Review Article

Robert Koch redux: malaria immunology in Papua New Guinea

D. I. STANISIC,1 I. MUELLER,2 I. BETUELA,2 P. SIBA2 & L. SCHOFIELD1

1The Walter and Eliza Hall Institute of Medical Research, Parkville, Vic, Australia, 2Papua New Guinea Institute of Medical Research,Goroka, PNG

SUMMARY

Over a century ago, the malaria expedition of the brilliantmicrobiologist Robert Koch to the Dutch East Indies(Indonesia) and German New Guinea (now Papua NewGuinea, or PNG), resulted in profound observations thatare still central to our current understanding of the epide-miology and acquisition of immunity to the malaria para-site Plasmodium. The tradition of malaria research inPNG pioneered by Koch continues to this day, with a num-ber of recent studies still continuing to elucidate his origi-nal concepts and hypotheses. These include age andexposure-related acquisition of immunity, species-specificand cross-species immunity, correlates of protective immu-nity and determining the prospects for anti-malaria vac-cines.

Keywords immunity, Plasmodium, Papua New Guinea

INTRODUCTION

Of the five documented species of Plasmodium that infecthumans, four are found in Papua New Guinea (PNG)(Plasmodium falciparum, Plasmodium vivax, Plasmodiumovale and Plasmodium malariae), with P. falciparum andP. vivax the dominant species. The high prevalence ofP. vivax is a unique feature of the malaria species com-position compared with Africa. The geography of PNGis extraordinarily diverse, varying from coral atolls, low-land coastal areas and coastal volcanoes through to high-land veldts, montane habitats and snow-capped mountain

peaks. Concurrent complexity in rainfall and temperatureresults in extensive variation in the ecology of both mos-quito and human hosts. Cultural and linguistic diversityin human populations is mirrored in a comparativelyhigh degree of genetic variability over small distances.PNG therefore offers in microcosm the broadest rangesof malaria epidemiology. Changing temperature withincreasing altitude is the main climatic determinant ofmalaria endemicity (1). In lowland regions, there isperennial transmission of malaria with limited seasonali-ty, and malaria can be the main cause of morbidity inhealth facilities (1). As in other areas of high malariaendemicity, the burden of infection and disease falls onyoung children (2) and pregnant women. In the high-lands, malaria is absent, but can occur in local epidem-ics.

Malaria has been the focus of much research in PNG,with the first systematic studies of malaria epidemiologycarried out in the late 1800s by the microbiologist RobertKoch. Reports were sent back to the Colonial Departmentof the German Foreign Office and were published in theGerman Medical Weekly (Deutsche Medizinische Woc-henschrift) with English translations published in the Brit-ish Medical Journal (3–5). These reports containedobservations he made during the expedition, some ofwhich were to form the basis for understanding the epide-miology of malaria and also gave critical insights into theacquisition of immunity. The key observations have beenoutlined in a review of Koch’s contribution to malariaresearch (6) and include:1. In endemic areas, malaria is most apparent in young

children and the parasites may be completely absentfrom the blood of adults.

2. Individuals who are constantly exposed to infectiondevelop immunity to malaria.

3. Immunity is species specific.4. It may be possible (when enough is known about the

mechanisms involved) to immunize against malaria tox-ins.

Correspondence: Dr Louis Schofield, Infection and ImmunityDivision, The Walter and Eliza Hall Institute of MedicalResearch, 1G Royal Parade, Parkville, Vic, 3052, Australia(e-mail: schofield@wehi.edu.au).Disclosures: L.S. declares a financial interest in AncoraPharmaceuticals Inc. that might benefit from this publication.Received: 09 April 2010Accepted for publication: 08 May 2010

Parasite Immunology, 2010, 32, 623–632 DOI: 10.1111/j.1365-3024.2010.01238.x

� 2010 Blackwell Publishing Ltd 623

5. Malaria may break out anew in an area each time alarge number of nonimmune people are introduced tothe area.

ACQUISITION OF IMMUNITY

While in PNG, Koch observed that among Chinese andMalay immigrants, malaria parasite rates were muchhigher in recent arrivals than in those who had been inPNG for many years (4). Additionally, he observed thatchildren from the indigenous Melanesian populations werecommonly infected, whereas parasites were not as fre-quently found in adults (4). From this he deduced, andnumerous subsequent studies have confirmed, that acquisi-tion of immunity to malaria requires constant exposureover a number of years and never results in sterile protec-tion. The immunity that does obtain manifests as a reduc-tion in parasite densities and reduced risk of disease.Clinical immunity can be defined as the absence of diseasein the presence of infection (7,8).

