Background
Nearly half of the world’s population lives in malaria endemic zones.
Plasmodium falciparum dominates in sub-Saharan Africa and accounts for the very high prevalence rates associated with this continent.
Plasmodium vivax is mostly prevalent in the American continent and Asia [
1]. Prevalence rates can mask the absolute amount of the disease. For instance, India has low endemicity due to the large population size but, as a country, India has the second highest number of malaria deaths in the world [
1].
Within India, individuals are exposed to varying levels of endemicity and many develop partial protection against infection [
2‐
4]. This naturally acquired immunity (NAI) is expected to vary with age, host genetic makeup, parasite species, and level of endemicity. In high transmission settings, individuals almost always have some levels of parasites in their blood. These asymptomatic carriers develop partial protection from severe disease at an early age, but rarely develop sterile immunity. In areas of low transmission, most individuals exhibit moderate to high-grade parasitaemia and develop clinical symptoms. Children are at a greater risk and are susceptible to severe disease until the age of about 2–3 years. Clinical immunity, although far from optimal, gradually develops with age, but lasts only as long as individuals are continuously exposed to and repeatedly infected with malaria parasites [
5].
NAI is mediated by acquired immune mechanisms, primarily directed against blood stages of malaria parasites. Such immunity involves circulating immunoglobulins, antibody-producing plasma cells, and memory B cells [
6‐
9]. Knowledge of dynamics underlying NAI is vital in identifying the most effective infection monitoring strategy for specific epidemiological settings. NAI can also be used to study the impact of control strategies on exposure and transmission. Antibodies exhibit several advantages over entomological and parasitological methods to estimate malaria exposure, prevalence, and transmission. In particular, antibodies provide evidence of exposure history since they persist in the body for some time. In an important technical advance, high-throughput protein arrays have been designed to study immunity against both
P. falciparum and
P. vivax in a cost-effective manner [
10,
11]. A large collaborative effort was promoted on a global scale through the US NIH International Centres of Excellence for Malaria Research (ICEMRs). The efforts were expected to identify biosignatures of immunity, including the most informative antibody responses across a large number of malaria endemic sites. Using this approach, several groups coordinated investigations of antibody reactivity to hundreds of parasite antigens [
10‐
19]. The studies were first performed using samples from 14 regions of the world with different epidemiological settings, transmission intensities and parasite prevalence. The trans-ICEMR survey used a standardized platform to measure antibody responses which was also used by other groups to study exposure and acquired immunity [
15,
17,
18]. Using the standardized genome-scale protein arrays, the list of potential antigens that inform on malaria prevelance in global communities, prior exposure to malaria in individuals, and antibodies that affect infection, disease and transmission through mosquitoes has been greatly increased.
Immunity against malaria in India is less well understood compared to in Africa, SE Asia, or neighbouring islands. The effort above with the first protein arrays included a promising start to genome-scale probing of malaria immunity in India. An early comparative analysis of the seroreactivity profile from three sites, Raurkela, Nadiad, and Chennai in India indicated varied immune responses across the epidemiologically diverse endemic settings [
19]. Raurkela and Nadiad are near rural sites with limited immigration from other parts of India. Chennai, which is in the deep south, has low level
P. vivax and no local
P. falciparum. Yet, importantly, asymptomatic carriers revealed higher reactivity to 19
P. falciparum antigens compared to symptomatic patients and they included PHISTc, MSP11, RH2b, RON2, SERA4, two VARs, RESA, ETRAMP5, MSP2, SEMP1, HSP70-x, GEXP18, GCN5, LSA3, ETRAMP14, MSP4, ring infected erythrocyte surface antigen and a conserved protein.
To assess the generality of the early findings from the studies in India and the world, and to understand and measure the level of exposure in low transmission areas, protein array-based immune surveillance was performed in a low epidemiological setting in Southwest India (Goa). While Goa is considered a non-malaria endemic region of India, it has significant local seasonal transmission of both
P. falciparum and
P. vivax. Most importantly, Goa Medical College and Hospital offers free health care to all and attracts a very wide spectrum of malaria patients. Many participants involved in the present study were immigrants from different parts of India, with different genetic backgrounds, and different histories of prior exposure to malaria [
20]. These complex dynamics are typical of many large cities in central India and offer a chance to capture immune status of a wider population of infections in India. Collections at a hospital setting assured a broader set of disease presentations. In order to differentiate between recent and cumulative exposure, Immunoglobulin G (IgG) Ab responses in young children and adults were studied. Additionally, for the first time, a comparative analysis of IgG Ab levels in non-severe and severe malaria patients was performed to potentially discover new antigens of importance in the context of protection from severe disease.
In summary, the present study reveals, (i) seroreactivity profiles against both P. falciparum and P. vivax antigens of symptomatic malaria patients in Goa and (ii) the relationship of IgG antibody levels of this population in Goa to multiple factors that could contribute to differential immunity. These include their age, disease severity status, and season of sample collection. Such insights can potentially identify markers of exposure and disease severity that would be useful for malaria surveillance. The present results compare favourably with reactivity from non-urban areas of India and to the other ICEMR studies from around the world. This suggests that such antigens may have value as general sero-surveillance tools in very different epidemiological settings.
