisition of immunity, for men and women separately in a manner similar to that described by Kocherlakota and Kocherlakota [35] (developed by K. Dietz, Department of Biomedical Biometry, University Eberhard-Karls of Tübingen, Tübingen, Germany). Even though a Poisson distribution also fitted the data, in the end a negative binomial was chosen, because it is frequently used to describe this type of data. Further assumptions were that every individual has a specific risk for a malaria attack. The distribution for the risk is, as conventionally used, a gamma distribution. The gamma distribution was standardized and had a mean of 1. The bivariate negative binomial distribution for both years had three parameters: a term for the variances (1/k) for the gamma distribution and the mean values for the two different years. The real risk of malaria in the different year is multiplicatory—and was given by a product term that consisted of an individual term and a year specific term, the latter being the same for all individuals. It was assumed that, for a particular individual, the number of attacks in the subsequent years was independently distributed according to a Poisson distribution with different means, which were proportional to the individual exposure risk.

The statistical models used for analysis of data were multivariate models. The potential relationship between malaria attacks (the variable to explain) and selected explanatory variables was tested using a log (1 + x) transformation of the number of attacks. The explanatory variables were selected on the basis of our existing knowledge of the potential involvement of these different parameters (with reference to our previous observations in different malaria-endemic areas). Age and age², sex, haemoglobin phenotype and G6PD deficit as well as the individual immunoglobulin isotypes (i.e., IgG1, IgG2, IgG3, and IgG4) that were available and suspected of having a potential impact on the dependent variable (the number of malaria attacks), were tested in backward stepwise regression, eliminating, by hand, the variables which were not significant. The criterion for inclusion or elimination of the explanatory variables was that a predictor variable was included when its partial regression coefficient was significant at the 0.05 level and eliminated when its partial regression coefficient failed to be significant at the 0.1 level. The assumption that the regression equation accurately summarized the data was visually verified by examination of the residuals plotted against the fitted values. The significant terms were retained for further analysis and retested again to avoid exclusion due to interactions.

A negative binomial distribution is conventionally accepted for most disease data and has therefore been used to test the significant variables. Nominal and ordinal logistic regression (with no attack, one malaria attack, and two malaria attacks for the adults) and multivariate regression using a negative binomial distribution were used to analyse data with JMP software (SAS Institute, Cary, NC). For this analysis, IgG1 to IgG4 values were log-transformed or dichotomised as indicated in the ELISA section. A nonlinear model was also fitted to the significant variables of interest used to establish indications of long-term protection (between 1 and 6 y after the blood samples were tested) in children and adults of Dielmo. Chi-square and significance values were calculated looking at the difference between the full model and the model without the respective term. We used a level of significance of 0.05.

Characterization of the Study Population

Of the 247 individuals constituting the whole village population of Dielmo, 217 were enrolled for the present study. The study cohort included 102 females (mean age ± SD 25.9 ± 21.1 y) and 115 males (21.9 ± 18.5 y). During the year preceding the blood sampling, the inhabitants of Dielmo involved in this study were present in the village during 78.2% ± 27.6% of the time (95% confidence interval [CI] 74.5%–82.0%). At the time of blood sampling, 112 inhabitants (mean age ± SD 30.5 ± 18.2 y) had a negative thick smear, whereas blood parasitaemia was detected in 55 individuals (21.8 ± 16.6 y) with less than 5,000 parasites/μl, and in 50 individuals (10.6 ± 10.7 y) who had ≥ 5,000 parasites/μl. In total, 130 villagers (30.8 ± 18.2 y) never presented with a malaria attack during the 2 y of follow-up and had a mean parasitaemia of 3,251 ± 10,841 parasites/μl (95% CI 1,553–4,949) at the time of sampling, whereas 87 (mean age ± SD 14.4 ± 18.3 y) had at least one malaria attack with a mean parasitaemia of 41,814 ± 191,964 parasites/μl (95% CI 0–88,638) at the time of the attack. The haemoglobin phenotype was AA in 191 individuals, AS in 23, and AC in 3; 46 villagers had a G6PD deficit.