While the slow acquisition of anti-parasite immunity isoften thought to reflect the need for long-term cumulativeexposure to the parasite, it remains possible that age is animportant independent variable in this process. In malariaendemic areas, increasing age and cumulative exposure areinextricably linked, making it difficult to distinguish therelative contribution of each to the successful acquisitionof immunity. To address this important issue, Baird andcolleagues conducted several studies in Irian Jaya (WestPapua), examining infection and morbidity rates invillages containing both life-long Melanesian residentsnative to the region and recent malaria-na�ve Javanesetransmigrants. Importantly, both populations were co-resi-dents in the same villages and encompassed a wide agerange (neonates–adults). Transmigrant adults acquiredclinical immunity to malaria after relatively few infections,in contrast to children, who remained susceptible aftercomparable exposure (9). These findings support otherepidemiological studies observing age-dependent immunityin individuals with only brief exposure to malaria, andsuggested an alternative model for acquisition of immu-nity, where protection is linked to an intrinsic age-depen-dent variable that requires much less exposure in adultsthan children when both are exposed to hyperendemictransmission (10,11).

SPECIES-SPECIFIC IMMUNITY

Another important concept developed by Koch followedfrom observing the movement of adults from a regionwhere only P. malariae was endemic to a region whereother species of malaria were present. They became sick

soon after arrival with non-P. malariae malaria, leadingKoch to hypothesize that immunity against one Plasmo-dium species offers little protection against the other spe-cies. This in turn leads to several important relatedquestions, namely: do the mechanisms of naturallyacquired immunity to various malaria species differ, and ifso, what are the distinct mechanisms, and what influencedoes one species exert on acquisition of immunity to theothers? Cross-species regulation of malaria parasites hasimplications for the implementation of single-species vac-cines. Reduction in the prevalence of one species (througha species-specific intervention such as a vaccine) couldresult in an increase in infection and disease caused byother species that may have been previously suppressed bynonspecific and ⁄ or specific mechanisms. This is particu-larly relevant for PNG, where the four main species infect-ing humans co-exist and the knowledge of anti-malarialresistance in the nondominant species is not as wellknown. For these reasons, the issue of species-specificimmunity and mixed species interactions in relation toinfection and morbidity within the human host has beenstudied more extensively in PNG than elsewhere.

Mixed species interactions

Using PCR and light microscopy (LM) in cross-sectionalsurveys in three different populations in PNG, Melhotraet al. (12,13) showed that while use of PCR greatlyincreased the frequencies of mixed species infection com-pared to LM, both methods were consistent in observinga random distribution of different types of infection. In amore recent larger study, mixed species infection wasobserved less commonly than expected by chance whendiagnosed by LM but significant excess of mixed infec-tion was observed by PCR (14). Analysing blood smearsfrom seven cross-sectional surveys carried out in tenWosera villages at approximately four monthly intervals,Smith et al. (15) found infections with P. falciparum to beindependent of P. vivax or P. malariae, but a strong nega-tive correlation was observed between infections withP. vivax and P. malariae. There were significant differ-ences in age-specific prevalences among the different spe-cies with P. vivax most frequently detected in youngchildren and P. malariae in older children and adults(2,16) After adjusting for age and sex, a significant nega-tive association between P. falciparum and P. vivax wasobserved [Odds ratio 0Æ80, CI (0Æ72, 0Æ89)]. However,when the data were analysed longitudinally, positive cor-relations were found for all pairs, except forP. vivax ⁄ P. malariae, presumably reflecting correlations inexposure rather than positive interactions between species.In a third study, Bruce et al. (17) looked at changes in

