Antigenic polymorphism is considered a significant confounder in the development of antibody-mediated protection against blood stage P. falciparum in the context of naturally acquired immunity and malaria vaccine development. The goal of this study was to determine whether variant-specific antibodies to MSP-119 were associated with haplotype-specific protection in a cohort of Kenyan adults and children who participated in a treatment time-to-infection study. These data showed that i) baseline infecting MSP-119 haplotype had no effect on the subsequent infecting haplotypes; ii) variant-specific IgG antibodies measured serologically and functionally had no association with follow-up infecting haplotypes or density; iii) variant-specific antibodies correlated with delayed time-to-infection among children but not adults; and, iv) variant-specific antibodies were associated with protection in a haplotype-transcending manner. Considered together, these data found no evidence for haplotype-specific immunity to MSP-119 in this study of naturally infected individuals living in a malaria holoendemic region.
EKNG and QKNG were the most prevalent MSP-119 haplotypes in this population and region of western Kenya in 2003 when this study was conducted. From other surveys, it has been found that this distribution of MSP-119 haplotype distribution pattern is stable (Yeo, unpublished). Takala et al. found comparable results with EKNG and QKNG being the most prevalent haplotypes in Mali from 1999 to 2001
. Others have shown similar dominances of EKNG and QKNG in Kenya, Brazil, Vietnam, Thailand, Tanzania and Vanuatu
It was found that within a single individual, infecting haplotypes frequently changed from week to week. This could be due in part to sampling effect. A finger-prick blood sample does not accurately reflect total body haplotype prevalence or parasitaemia. Additionally, this study showed that with low parasitaemia haplotype detection may not be optimal (Figure
8). In several studies, parasite densities and parasite genotypes varied significantly within a 24-hour time period and over days to weeks
[37–40]. Furthermore, venous blood may reflect different densities than finger-prick blood as the latter would be expected to have a greater frequency of capillary-sequestered infected erythrocytes. Examining weekly infections may heighten the variability detected, but with repeated measures in 176 participants over a 12-week study period, overall trends should persist. To this end, it was found that haplotype complexity was associated with P. falciparum density and possibly symptomatic infections. This contrasts with others’ finding that complexity of infection was associated with increased age and decreased frequency of symptomatic infection
. An important point in this regard is that children had higher parasite densities than adults, and this most likely led to better detection of multiple haplotypes.
Analysis of serologic responses was performed using recombinant MSP-119 (four variants) and MSP-142 (three variants). Variability in protein folding and expression systems used to produce these products, e g, yeast and E. coli, and serology techniques (traditional ELISA vs Luminex® multiplex) could account for differences in determining antibody responders vs non-responders. Nevertheless, these data indicated that both approaches produced a similar overall result — no discernible variant-specific immune correlation, consistent with the notion that variant-specific antibodies cross react with heterologous variants. Using immunodepletion assays, Zakeri et al. found evidence of antibody cross-reactivity among several MSP-119 variants, consistent with these findings
A significant limitation to the approach of detecting the infecting haplotype is the assumption made to assign two haplotypes to an infection that contained all four alleles. Although in most cases a predominant haplotype could be differentiated from a minor haplotype with the MFI of each detected allele, it is not certain that only two haplotypes were present
. For example, if an infection composed of all four alleles had higher MFIs for Q > E and KNG > TSR, the haplotypes would be assigned as QKNG and ETSR. However, it is possible that the individual was actually infected with QKNG, EKNG and ETSR. The only way to definitively determine this would be to have a larger blood volume and clone and sequence multiple PCR products, an approach which was not feasible for this study. Other methods such as pyrosequencing are advantageous in that direct sequencing of amplicons is possible.
This study had limited power to detect associations between antibody responses and infecting haplotypes in the context of susceptibility to symptomatic malaria. The 18 individuals who developed clinical malaria during the follow-up period did not have a discernible variant-specific antibody pattern. They did, however, lack MSP-119 IIA antibodies, which has previously been shown to increase with haplotype-specific (ETSR) infection
. Interestingly, 15 of these symptomatic individuals had P. falciparum detectable by PCR at least one week after treatment. Although this study was not designed to examine the efficacy of Coartem® treatment, previous observations demonstrate that parasites are cleared from the blood within 48 hours
. Although the possibility that detection of P. falciparum after treatment resulted from residual P. falciparum DNA cannot be excluded, it is most likely that detection resulted from the progression of pre-existing liver stage P. falciparum to the blood stage as Coartem® does not eliminate the former. Inadequate adherence to treatment regimen and/or lack of food intake with medication consumption could also result in incomplete parasite clearance
[42, 43], but is unlikely with this study as all six doses of Coartem were directly observed by project staff. All but one participant with symptomatic malaria during the follow-up period had infections containing three or more MSP-119 alleles. Increased COI may be associated with increased risk of symptomatic malaria, as has been observed previously
[44, 45]. Malaria transmission intensity and seasonality may also affect COI, but this study was not designed or powered to detect this association.