Introduction

Plasmodium falciparum parasites that accumulate in the placenta express unique variant surface antigens (VSA) including the protein VAR2CSA, which enables the parasitised erythrocyte to adhere to placental chondroitin sulphate-A (CSA) [1]. IgG antibodies towards CSA-VSA (CSA-binding phenotype VSAs) are gravidity dependent and have been associated with protection from malaria-related low birth weight and maternal anaemia [1]. The development of these antibodies may depend on a combination of factors, including maternal genetics, HIV status and parasite exposure, which in turn could vary with the use of intermittent preventive treatment in pregnancy (IPTp) and bed nets. Published data suggests IPTp may decrease CSA-VSA IgG acquisition [2], [3] and that this may be dependent on the presence of HIV infection [2]. The effect of bed net use on development of these antibodies is unknown. Antibody development may be further affected by changes in transmission intensity that decrease exposure to the antigen; and declines in pregnancy malaria prevalence have been recently reported [4]. To examine the effect of bed nets and IPTp on CSA-VSA IgG levels we measured antibody towards a CSA-binding parasite line during pregnancy and post partum in pregnant Malawian women participating in a randomised clinical trial of different IPTp regimes.

Materials and Methods

Results

CSA-VSA IgG was measured in 538 women at enrolment (14–26 gw) and 364 women at 1 mpp; 355 women had samples assayed at both time points. CSA-VSA IgG at enrolment and 1 mpp was positively associated with gravidity (Table 1). The mean (SD) age of the participants was 24.9 (6.7) years. Neither participants' age, gravidity, HIV nor socioeconomic status varied significantly with bed net use or IPTp regime; bed net use and ITN regime were also not associated with each other.

Download:

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

Table 1. Association between specified factors and IgG to CS2-VSA, multivariate analysis.

https://doi.org/10.1371/journal.pone.0029874.t001

At enrolment, bed net use (prevalence 63%) and HIV infection (prevalence 14%) were associated with lower CSA-VSA IgG. Concurrent P. falciparum parasitemia (prevalence 9%) was associated with a non-significantly higher CSA-VSA IgG level (Table 1). When the data were stratified by gravidity, there was a significant negative association between bed net use and CSA-VSA IgG at enrolment in primigravidae (coefficient −0.672; 95%CI −1.245, −0.099, P = 0.022), but not in multigravidae (coefficient −0.307; 95% CI −0.632, 0.018 P = 0.064) or secundigravidae (coefficient −0.588; 95% CI −1.442, 0.266 P = 0.18). HIV infection was associated with non-significantly lower CSA-VSA IgG at enrolment in primigravidae (coefficient;−0.709 95% CI −1.869, 0.276 P = 0.14) and multigravidae (coefficient −0.404; 95% −0.826, 0.016 P = 0.059) but not secundigravidae (coefficient −0.060; 95% −1.237, 1.118 P = 0.92).

At 1 mpp, among all women tested, levels of CSA-VSA IgG were non-significantly lower in women who received SP+AZI compared to women who received two-dose SP (Table 1), and CSA-VSA IgG levels in women who received monthly SP were intermediate (compared to the other two regimes). When stratified by gravidity, there were significantly lower levels in primigravidae who received SP+AZI than those who received two-dose SP (coefficient −1.447; 95% CI −2.800, −0.094 P = 0.036); levels in primigravidae who received monthly SP were intermediate (coefficient −1.189; 95% CI −2.613, 0.233 P = 0.1) (Figure 1A). Secundigravidae who received SP+AZI had non-significantly lower levels compared two-dose SP (coefficient −0.989; 95% CI −2.213, 0.235 P = 0.11); and similar levels to secundigravidae who received monthly SP (coefficient −0.371; 95% CI −1.634, 0.893 P = 0.56). In contrast to primigravidae, multigravidae who received two-dose SP had similar CSA-VSA IgG levels to those who had monthly SP (coefficient −0.088; 95% CI −0.729, 0.552 P = 0.79) and to those who had received SP+AZI (coefficient 0.118; 95% CI −0.495, 0.731 P = 0.71) (Figure 1B).

Download:

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

Figure 1. Relationship between antibody to CS2 variant surface antigens and IPTp regime, stratified by gravidity.

https://doi.org/10.1371/journal.pone.0029874.g001

Because HIV infection has been associated with decreased antibody response to IPTp, we assessed whether the relationships between both IPTp and bed net use and antibody development were also seen in HIV negative women. Among HIV negative women who received SP+AZI, only primigravidae had significantly lower CSA-VSA IgG than those who received two-dose SP (coefficient-1.400; 95%CI −2.769- −0.032, P = 0.045). HIV negative women who used bed nets had significantly lower CSA-VSA IgG at enrolment compared to those who did not use bed nets (coefficient −0.501; 95%CI −0.802, −0.200, P = 0.001), and this did not appear to be explained by differences in schooling or literacy (data not shown). Due to the low HIV prevalence, HIV positive women were not separately assessed for the effects of IPTp regime or bed net use on antibody levels.