D. I. Stanisic et al. Parasite Immunology

624 � 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632

species composition and densities in a group of asymp-tomatic, semi-immune children in Madang. Similar to thestudy by Smith et al., a negative association betweenP. falciparum and P. vivax was found for cross-sectionalanalyses, but not for the full data set of 21 bleeds (17).In addition, Bruce et al. (18) found that total Plasmodiumdensities were limited and concluded that the dynamics ofparasite densities were consistent with density-dependentregulation, indicating a negative interaction between thespecies that prevented simultaneous high levels of infec-tions with more than one species.

Much less data are available on possible cross-specieseffects on malarial morbidity. In their study, Smith et al.(15) calculated the prospective risk of morbidity followingearlier malaria infections and found that the risk of healthcentre attendance with a P. falciparum-associated feverdecreased following an (asymptomatic) P. vivax infection[OR: 0Æ52, CI95(0Æ26, 1Æ02), P = 0Æ04]. However, onceadjusted for covariates, significance was lost. In a recentanalysis of blood smear readings from 24 670 cases withpresumptive malaria investigated during routine morbiditysurveillance 1990–2002 at the two health centers in PNG,we found strong negative associations between malarialspecies in symptomatic cases and a highly significantshortage of mixed infections compared to asymptomaticcontrols (Table 1).

In summary, there is little firm evidence that anyPlasmodium species is consistently associated with protec-tion against heterologous infections. However, there maywell be interactions among individual infections withinthe host that result in density-dependent regulation ofparasitaemia in asymptomatic individuals and protectionagainst clinical disease, in particular disease associatedwith P. falciparum. Given that PNG is the designated sitefor future trials of vaccines directed against P. vivax (19),

it is critical to understand whether P. vivax decreases thefrequency of clinical P. falciparum episodes. Furthervalidation in well designed, longitudinal studies usingPCR-based methods for species identification is thereforeurgently required.

Species-specific immunity

A prospective treatment–reinfection study carried out inthe Mugil area on the North coast of Madang in 2004(referred to hereafter as the Mugil cohort) involved a 6-month follow-up of a cohort of 206 5- to 12-yr-olds (20).The high rates of re-infection with both P. falciparum andP. vivax together with the sensitive molecular methodsused to detect re-parasitization made this an excellentstudy for examining species-specific immunity. Despitecomparable levels of exposure to both parasites in thisstudy, different patterns of immune acquisition wereapparent with a more complete clinical immunity observedfor P. vivax than for P. falciparum. Unlike P. falciparum,there were very few clinical episodes attributable toP. vivax during follow-up, indicating that by this age, ahigh degree of clinical immunity had been attained toP. vivax.

Another longitudinal cohort study carried out in theEast Sepik Province in 2006 followed 264 children aged1–3 for 16 months (referred to hereafter as the Ilaitacohort) (21). In this age group, despite similar exposure toboth parasites, a difference in the acquisition of immunityto P. vivax and P. falciparum was also observed, specifi-cally the ability to control symptomatic illness in P. vivaxbut not P. falciparum infection and a decrease in the inci-dence of P. vivax infection with age. This study confirmsthe findings in the Mugil study and others (16,20) withrespect to faster acquisition of immunity to P. vivax. Thisis an intriguing observation and naturally leads to specula-tion as to the reasons for this difference. The answer mustlie within distinctive biological features of P. vivax.Although there are many such unique features, a simpleexplanation may lie in the expression by P. falciparum andnot P. vivax of dominant variant surface antigens (VSAs)known as the P. falciparum erythrocyte membrane protein-1 (PfEMP-1) family encoded by 50–60 different var familygenes. PfEMP-1 mediates sequestration of infected ery-throcytes, and antigenic variation (and thus immune eva-sion) is mediated by switching between different PfEMP-1types. Certain PfEMP-1 types are associated with placen-tal malaria (22), while others may mediate severe diseasein children (23). Recently, PfEMP-1 was shown to sup-press cellular immune responses (24). Lack of a PfEMP-1orthologue may explain in part the very differenthost ⁄ parasite relationship in P. vivax.

