Background
In 1990, histidine-rich protein 2 (HRP2) was identified as a potential target for diagnosis of
Plasmodium falciparum infections [
1]. Soon thereafter, the approach for diagnosing malaria by detecting HRP2 in blood was validated when a field trial in Thailand (1993) reported that an antigen-capture ELISA for HRP2 [
2] had 98% sensitivity and 96.2% specificity [
3]. Since then, numerous rapid diagnostic immuno-chromatographic tests (RDT) for detecting HRP2 have been developed and used worldwide as an alternative to microscopy for diagnosing malaria. The WHO initiative launched in 2012 for control of malaria, T3: Test, Treat, Track, relies heavily on the use of RDT for diagnosis of malaria [
4]. In addition, recent studies have measured the amount of HRP2 in blood as an index of whole-body parasite burden [
5,
6] and to estimate disease pathology [
7]. HRP2 has proven to be a valuable protein for diagnosis of malaria since it is produced by ring and trophozoite-stage parasites and secreted into plasma [
6,
8]. Because infected erythrocytes containing trophozoite-stage parasites sequester in deep vascular tissues, detection of soluble HRP2 in blood plasma allows for diagnosis of malaria in individuals who are slide-negative by microscopy.
Since rabbits and mice readily produce antibodies (Ab) when immunized with recombinant HRP2, it is assumed HRP2 is immunogenic in people infected with
P. falciparum. Accordingly, a number of authors have speculated that Ab to HRP2 might reduce the sensitivity of HRP2-based assays [
3,
9,
10], since Ab could complex with soluble or parasite-associated HRP2 and block its detection in antigen-capture assays. Surprisingly, few groups have studied Ab to HRP2 [
9‐
11]. In a drug-treatment study in India, Biswas et al. reported that soluble HRP2 and IgM Ab to HRP2 were present in
P. falciparum-infected patients at enrolment, that both HRP2 and IgM to HRP2 declined following drug treatment, and that anti-HRP2 IgG Ab were present on days 15 and 28 days after drug treatment [
10]. However, in a brief report, Das et al. tested plasma by ELISA from individuals living in a malaria-endemic region of India for Ab to HRP2 and found no difference in optical density (OD) values between 15 naive and 45 exposed individuals [
11]; however, the lack of a positive control for the assay complicated interpretation of the results. More recently, Ho et al. using a similar ELISA reported that 25% of plasma samples (n = 75) collected from patients with acute malaria in Cambodia, Nigeria and the Philippines had anti-HRP2 Ab [
9], with the majority of samples having OD values only one to three times above the cut-off. These results suggest that HRP2 is either not highly immunogenic in humans or the Ab response is short-lived.
Because of the potential for anti-HRP2 Ab to reduce the sensitivity of HRP2-based diagnostic assays, this study sought to establish the prevalence, class, subclass, affinity, and age distribution of Ab to HRP-2 in Cameroonian children and adults, aged 5 to 82 years of age, residing in a malaria-endemic region with high P. falciparum transmission. For comparison, Ab to the merozoite antigens MSP1, MSP2, MSP3 and the pregnancy-associated antigen VAR2CSA were measured using a bead-based multiplex assay. The study design tested the hypothesis that if Ab to HRP2 were produced by an antigen-specific acquired humoral immune response, then Ab would be present in individuals who had been exposed but not in unexposed individuals, and that the Ab response would be boosted by re-exposure. Plasma samples from 181 Cameroonian children and adults (45% of whom were slide-positive from malaria) were tested and compared with data from 112 Americans who had not been exposed to malaria. Overall, results for MSP1, MSP2, MSP3 and VAR2CSA were as expected; however, there was no solid evidence that individuals living in a malaria-endemic area had developed an acquired humoral immune response to HRP2.
Discussion
HRP2 is a unique protein, both in amino acid sequence and protein structure. It is 65–85 kDa protein made up of 35% histidine, 40% alanine and 12% aspartate [
19,
23] that consists primarily of the repeat sequences AHH and AHHAAD [
19,
23]. Once synthesized, HRP2 is transported via vesicle to the infected erythrocyte membrane, where [3H]-labelling studies show that galactose is attached either during, or just before, HRP2 is released into plasma [
23]. Although the structure of HRP2 has been difficult to determine, data predict the repeat sequences form a 3/10 helix or a tightly coiled α-helix [
23]. Thus, HRP2 is very different from known protein antigens.