Identification of Clinical Cases and Evidence of Rapidly Acquired Protection in Young Children

The relevance of the analysis of immune responses critically depends on the reliability of clinical data. The daily clinical survey was satisfactorily carried out on 217 of the 247 individuals constituting the whole village population, who showed a classical age-dependent acquisition of clinical and parasitological immunity (Figure 1, shaded area). However, it was possible to distinguish, among the children, a subset of 14 individuals (20.9%) who did not present with an attack of malaria during the 2 y of follow-up (4/35 and 10/32 in the 0–5 y and 6–10 y age groups, respectively). Conversely, the study design also led to the identification of a small subset of 14 adolescents and adults (10.8%) occasionally susceptible to malaria attacks (3/28 and 11/101 in the 15–20 y and >21 y age groups, respectively), though the majority had acquired protection. However, in adults, symptoms were of short duration and resolved spontaneously, i.e., without requiring treatment in most cases [7,16,36].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Means and Standard Deviations of Anti-MSP3-b IgG1, IgG2, IgG3, IgG4, and IgM Antibody Responses per Age Group

Concentrations of antibodies are estimated as described in the Methods section. For each age group the mean number of malaria attacks recorded during the first year of follow-up is indicated by the shaded area. The numbers of individuals in each age group were: 5 (0–1 y), 14 (2–3 y), 17 (4–5 y), 14 (6–7 y), 12 (8–9 y), 13 (10–11 y), 11 (12–14 y), 18 (15–18 y), 13 19–21 y), 100 (>21 y), respectively.

https://doi.org/10.1371/journal.pmed.0040320.g001

The existence of such rapidly acquired protection in very young children and, conversely, the occurrence of brief malaria attacks in adults, have seldom been reported previously, possibly due to the use of less-stringent surveillance setups. The pattern of clinical incidence in Dielmo remains similar to that described in other high-endemicity African areas [37] with a far greater number of attacks in younger individuals, e.g., 3.45 clinical attacks per year in children aged 1–5 y versus 0.1 attack per year in adults. Nevertheless, the daily surveillance of highly exposed villagers made it possible to distinguish in every age group individuals with or without malaria attacks who can therefore be referred to below as “nonprotected” and “protected,” respectively, for the duration of the 2 y follow-up period. The monthly surveys of entomological inoculations, conducted during the clinical follow-up period, showed a very high level of transmission, with an estimate of 260 infective mosquito bites per person per year, i.e., 520 parasite inoculations over the 2 y survey, with little house to house variation, leading us to exclude the hypothesis that villagers without malaria attacks were not exposed.

This clinical situation was in most instances stable since only 13% of participants changed clinical status from the first to the second year of follow-up. The cases recorded are therefore well suited to the identification of immune correlates of protection and, furthermore, for the first time this analysis can be performed in an age-independent manner.

Cytophilic Anti-MSP3 Antibodies Correlate with Clinical Immunity

The overall prevalence of IgG and IgM antibodies against the MSP3-b epitope in the 217 individuals studied was high (97.2% and 93.1%, respectively), whereas the prevalence of IgA was low (19.3%). MSP3-specific IgE antibodies could not be detected (unpublished data). Among the IgGs, the two cytophilic classes, IgG1 and IgG3, were the most abundant (Figure 1) and they both increased as a function of age, whereas the increase was modest for IgG2 and IgM antibodies, and low for IgG4 antibodies in all age groups.

Responses to other malaria antigens, including MSP1, MSP2-FC27, MSP2-3D7, RESA, and AMA-1, were also found to increase with age in this study (unpublished data) as has been observed in previous studies [38–42]. Therefore, due to this age-dependent increase, all antibodies measured showed an overall inverse relationship with the prevalence of clinical attacks (Figure 1, shaded area), a phenomenon reported in most previous studies in African settings and sometimes taken as indicating that protection is afforded by those antibodies [33,38,43–45]. The prevalence of responses to pre-erythrocytic antigens, such as CS, LSA1, and LSA3, also showed an age-dependent increase and confirmed the high degree of exposure to infected mosquitoes (see Figure S1). Antibodies against the pre-erythrocytic antigens were not included in the present statistical analysis.