Odds-ratio

Odds-ratio

Volume 32, Number 8, August 2010 Malaria Immunology Research in PNG

� 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632 625

IMMUNITY TO MALARIA TOXINS

Following his studies on acquired immunity in Madang in1899–1900, Koch wrote ‘One may assume an anti-malarialimmunity, which under ordinary conditions is acquiredonly in 4–6 years and after repeated attacks, to be con-ferred artificially and in a short time. Seeing, however, thatas yet we have no idea how to obtain the toxins necessaryfor the immunising process, chances of progress in thisdirection are small’. (5). The proposal by Koch that itmay be possible to immunize against malaria followed thework of Golgi, who in 1886 observed that the acutemalaria paroxysm followed closely the 48 or 72 h blood-stage developmental cycle of various malaria parasites andsuggested that one or more toxins were released uponschizont rupture. Since then, parasite glycosylphosphatidylinositol (GPI) was shown to have the properties predictedof a toxin (25,26). Immunizing against GPI prevents sev-eral severe disease syndromes in pre-clinical models (27),indicating that more than a century after Koch it hasfinally been possible to fulfil his postulate on this particu-lar matter.

Several studies have addressed the question of whethernaturally acquired immune responses to GPI could medi-ate anti-toxic and clinical immunity to malaria. Address-ing this issue properly has been confounded by thedifficulty of purifying GPI to compositional homogeneityin sufficient yield. Purifying GPI requires multiple sequen-tial and independent fractionation procedures, especiallyin Plasmodium, where GPI copy number is not high andlarge quantities of phospholipids are expressed. Despitethis, several studies have undertaken GPI serology onmaterial prepared by a one-step HPLC fractionation ofthe total pool of organic solvent and water extractableparasite material, where GPI material elutes in a pool rep-resenting over 20% of the total fractionation gradient (28).Definitive compositional purity analyses were not under-taken. Studies using this material in ELISA endpointshave provided contradictory findings. A well-designedstudy in Javanese transmigrants to West Papua found anti-GPI antibodies were associated with reduced risk of symp-tomatic malaria in children (29) consistent with studies inAfrica reporting negative association with disease states(28). However, cross-sectional surveys in PNG reported noassociation of antibodies measured in this way with toler-ance (30). More recently, GPI antibodies in the Mugilcohort were measured by ELISA using compositionallypure synthetic GPI glycan and were associated withreduced risk of morbidity (L. Robinson and L. Schofield,unpublished observations). Such approaches may thereforeyield a more accurate correlate of anti-toxic or clinicalimmunity.

IMMUNE RESPONSES TO MALARIA BLOOD-STAGES AND THEIR ANTIGENS

Despite Koch’s elucidation of acquired immunity tomalaria over 110 years ago, the mechanisms and targets ofanti-parasite immunity remain unclear. Antibody-mediated,innate and acquired cellular immune responses appear tobe involved. Indeed, it is likely that all three compartmentscollaborate in mediating the immune state. The main tar-gets of antibody are thought to be merozoite surface anti-gens and VSAs expressed on infected erythrocytes. In bothcases, IgG may act via Fcc Receptor (FccR) interactions inconcert with the innate system such as phagocytic cells.This system would in turn be highly regulated by Th1 cyto-kines from the innate and acquired T-cell arms. WhileFccR-mediated interactions seem intuitively likely to con-tribute to immunity, this important area remains veryneglected in application to malaria immunology.

Cell-mediated immunity to blood-stage Plasmodium fal-ciparum

The precise nature and role of cellular immune responsestargeting the malaria parasite are not clearly understood,although some studies have implicated adaptive effectorcells as playing key roles through the release of mediatorssuch as interferon-c (31). Protective immunity may in factreflect a network of highly regulated interactions betweeninnate and adaptive cellular pathways, with immune regu-lation critical because of the ability of cytokines to pro-mote both immunopathology and protection.