Discussion

We examined the association between CSA-VSA IgG antibodies and use of bed nets and IPTp in the context of a randomised clinical trial in pregnant women. We found that bed net use and intensive IPTp regimes were associated with lower levels of these antibodies. These associations were not dependent on the presence of HIV infection, but the association between IPTp regime and antibody was clearly dependent on gravidity. At antenatal booking, antibody levels were lower in women who used bed nets or who were HIV positive. At 1 mpp the most intensive IPTp regime was associated with decreased antibody in primigravidae. The observation that the difference was between the SP+AZI and two-dose SP regimes, not two-dose SP and monthly SP regimes, suggests that it is the drug AZI which accounts for the lower CSA-VSA IgG levels. Interestingly, a high frequency (90%) of mutations associated with SP resistance has been found in P. falciparum infections in Malawi [6].

That bed net use is associated with decreased antibody is a novel finding, and may indicate that bed nets decreased exposure to CSA-binding parasites before antenatal clinic attendance. Whereas most antimalarial drugs are not recommended in the first trimester, using bed nets from conception may decrease parasite exposure in early pregnancy. We found that bed net use was not associated with antibody at 1 mpp, while the monthly SP+AZI regime was, suggesting that intensive IPTp, which may both clear existing infection and inhibit new infection, may be a more important determinant than bed net use of the development of antibody in the latter part of pregnancy.

Previous studies have yielded conflicting findings on the relationship between IPTp and pregnancy-specific immunity [2], [3]. Staalsoe et al showed that CSA-VSA IgG levels were decreased among primigravidae receiving SP over placebo. In Mozambique, women of all gravidities did not show decreased CSA-VSA IgG with two doses of SP-IPTp, unless they were HIV positive [2]. HIV infected women may require more than two doses of SP to be protected from pregnancy malaria [7]. In the present study, the association between IPTp and antibody development was not restricted to HIV positive women. Differences between studies may be due to variation in SP resistance, both Mozambique and Malawi have SP resistant parasites [8], [9]. Given the recommendation that all pregnant women in areas like Lungwena receive IPTp [10], we had no placebo group. Instead, we observed a clear relationship between more intensive IPTp regimes and antibody levels in primigravidae, suggesting that the more intensive regimes decrease exposure to the antigen, decreasing antibody development. In keeping with this, we observed a lower prevalence of P. falciparum infection (detected by microscopy of Giemsa stained peripheral blood smears) at 32 weeks and at delivery in women receiving monthly SP ± AZI (2% at each time point) than in women receiving two-dose SP (5% at each time point) [6]. Parasite prevalence at delivery estimated by quantitative PCR was also significantly higher in women receiving two-dose SP (Luntamo et al, in preparation).

The observation that IPTp use was only associated with antibody levels in primigravidae may reflect higher existing levels of immunity in multigravidae, such that decreased antigen exposure with intensive IPTp regimes had minimal effects on immunity. The protective effect of IPTp may also be more modest in women with existing immunity, and these women, usually multigravidae, may be more capable of boosting antibody levels [5] in response to sub microscopic infections, which were relatively common [11].

At enrolment, history of bed net use was associated with decreased antibody in both primigravidae and multigravidae, although the association only reached statistical significance in primigravidae. By contrast, IPTp regime was only associated with antibody level at delivery in primigravidae. This difference in associations between bed nets and IPTp with antibody may be due to the total time that the intervention can decrease antigen exposure (as bed net use might reflect decreased exposure over a prolonged period, not just the current pregnancy), or to the ability of bed nets to block boosting of antibodies by infection in early pregnancy.

Our study had a number of potential shortcomings. The low prevalence of HIV and parasitemia compromised our ability to identify relationships between these factors and antibody. We only measured antibody at two time points during pregnancy, but antibody levels may rapidly decrease following clearance of infection, and are in any case highly dynamic [5]. Being a community based study, it was usually not possible to document placental infection and as peripheral parasitemia was measured by microscopy [11] and at limited time points, parasitaemia prevalence was likely underestimated. (Only a subset of women in the present study also contributed to the comparison of qPCR and microscopy). In choosing samples from a sequential set of women for this study, the probability of selection bias was minimised, and the LAIS used population based sampling [6]. Socioeconomic factors that could indicate altered risk of P. falciparum exposure, potentially compromising analyses, were not associated with HIV or bed net use. We could, of course, only determine statistical associations, not causal relationships, between interventions and antibody development.

If bed nets and intensive IPTp decrease pregnancy-specific immunity, this may have important implications for the ongoing susceptibility of pregnant women to malaria. As coverage with bed nets and IPTp rises, more women may receive their benefits, but they may be increasingly vulnerable in subsequent pregnancies. The only study that examined this found that chemoprophylaxis in the third trimester did not affect malaria susceptibility in the subsequent pregnancy, but participants may have already experienced substantial exposure and developed immunity by this time [12]. If bed nets and IPTp cease to be widely available, many highly-susceptible pregnant women could be again exposed to high risk of pregnancy malaria, with potentially devastating consequences. Whether the relationships between bed nets and IPTp use and development of immunity are similar in higher transmission settings, is currently unknown. The effects of decreased exposure and decreased antibody on susceptibility to disease need to be properly addressed by longitudinal studies over successive pregnancies.

Acknowledgments

We thank the Australian Red Cross Blood Bank for the provision of human blood for P. falciparum culture. Most importantly, we also thank all the study participants.