We first sought by stepwise regression analysis if it was possible to select a subset of antibody responses that would tend to predict the number of malaria attacks when controlling for age. We then examined the predictive value in terms of clinical protection of a positive antibody response for each IgG isotype, to each of the five molecules studied, for both each year separately and combined. A highly consistent association was observed between protection and anti-MSP3 IgG3 antibodies for each of the years studied, as well as for the two years combined, and this indication persisted even when controlling for age (age-adjusted odds ratio [OR] and 95% CI = 7.19 [2.70–22.85], p < 0.001 for the 2 y of follow-up), whereas no similar association was found, when age was accounted for, for any of the other isotypes, nor for any of the various antibody responses to the other vaccine candidates (see Table S1). For example, for 2 y of survey, age-adjusted ORs and 95% CIs were, for anti-MSP1 IgG1, 0.36 (0.076–12.36), p = 0.134; for anti-MSP-1 IgG3, 1.38 (0.60–3.15), p = 0.437; for anti-AMA1 IgG3, 1.01 (0.41–2.40), p = 0.975; for anti-RESA IgG3, 1.18 (0.59–2.35), p = 0.637; for anti-MSP2 (FC27) IgG3, 0.74 (0.135–3.952), p = 0.727; and for anti-MSP2 (3D7), 0.81 (0.13–49), p = 0.823. The protected group had mean anti-MSP3 IgG3 values 3.75 times and 4.3 times higher than individuals having experienced either one or two or more attacks, respectively. The malaria attacks recorded over 2 y among the 177 individuals with anti-MSP3 IgG3 were 2.52-fold less frequent than among the 40 participants without detectable anti-MSP3 IgG3 responses.

We have reported that it is the relative proportion of cytophilic antibodies to noncytophilic antibodies (the C:NC ratio) that is the most important surrogate marker of protection to date [32,46]. This is because anti-MSP3 antibodies act in a monocyte-dependent manner, requiring binding to Fc-γ receptors on monocytes [6,32]. The C:NC ratio takes into account the possible, and demonstrated, competition of noncytophilic antibodies directed towards the same epitope [15,32]. We found a consistently higher C:NC ratio, that is, a quantitative dominance of cytophilic classes, among protected as compared to nonprotected participants. This higher ratio was found in each age group (Figure 2), though it reached statistical significance only in the two youngest and the oldest age groups (0–5 y, p < 0.01; 6–10 y, p < 0.03; and >20, p < 0.007), probably due to group size. The overall C:NC ratio was significantly associated with protection for each year studied (OR [95% CI] = 2.79 [1.50–5.22], p = 0.001 for the year before sampling and 2.82 [1.51–5.30], p = 0.0012 for the year following the sampling), and for the two years combined (age-adjusted OR [95% CI] = 3.34 [1.72 – 6.67], p < 0.001). For the 2 y of survey, the other parameters were not significant, e.g., age-adjusted ORs (95% CIs) for the C:NC ratios were, for MSP1, 1.08 (0.57–2.11), p = 0.81; for AMA-1, 0.83 (0.43–1.61), p = 0.598; for RESA, 0.54 (0.13–2.04), p = 0.387; for MSP2 (FC27), 1.20 (0.43–3.38), p = 0.723; and for MSP2 (3D7), 1.35 (0.50–2.33), p = 0.551.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Ratios of Cytophilic to Noncytophilic Anti-MSP3 Responses Found Among Protected and Nonprotected Individuals of Dielmo

The mean ratios of (IgG1 + IgG3) to (IgG2+ IgG4+ IgM) antibody responses (i.e., the cytophilic to noncytophilic [C:NC] ratios) were calculated for protected (open circles) and unprotected individuals (closed circles) within each age group. Error bars indicate the SD calculated for each age group. The number of individuals within each group is indicated in parentheses. A protected person is defined as an individual in whom no malaria attack was recorded during the first year of follow-up.

https://doi.org/10.1371/journal.pmed.0040320.g002

IgG3 Anti-MSP3 Antibodies Are Strongly Associated with Clinical Protection against Malaria

Improved statistical analyses necessitated the use of larger group sizes. Since there was a clear age-dependent decrease in the number of attacks in children, whereas in adults the number of attacks was no longer age-dependent, we next determined the age threshold that informed the separation of individuals into two groups according to this criterion: for women this was 13.4 y and for men 17.9 y, over the 2 y observation period. These two groups were thus analysed in detail separately.