The first studies of cellular immunity to malaria inPNG were undertaken to provide baseline data for subse-quent vaccine trials. Al-Yaman and colleagues obtainedvenous samples from children 0Æ5–15 years, with morbiditysurveillance carried out for 1 year through community-based and self-reported case detection (32). Immunologicalmeasures included T-cell proliferation, IL-4 and IFN-cproduction specific for the RESA antigen and schizontextract. RESA-specific IL-4 elicitation and IFN-cresponses to schizont extract were significantly higheramongst those who experienced no episodes of malariacompared with those who had two or more episodes ofany ⁄ high parasitaemia over the subsequent year (32). In abroadly similar study design (33), there was no significantassociation between T-cell proliferation, IL-4 and IFN-cproduction specific for MSP2 FC27, 3D7, d3D7 (withoutthe central repeat region) and schizont extract, and subse-quent parasitological and clinical indicators, during the1-year follow-up.

The high rates of re-infection with both P. falciparumand P. vivax together with the sensitive molecular methods

D. I. Stanisic et al. Parasite Immunology

626 � 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632

used to detect reinfection make the Mugil and Ilaita longi-tudinal cohorts outlined above excellent studies for exam-ining acquisition of immunity and identifying correlates ofprotection and risk (34). Thus, early cellular responseswere examined as correlates of immunity ⁄ risk in bothcohorts. In the Mugil cohort (5- to 12-yr-olds), peripheralblood mononuclear cells isolated at baseline were stimu-lated with live, intact P. falciparum parasites and theinduced cytokine responses (IL-2, IL-4, Il-6, IL-10, TNFand IFN-c) were assessed in relation to the risk of clinicalmalaria episodes during the 6-month follow-up period.Interestingly, there was an age-dependent increase in earlycytokine responses that paralleled an age-dependentdecrease in the incidence of clinical P. falciparum episodes.When considering individual cytokines, only early produc-tion of IFN-c by cd T cells and ab T cells was associatedwith protection from high density P. falciparum clinicalepisodes of malaria (35). In a multivariate statisticalmodel, high levels of IL-6 relative to TNF and IFN-c wereassociated with an increased incidence of clinical P. falci-parum episodes, while high levels of TNF and IFN-c rela-tive to IL-6 were associated with a reduced incidence (36).Monocytes ⁄ macrophages and cd T cells were shown to beimportant cellular sources of IL-6 and TNF. None of thecytokine responses were associated with risk of re-infec-tion. Thus, cytokine ratios predict risk of higher densityparasitaemias but not incidence of infection, consistentwith their being elicited in response to infection.

Similar trends were seen with TNF and IFN-c in rela-tion to protection from high-density clinical episodes inthe younger Ilaita cohort of 1- to 3-yr-olds (L. Robinson,I. Mueller and L. Schofield, unpublished observations).Additionally, IL-10 and not IL-6 was associated with anincreased incidence of clinical episodes. Interestingly, inthis age group, IFN-c was predominantly derived from cdT cells with only a minor contribution by NK and ab Tcells. The greater contribution of the more innate-typeimmune cells in this younger age group reflects the lack ofacquired immunity in this cohort compared with the olderchildren in the Mugil cohort. Taken together, the resultssuggest key roles for TNF and IFN-c in protection fromclinical P. falciparum malaria episodes. These cytokinesmay act by limiting parasite replication and thus develop-ment of symptomatic malaria. Importantly, the relativecontribution of the different effector cell types to theseprotective immune responses appears to change with ageand this may reflect transition from a nonimmune statusto active acquisition of immunity. In parallel with the epi-demiological findings from this study, the data suggestthat immunity against P. falciparum is still being acquiredat this age and that the acquired cellular immune com-partment may be insufficiently primed.

Cellular responses to Plasmodium vivax

The hyperendemic transmission of P. vivax in PNG makesit an ideal setting in which to examine P. vivax biologyand immunology. This parasite has been the subject of anumber of studies over the last few years, but the recentrecognition of the importance of this parasite in relationto burden of disease and the need for a species-specificvaccine has ensured that P. vivax will be high on theresearch agenda in PNG in the coming years.