HRP2 was initially selected for diagnosis of falciparum malaria because it is highly conserved, synthesized by ring-stage parasites, and secreted into plasma. Lee et al. sequenced 448 global isolates and reported that the sequences DAHH
AAHAHH (recognized by mAb 2G12) and AHHAHHV (recognized by mAb 1E1) were present in 448/448 (100%) and 45/448 (99.3%) of the isolates, respectively, and that multiple copies of these repeat sequences were present (average 18.4 copies/isolate) [
16,
17]. The binding of monoclonal Ab to HRP2-coupled microspheres demonstrates that epitopes in the repeat sequence were correctly displayed. Since individuals infected with
P. falciparum are exposed to the repeat sequences of HRP2, they should have Ab to HRP2 if the protein is immunogenic.
The initial goal was to determine the prevalence, class, subclass, and avidity of Ab to HRP2 in different age groups living in a malaria-endemic region of Cameroon. The samples, collected between 1994 and 1999, were from a cross-sectional malaria prevalence study in Simbok village, at which time the estimated entomological inoculation rate was ~556 infectious bites annually [
12]. Study subjects had been infected with
P. falciparum multiple times. Based on the characteristic of acquired B cell responses, one would expect to find Ab to HRP2 in Cameroonian, but not American adults and that Ab prevalence, amount and avidity would increase with age. Interestingly, the results found no evidence of an acquired immune response to HRP2, since values for IgG to HRP2 were similar in exposed and unexposed individuals and IgG levels did not increase with age (Figs.
2,
3). The lack of IgG was not due to poor quality of the plasma samples, because the samples contained high levels of Ab to the other malarial antigens (MSP1, MSP2, MSP3) when tested simultaneously in the multiplex assay (Fig.
2). Because HRP2 has tandem repeats, it might function as a T-independent antigen. This possibility seemed unlikely since a related plasmodial parasite,
Plasmodium yoelii, lacks both type 1 or 2 T-independent antigens, i.e., infected athymic mice fail to produce Ab to any malarial antigen [
24]. The mice, however, had activated B-cells that did not secrete Ab in the absence of T cell help. Nonetheless, plasma from 81 Cameroonians who were blood-smear positive of
P. falciparum were tested for IgM Ab to HRP2 (Fig.
4). A few individuals had borderline-positive IgM values for HRP2 compared to the US negative controls but MFI were very low. Since total IgM levels are elevated during
P. falciparum infection due to antigen-specific and polyclonal B cell activation, it was difficult to know if values above the arbitrary cut-off were due to IgM Ab, non-specific binding, or cross-reactive Ab. Therefore, among the 112 American plasma samples, the six samples with the highest MFI for HRP2 were tested for IgG and IgM binding to
P. falciparum-infected erythrocytes by IFA (Fig.
6). None of the US samples tested positive, suggesting that the Americans did not have Ab that cross-reacted with
P. falciparum; rather, the samples simply had high backgrounds. The finding that antibody avidity was similar between Americans and Cameroonians for HRP2 (Fig.
6b) supports the conclusion that Cameroonians adults did not have an acquired immune response to HRP2, that is, a response that had undergone somatic hypermutation, affinity maturation and development of B cell memory.
Results from the current studies are similar to those reported previously. Using an IgG ELISA assay, Das et al. reported that OD of plasma samples were the same between malaria-naive individuals and those who had acute symptomatic infections, asymptomatic infections and uninfected adults residing in a malaria-endemic area in India [
11]. Biswas et al. used a similar ELISA with a cut-off based on replicates of a pool of five negative plasma samples and reported detecting IgM Ab to HRP2 in patients at enrolment. However, following drug treatment, IgM Ab quickly declined and IgG Ab to HRP2 were detected 15–28 days post treatment. Patients in the above study had asymptomatic infections, showing they had sufficient immunity to control their infections. Yet they tested positive for IgM at enrolment and only produced IgG several weeks after drug treatment, thus displaying characteristics of a primary Ab response, suggesting they lacked a memory B cell response to HRP2 prior to the ongoing infection. Ho et al. used a similar ELISA and their ‘raw data’ looks similar to that found in Fig.