Discussion
Naturally acquired immunity (NIA) depends on previous infection and builds with every pathogen encounter [
22]. It is generally proportional to the duration and degree of exposure to parasites, wanes rapidly in the absence of an active infection [
23,
24] and is specific to the infecting
Plasmodium species. To an extent, it is also specific to different life stages of the parasite [
22]. To disease control experts, NAI, especially in children can be useful for estimating areas with on-going transmission and varying disease burden [
25]. In principle, it may be used to determine whether populations are at a high risk of severe disease.
India contributes significantly to the overall global malaria burden. It is important for the study of malaria immunity due to (a) the diverse eco-epidemiological profiles across the country, (b) co-existence of multiple
Plasmodium species and vectors, (c) changing climatic patterns that may have an impact on malaria transmission and (d) emergence of anti-malarial drug resistance [
26]. India has an extensive integrated public health surveillance system to identify disease burden, as well as morbidity and mortality in a community [
27]. As a part of an early founding inter-ICEMR collaboration, a study described the serological profiles of 236 malaria positive patients from three different sites in India. Indian samples showed seroreactivity to 265
P. vivax and 373
P. falciparum antigens. A significant difference in the levels of seroreactivity and breadth of antibody response was observed across the three study sites [
19]. Additionally, a linear correlation was observed between the breadth of Ab response and malaria prevalence with respect to
P. falciparum, but not
P. vivax. These findings suggest that indicators of exposure may vary across diverse endemic settings and there may be a need for establishing pan-specific general antigens that may be of use for surveillance campaigns in India [
19]. Of particular interest are urban and semi-urban areas that have local malaria transmission but also large movement of workers from rural and poorer states of India [
20].
The current study focuses on patients from Goa, a small prosperous coastal state in western India where the epidemiology of malaria includes the presence of migrant workers and transient communities. Goa is classified as a non-endemic region with low transmission intensities where parasitaemia are usually low (< 1%). Using the shared ICEMR protein array platform, the present study describes the profiles of NAI in malaria-positive patients residing in Goa. The primary goal was to explore species-specific and pan-species serological correlates of exposure. Further, differential seroreactivity was identified in severe versus non-severe malaria patients to help identify antigens likely to offer protection from severe manifestations. This cross-sectional serological study was performed using a small subset of the total number of malaria-positive individuals referred to the MESA-ICEMR study team [
20]. Of over 1000 confirmed malaria cases, a majority (88.2%) were born outside of Goa (primarily Uttar Pradesh and Bihar) and 51% were construction workers. A larger proportion of the patients were diagnosed with
P. vivax (77%), while 21% had
P. falciparum malaria. Of these 1000 confirmed malaria cases at GMC, 96 falciparum and 100 vivax patient sera were used in the present serology study.
Consistent with previous ICEMR protein array data from around the world,
P. falciparum patients showed a remarkably stronger immune response compared to
P. vivax. The reasons for this could be many, the major one being the significantly different biology of the two parasites [
16]. Secondly, the protein arrays were designed against the genome of a single strain for each species (
P. falciparum 3D7,
P. vivax Sal1), allowing the introduction of several important biases. Genetically,
P. vivax is extremely diverse and the protein array is made based on a South American Strain. This may explain the lower general reactivity seen in Goa. Overall, 248
P. falciparum and 73
P. vivax seroreactive proteins were identified. Interestingly, most of the top seroreactive antigens of
P. falciparum (Table
1A) were previously identified in global studies using
P. falciparum protein array platforms [
11,
15,
28,
29]. Many were also found to be immunogenic in more rural endemic regions of India [
19]. Similarly,
P. vivax antigens have also been seen in previous global microarray studies from other parts of the world. According to a previous study, a few
Plasmodium exported proteins (PVX_083560, and PVX_121930), a hypothetical protein (PVX_118705) and MSP10 could discriminate between naïve and semi-immune individuals [
12]. The immuno-proteome of
P. vivax recently published using a protein array with 1936 genes encoding
P. vivax proteins identified 151 highly seroreactive
P. vivax antigens, a few of which (PVX_115450, PVX_090230, PVX_085025 and PVX_087670) were also recognized in our study [
14]. In conclusion, for global prioritization of antigens, select proteins transcend national and continental variations in malaria host-parasite biology.
Importantly, the malaria patients in Goa showed differential levels of IgG reactivity to different polypeptides from the same protein, in both
P. falciparum and
P. vivax. Individual proteins, or even different peptides in a protein, can elicit varying levels of IgG reactivity. Protein-specific features such as subcellular location, protein abundance, degree of polymorphism and presence of human orthologs may influence the magnitude of antibody responses during natural malaria infections [
17]. Overall, once an antigen receives high priority based on epidemiological investigations, a deeper hunt for the best antigenic peptide from that protein may be fruitful for signal intensity and breadth of responses in target patient populations.