As previously noted, the Mugil and Ilaita cohortsrevealed a much more rapid and complete acquisition ofimmunity for P. vivax when compared to P. falciparum.Therefore, cellular responses to P. vivax were investigated,specifically to Region II of the Duffy Binding Protein.Three dominant T-cell epitopes have been identified inPvDBPII (37), with two of these epitopes being polymor-phic with distinct variant-specific T-cell responses, whilethe third is conserved. Recent studies have examined cyto-kine responses to these epitopes in relation to protectionfrom re-infection in the Mugil cohort. Children thatacquired a predominantly Th2-type cytokine response (IL-5 and ⁄ or IL-13) to one or more PvDBPII peptides showeda 35% increased susceptibility to P. vivax infection, while atrend towards protection was seen when the cytokineresponses were predominantly Th1-type (IFN-c andTNF). Those children with a high ratio of Th1- to Th2-type cytokine responses (IFN-c-TNF ⁄ IL-5-IL-13) showeda 50% protection against re-infection with P. vivax(C. King, unpublished observations). Importantly, in chil-dren who acquired high levels of binding inhibitory anti-bodies directed against PvDBPII, which alone correlatedwith 50% protection against re-infection (38), along with ahigh ratio of Th1- to Th2-type cytokine responses, protec-tion increased to 70% suggesting a cooperative effectbetween humoral and cellular immunity for protectionagainst re-infection with P. vivax. Exactly how the cellularimmune responses augment humoral immunity toPvDBPII is unclear and warrants further investigation.

Antibodies to Plasmodium falciparum antigens

Passive transfer of c-globulin from immune adults success-fully treated children suffering from severe malaria, dem-onstrating the importance of IgG against blood stageparasites in protection against disease (39). There aremany targets of the humoral immune response, and indepth discussion of these is beyond the scope of thisreview. However, the role of antibodies specific for mero-zoite antigens has been the subject of a number of studiesin PNG. Initially, such studies were aimed at providingbaseline data for antigen selection in relation to vaccine

Volume 32, Number 8, August 2010 Malaria Immunology Research in PNG

� 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632 627

trials [reviewed in Ref. (40)] Cross-sectional surveys wereused to assess naturally acquired immune responses to themalaria antigens MSP1, MSP2, RESA, LSA1, LSA3 andSPf66 under consideration as vaccine candidates(32,33,41–48). Further studies with a longitudinal designwere then undertaken to assess the role of these specificantibodies in protection. Baseline bleeds were undertakenin children 0Æ5–15 years with morbidity surveillance car-ried out for 1 year through community-based and self-reported case detection (48). The immunological measureswere IgG to MSP2, RESA, SPf66 and crude schizontextract. After controlling for age and previous exposure,antibody levels to MSP2, 3D7 and d3D7 (without the cen-tral block of repeats), and RESA were predictive of areduction in incidence rate of episodes of clinical malaria.For MSP2 3D7 and RESA, antibody levels were signifi-cantly higher in the groups who did not experience anymalaria episodes.

A second study with a similar design was undertaken(47) that assessed IgG responses to parasite-derivedMSP-1, N-terminal yeast-expressed (195A) and baculovi-rus-expressed C-terminal region (BVp42) of MSP1 andschizont extract. Without adjustment, IgG to schizontextract and all MSP1 proteins was negatively correlatedwith the number of subsequent malaria episodes. Whenadjustments were made for age and previous exposure,IgG levels to BVp42 and to a lesser extent MSP-1 werepredictive of a decrease in incidence rate of clinicalmalaria. Reduction in risk was most marked for BVp42and was significant even at high parasite densities.

More recently, subclass responses (specifically, IgG1 forMSP1-19 and IgG3 for AMA-1) were predictive of areduced risk of symptomatic malaria and high densityP. falciparum infections in young children in the Mugilcohort (49). With the current AMA-1 vaccine inducing apredominantly IgG1 response (50), this study highlightsthe importance of measuring subclass-specific antibodyresponses to merozoite antigens. Antibodies specific forP. falciparum merozoite surface antigens were not protec-tive in the younger Itaita cohort (D. Stanisic, I. Muellerand J. Beeson, unpublished observations), suggesting thatimmunity against P. falciparum is still being acquired atthis age. Antibodies may be of insufficiently broad reper-toire or concentration to afford protection. Alternatively,the cellular immune compartment may be insufficientlyprimed or regulated.