2 of the current study. The biggest difference between the two studies is the method used to establish the cut-off for positivity. Ho et al. used a pool of six samples from non-malaria exposed individuals as a negative control and converted OD into arbitrary units based on a pool of five samples with the highest OD to HRP2 [
9]. They reported that 19/75 individuals with acute malaria in Cambodia, Nigeria and Philippines were Ab-positive and three/28 asymptomatic Solomon Island residents were positive. However, most positive samples had antibody levels only one to three times higher than the cut-off, with only two plasma samples having Ab levels fivefold and 16.7-fold above cut-off. In the current study, 112 malaria-naive control individuals were used, providing data for a wide range of malaria-naive individuals, including individuals with other infections (Fig.
2). Compared to MSP1, MSP2 and MSP3, a highly variable background level of reactivity was found for HRP2 in malaria-naive adults, with some individuals having 20-fold higher levels than others. Using plasma from a limited number of healthy controls to establish a cut-off for malaria-infected individuals with hyper-gammaglobinaemia may not be optimal. Overall, none of the above studies found high IgG antibody levels to HRP2 in people living in malaria-endemic regions. It is unlikely that IgG Ab to HRP2 significantly alter the sensitivity of HRP-2-based diagnostic assays.
The lack of Ab may help explain why HRP2-based assays work as well as they do. First, if Ab were present, soluble HRP2 would rapidly form immune complexes and quickly be cleared by the spleen. HRP2 is detectable in plasma for >28 days after parasite elimination, suggesting that the protein circulates freely in plasma and is slowly removed by glomerular filtration and proteolysis. Second, the detection thresholds in RDTs is very low, for example Marquat et al. reported that the lower limit of detection for four RDTs was between 6.9 and 27.8 ng/ml [
25], an amount of antigen that would be neutralized by microgram levels of Ab. Finally, absence of Ab could explain why an association exists between parasite burden and disease severity, i.e., the more parasites, the more HRP2 in circulation. The uniqueness of HRP2 and its apparent poor immunogenicity would help explain why HRP2-based tests work the way they do.
Finally, the question is, why do people not produce Ab to HRP2? A number of possibilities exist. First, soluble proteins are usually poorly immunogenic in the absence of adjuvants. This explanation seems unlikely since HRP2 is also associated with the erythrocyte membrane. Secondly, endogenous histidine rich glycoprotein (HRG) exists in human plasma that also binds to divalent metal ions. However, Ab to HRG do not bind to HRP2 and mAb to HRP2 do not bind to HRG [
23], making it unlikely humans are immunologically tolerant to HRP2. Third, Das et al. found that peripheral blood mononuclear cells from malaria-exposed adults proliferated and secreted IL-4 when stimulated in vitro with MSP1-42, but not with HRP2 [
11]. They suggested that HRP2 might suppress some immune responses, including T cell proliferation, IFNɤ secretion and CD69 expression [
26]. Fourth, HRP2 may lack an adequate number of T cell epitopes to induce an acquired B cell response in most individuals. Using predictive models, only one predictive Class II T cell epitope was identified, namely, the core sequence LNLNKRLLH located before the repeat sequences that only has high affinity for DRB1*011.01. Within the repeat region, which contains the majority of amino acids, only one T epitope with intermediate-binding was predicted, i.e., amino acids 45 to 113 with HLA-DQA1*05:01/DQB1*03:01 and one weak interaction with DRB3*01:01. A search of the Allele Frequency Net Database showed these alleles are not common in Africa, e.g., DRB1*11.01-DQA1*05:01 -DQB1*03:01 is present in 1.8–2.6% in Ethiopia and HLA-DRB3*01:01 was not reported. Since HRP2 has a unique amino acid composition and structure, it is plausible it may not be highly immunogenic in humans.
Authors’ contributions
DWT designed the study; NB conducted the experiments and assisted with data analysis; VSK performed the bio-informatic analysis; RGFL and IAQ conducted the field studies and provided the plasma samples, parasitological and clinical information. DWT wrote the first draft of the manuscript, which was edited by RGFL and IAQ. All authors read and approved the final manuscript.