A majority of the 248
P. falciparum and 73
P. vivax seroreactive antigens identified were understandably either proteins normally exported to the erythrocyte cell surface during the parasite life cycle or present on the merozoite surface. Yet, some nuclear, cytoskeletal and cytoplasmic proteins also triggered high IgG binding in the majority of the infected population. A 1990s study from India, provided early evidence for immune reaction against intracellular proteins [
30]. A differential immuno-screening of an erythrocyte-stage cDNA expression library revealed novel protein targets that were exclusively recognized by immune sera and not acute patient sera. Surprisingly, one was a conserved ribosomal protein P0 and another was an endonuclease, alongside several conserved hypothetical proteins. Antibodies against four of these proteins also inhibited
P. falciparum growth in culture and correlated with protection in mice [
30]. Most important invasion proteins, and also those strongly implicated in correlates of protection, are highly polymorphic. This may reduce seroreactivity and introduces a bias in protein screens. The latter may favour highly conserved genes which are often intracellular but not necessarily relevant for protection.
The basis of the major cross-reactivity observed in the study could be the large number of orthologous
P. falciparum and
P. vivax proteins printed on the chip, but it is not that simple. Of 123
P. falciparum antigens exclusively seroreactive in the
P. falciparum study group, 93 were found to share orthologs in
P. vivax (Additional file
5: Table S4). Reciprocally, there were also non-orthologous antigens which did not show any species specificity. These observations highlight the need to better understand antigenic cross-reactivity. It should be noted that differentiation between
P. vivax and
P. falciparum co-infections was based on two independent microscopy determinations. It is possible that cross-reactivity due to sub-microscopic infection could affect antibody levels in areas such as Goa, where both
P. falciparum and
P. vivax co-exist. Finally, residual seroreactivity from past exposures to either
P. falciparum or
P. vivax could also be responsible for unaccounted cross-reactivity.
In areas of high transmission, parasitaemia and risk of morbidity and mortality have an inverse relationship with age: Disease versus parasitaemia can be distinct, and dependant on age. Children often exhibit anti-disease immunity, which confers protection from the risk and extent of morbidity associated with a given parasite density [
22]. In contrast, adults exhibit anti-parasite immunity which protects against high parasite density and the risk of severe disease. Compared to the above patterns in high-transmission areas, immunological patterns can be different in areas with low transmission. In our study, a few antigens showed remarkably higher signal intensities in older adults. Higher reactivity in adults can arise from legacy antibody profile derived from decades of previous intense exposure, which is durable or can arise with age. To further complicate matters, our study is dominated by a migrant population, so this finding could be the product of both historic exposure from a different epidemiological setting and recent exposure boosting select parts of the NAI. It is noteworthy that the common responses transcend epidemiological settings and are age dependent, and not the result of several years of heavy local exposure in Goa [
22].
To explore the possibility of identifying markers of disease protective immunity, participants were segregated as in-patients and out-patients. The overall antibody response was higher in in-patients compared to out-patients, and this was presumably a response to the most recent infection (Fig.
5a). In-patients were further classified as severe and non-severe based on the signs/symptoms of severity. Importantly, a few
P. falciparum antigens were recognized in all sera from non-severe hospitalized patients, and were largely non-reactive in most of the severe hospitalized sub-group. These antigens could serve as correlates of protection against severe manifestations or may even confer protection against disease. Although further assessment is required to fully understand immune responses that can protect against disease, this study provides the first evidence of differential seroreactivity to
P. falciparum antigens in severe and non-severe patient cohorts using protein microarrays. This trend was not observed in case of
P. vivax malaria, possibly due to the overall low antibody response to
P. vivax. Deeper longitudinal cohort studies will be necessary to expand the body of evidence for such potential protective immunity.
With the success of malaria control programmes, transmission is likely to decline in the next few years. Low transmission areas require increasingly sensitive tools to check for malaria exposure. Antigens which are seroreactive across varying epidemiological and transmission settings, and exhibit species-specificity may represent the best markers of exposure. Protein arrays provide a suitable platform for the identification of a manageable subset of antigens eliciting the strongest response. Antibodies against these antigens may be measured using basic immunoassays, such as enzyme-linked immunosorbent assays (ELISA). The biggest benefit is the ease with which antibodies can be stored using dried blood spots, making sample collection and storage relatively simple [
31].
To build on present advances, future follow-up studies may improve in four areas: first, more effort is required to enroll children and adolescent patients. Second, longitudinal serology studies will track exposure to malaria, protective immunity, and waning of such immunity after an infection is resolved. Third, more detailed information will be sought on the origin of migrant workers and prior history of infection in their place of origin. Finally, the array platform itself will evolve to test sera against some shorter exon fragments on the array, some full-length proteins, all accepted blood-stage antigens, and antigens from non-erythrocytic stages of malaria parasites to assess their contribution to local transmission.
Authors’ contributions
AV, PR, SS, and SP conceived and designed the experiments. AV performed the experiments and analysed data, AJ did initial training and sample analysis, HD and AJ normalized data. EG directed patient sample collections. LP and JM managed patient sample inventories, patient data, and patient sample movements. Co-author PF provided the protein arrays and reviewed drafts of the manuscript. AV and PR wrote the manuscript. All authors read and approved the final manuscript.