Antibodies specific for VSA on the surface of infected redblood cells are also thought to play a role in protection.Studies in PNG in the 1980s examined the recognition ofthe unknown antigens on the surface of infected erythro-cytes using clinical isolates that had been recently cultureadapted (51). Extensive serological diversity was observed

as well as an age-related increase in serological recognitionindicating that these responses may be associated with pro-tective immunity. It is now known that the major VSA isPfEMP-1 that mediates the binding of parasitized erythro-cytes to various receptors on the endothelium, such as inter-cellular adhesion molecule-1 (ICAM-1). Antibodies to ageographically broad range of ICAM-1 binding parasiteswere examined in the Mugil cohort. All ICAM-1-bindingisolates from different geographic origins were recognizedby PNG children, and antibodies were associated with pro-tection from malaria (J. Chesson, J. Richards, I. Muellerand J. Beeson, unpublished observations). However, anti-body responses to the same targets were not associatedwith protection in the younger Ilaita cohort (D. Stanisic,I. Mueller and J. Beeson, unpublished observations).

Antibody responses to Plasmodium vivax

The interaction between PvDBP and the Duffy AntigenReceptor on erythrocytes is critical for successful invasionby P. vivax. As the receptor binding domain of DBP(PvDBPII) is highly polymorphic, the possibility that thismolecule contributes to strain-specific immunity was inves-tigated in the Mugil cohort. Antibody responses to fivedifferent variants were examined by ELISA and it wasfound that increasing age was associated with recognitionof increased number of variants (52). For the three mostprevalent PvDBPII variants, there was a delay in time tore-infection with parasites expressing the same allele by16–25% (when compared with parasites expressing alter-nate alleles) and reduced incidence-density parasitaemia.This suggests that antibodies to PvDBPII are in partresponsible for strain-specific immunity, so it would beimportant to take into account the relative prevalence ofthe different variants at a given site when considering thisantigen in relation to vaccine development.

PvDBPII has been shown to be the target of specificantibodies that inhibit P. vivax invasion of erythrocytes invitro (53). The role of these inhibitory antibodies wasexamined in the Mugil cohort using a recently developedfunctional assay. The presence of high-level inhibitoryantibodies at baseline was associated with both delayedtime to and reduced risk of P. vivax re-infection as well asa reduction in parasite densities when compared with low-level inhibitory antibodies (38). The high level inhibitoryantibodies tended to be variant transcending, reducingre-infection with different PvDBPII variants with similarefficiencies. Interestingly, the total antibody reactivityagainst PvDBPII as measured by ELISA demonstrated aweaker association with protection against P. vivax infec-tion. This illustrates the importance of assays that have amore functional-type readout. Similar investigations are

D. I. Stanisic et al. Parasite Immunology

628 � 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632

being undertaken in the Ilaita cohort to establish whethersimilar measures of immunity are also associated with pro-tection in younger age groups. In this younger cohort, thegreater number of clinical episodes attributable to P. vivaxmeans that it will be possible to look for associations withthese antibody responses and protection from clinical dis-ease.

THE COMBINATION B VACCINE TRIAL

The first field trial of an asexual blood-stage vaccine wasundertaken in PNG in 1998 (54). The Combination B vac-cine consisting of three asexual blood-stage proteins(MSP1, MSP2 and RESA) formulated in the adjuvantMontanide ISA 720 was tested in a Phase I trial in PNGadults. A Phase I ⁄ IIb double-blind randomized placebo-controlled trial was then conducted in a cohort of 120children 5–9 years of age in the East Sepik region ofPNG. Children received either the vaccine or a placebo(adjuvant alone) with half of each group receiving Fansi-dar at the commencement of the study. Two immuniza-tions were administered 1 month apart and follow-uplasted for 18 weeks. The outcome was a 62% reduction inparasite density in the vaccinated group that was notpre-treated with sulfadoxine-pyrimethamine (54). As thevaccine only contained the 3D7 allelic form of MSP2, allfollow-up bleeds were genotyped for MSP2. Vaccinees hada lower prevalence of parasites carrying the MSP2-3D7allelic form and had a higher incidence of morbid episodesassociated with FC27-type parasites.