Among the 121 individuals of the “adults” group, analyses of data by ordinal logistic regression showed that only MSP3 IgG3 and the C:NC ratio of MSP3 were associated with protection, i.e., negatively correlated with the number of malaria attacks (L-R Chi-square = 6.91; p = 0.0086 and L-R Chi-square = 10.84; p = 0.001 respectively). MSP1 IgG4 and AMA1 IgG3 were marginally significant if tested separately (p = 0.04 and 0.05 respectively), but were not significant if included in a model with MSP3 IgG3. No other variable showed any significance (unpublished data). In a nominal logistic regression, for MSP3 IgG3 the OR (95% CI) was 23.7 (3.15–238.36); and for the C:NC MSP3 ratio, it was 71.48 (3.65–1,339.68). The ORs for all other isotypes and other molecules did not reach significance. When the results were dichotomised, only MSP3 IgG3 (p = 0.022) and the C:NC ratio of MSP3 (p = 0.01) showed a significant relationship with protection. The surrogate value of anti-MSP3 IgG3 is graphically indicated in Figure 3, in which the risk of occurrence of one or two or more malaria attacks is shown to decrease in adults when the anti-MSP3 IgG3 levels increase.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. The Probability of No, One, or Two Malaria Attacks in the Adults of Dielmo as a Function of the Level of Anti-MSP3 IgG3 Responses

The chance of no attack increased markedly and almost reached a maximum in parallel with the gradual increase in anti-MSP3b IgG3 responses. The probability that one or two malaria attacks occurred in adults decreased as the anti-MSP3b IgG3 responses increased.

https://doi.org/10.1371/journal.pmed.0040320.g003

Among the 96 individuals included in the “children” group, all variables were tested in a backward stepwise multivariate model with log-transformed isotype concentrations and then a negative binomial model fitted to the significant variables. The final model contained: (i) For IgG3: age (χ² = 55.46, p < 0.001), MSP3 IgG3, (χ² = 14.30; p < 0.001) and haemoglobin AS phenotype (χ² = 5.56, p = 0.018), or (ii) For C:NC ratio: age (χ² = 67.51; p < 0.001), MSP3 C:NC (χ² = 12.21; p < 0.001), and haemoglobin AS phenotype (χ² = 7.39; p = 0.0066). No other variable was found to be significant. For dichotomous antibody concentrations, MSP3 IgG3 (χ² = 7.46, p = 0.0063) and MSP3 C:CN (χ² = 4.036, p = 0.045) were also significant. The OR (95% CI), when tested in a nominal logistic regression model, was for age, 51.15 (6.71–515.37), p < 0.001; and for MSP3 IgG3, 15.49 (1.57–184.65), p = 0.0219 (for age 96.93 [12.16–1093.25], p < 0.001 and for MSP3 C:NC 64.19 [2.94–2,204.40], p = 0.0118). Results from an ordinal regression model with no attack, one, or more than one attack were similar to the results from the previous model except that haemoglobin AS phenotype was no longer significant.

The Clinical Significance of IgG3 Anti-MSP3 Antibodies Extends to Several Years

In view of the above results, which show a strong association of the surrogate marker with clinical protection, and since individuals in the village were followed up clinically in the ensuing years, we next addressed the question of the long-term clinical significance of a single antibody determination.

Individuals resident in Dielmo for more than 75% of the time in the ensuing 6 y after the initial blood sampling were selected for this analysis. For this reason, the number of individuals included in the 6 y analysis is lower than in the 2 y analysis.

In the adult group, the association of the IgG3 anti-MSP3 responses with protection was significant for the year following sampling (OR 14.1, p = 0.019) but no longer significant, in the ensuing years (as shown in Table S2). This indication was not found for anti-MSP1 IgG3, with OR (95% CI) of 7.27 (0.85–145.35), p = 0.111; for anti-AMA1 IgG3, 2.92 (0.19–303.21), p = 0.527; for anti-RESA IgG3, 2.67 (0.68–10.62), p = 0.157; nor for anti-MSP2 (FC27) IgG3, 0.78 (0.04–6.16), p = 0.832.

In contrast, in the children group, a significant association of IgG3 anti-MSP-3 antibodies with clinical protection was observed in the following 12, 24, and 36 mo of follow-up (respective ORs 56.1, p = 0.0063; 34.8, p = 0.01, and 31.7, p = 0.009). Only by year 4 and onwards did the clinical value decrease, though it remained significant (OR, 15.49, p = 0.037; 18.2, p = 0.031, and 21, p = 0.036 on years 4, 5, and 6, respectively, as shown in Table S2). Of note, a similar analysis did not lead to the same observations for antibodies specific to other antigens (for example, after 1 y of follow-up, it was found in the children group, for anti-MSP1 IgG3, 1.30 [0.11–29.67], p = 0.84; for anti-AMA1 IgG3, 0.93 [0.18–4.9], p = 0.931; for anti-RESA IgG3, 1.89 [0.52–7.75], p = 0.345; for anti-MSP2 (FC27) IgG3, 0.20 [0.015–1.918], p = 0.183; and for anti-MSP2 (3D7), 0.71 [0.05–9.14], p = 0.785).

The above results were confirmed using a nonlinear model fitted to the variables demonstrated to be significant in the previous analysis, i.e., age, haemoglobin AS phenotype, and anti-MSP-3 IgG3. The maximum difference between the mean number of malaria attacks identified in children with anti-MSP-3 IgG3, and children without such a response (4.3-fold), was detected 3 y after sampling (χ² = 11.76; p = 0.0006) and remained marked by year 4 of the study (χ² = 9.76; p = 0.0017). The differential number of attacks was still significant by year 6 of the follow-up (χ² = 4.64; p = 0.031) when the model was compared with and without the antibody response. Hence, the two methods of analysis provided convergent conclusions indicative of the long-term clinical significance of a single IgG3 anti-MSP3b determination, in terms of the resistance of children to malaria attacks.

Results are graphically illustrated in Figure 4, which shows the cumulative number of malaria attacks over 6 y of follow-up in Dielmo, in relation to IgG3 antibodies to each antigen. When the analysis was performed comparing the same individuals with either “low” or “high” antibody levels (IgG3 levels below or above the median), similar conclusions were reached (see Figure S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Mean Cumulative Number of Malaria Attacks Identified over Six Years of Follow-up in Dielmo, in Relation to IgG3 Antibodies to Each Antigen

For each antigen, specific IgG3 responses were dichotomized (antibody ratio less than or greater than 1) and in each group the occurrence of malaria attacks was calculated. To avoid the confounding effect of age, a subgroup of 49 children was selected for this analysis so that the mean age was similar (5.5 ± 2.6 y) among children with or without antibodies to each antigen. The mean cumulative numbers of malaria attacks are indicated for anti-MSP3b (A), anti-MSP1 (B), anti-AMA1 (C), and anti-MSP2-3D7 (D) IgG3 responses with ratio values less than 1 (open circles) or greater than 1 (closed circles). Error bars indicate SD.

https://doi.org/10.1371/journal.pmed.0040320.g004

The Target Epitope of Antibodies Associated with Protection Is Fully Conserved in Various Parasite Isolates

Since MSP3 appeared at this point to be a strong surrogate marker of clinical protection, we deemed it important to investigate whether this antigen exhibits the polymorphisms reported for many malarial vaccine candidates [47–50]. The sequence of the C terminus part of the gene, encoding three B cell and four Th cell epitopes [23] was analyzed in 45 parasite strains and isolates and did not reveal any polymorphism, either at the amino acid level or at the nucleotide level (unpublished data). The importance of this result is supported by the parallel study of the nonrepetitive 3′ region of the CS gene (nucleotides 1,015 to 1,254), which revealed 16 nucleotide substitutions in the same 45 strains (ten within the Th2R and six within the Th3R regions), all resulting in amino acid substitutions, in agreement with previous reports [51].

Accession Numbers

The primer positions in the MSP3seqence are based on GenBank (http://www.ncbi.nlm.nih.gov/) accession number AF024624.