In this study, in terms of cellular responses only the IFN-c response against the MSP1 component was significant(55). Antibody responses, however, were induced against allcomponents of the vaccine with the biggest increase seen inthe MSP2-3D7-specific response. Importantly, while theN- and C-terminal conserved regions of the MSP2 compo-nent were included in the vaccine with the prospect ofinducing a strain-transcending response, they were notimmunogenic in this population (56). Instead, the antibodyresponse was directed primarily against the repetitive andsemi-conserved family region of the protein, which mayexplain the observed allele-specific protective effect. Becauseof the small group sizes, the study was not powered toidentify mechanisms of protection. The allele-specific vac-cine-induced selection pressure exerted on the parasitesnonetheless highlights the issues surrounding vaccines, con-served antigens, allelic diversity and selection.

MALARIA IN PREGNANCY

A number of longitudinal studies in the Madang area haveexamined antibody responses to PfEMP-1 and merozoite

antigens in pregnant women to evaluate immune measuresas prospective indicators of susceptibility to adverse out-comes. Placental-binding isolates originating from all partsof the world were commonly recognized by antibodiesamong pregnant women in PNG, reflective of limited glo-bal diversity of key surface antigens (57). However, therewas also evidence of regional diversification of var2csa,the major antigen expressed by placental-binding isolates,with significant differences seen in sequences from PNGcompared to Africa and elsewhere (57). This has implica-tions for the design of vaccines specific for pregnancy-associated malaria.

THE PNG NATIONAL MALARIA CONTROLPROGRAM

In late 2009 ⁄ early 2010, PNG received significant fundingfrom the Global Funds for AIDS, TB and Malaria(GFATM Round 8) in support of the national malariacontrol program. As in many other countries, the strategyis based on achieving universal bednet ownership throughthe free distribution of long-lasting insecticidal nets(LLIN), the introduction of Coartem (artemisinin–lume-fantrine) as national 1st line treatment and shift to parasi-tologically confirmed treatment through the roll out ofrapid diagnostic tests in rural areas and strengthening ofmicroscopy services in health centres and hospitals. Theseinterventions are expected to result in a significant andsustained reduction in transmission and consequently inthe burden of malarial infections and disease in PNG. Asthe national biomedical research institute, the PNGIMR isplaying a leading role in monitoring the implementationand impact of malaria control in PNG. After executing anindependent evaluation of the PNG GFATM Round 3LLIN distribution program in 2008 ⁄ 9, PNGIMR has beencontracted by the PNG Department of Health to conductthe monitoring and evaluation of the GFATM Round pro-gram.

Although mosquito bionomics, encapsulated in the con-cept of vectorial capacity, determines the broad level ofmalaria transmission, naturally acquired immunity inhumans is the controlling force that determines the relativeincidence and prevalence of malaria infection and diseasein areas of high endemicity. Acquired immunity deter-mines not only the age-specific incidence and prevalenceof disease, but also the expression of pathological pro-cesses and hence the clinical manifestations of malaria.Moreover, acquired immunity is a state of ‘premunition’that requires ongoing exposure to the pathogen to bemaintained. This premunitive basis of naturally acquiredimmunity has significant public health consequences. Inhighly endemic areas, interventions that reduce parasite

Volume 32, Number 8, August 2010 Malaria Immunology Research in PNG

� 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632 629

transmission may prevent the development of immunityand allow the loss of immunity because of insufficientongoing exposure, resulting eventually in an increasedoverall ‘rebound’ burden of severe morbidity. In theabsence of effective vaccines to replicate the premunitionstate, intensified control measures may yet result in chang-ing patterns of morbidity and mortality. Inconsistentapplication of control measures over the longer term or incertain settings may risk aggravated rebound morbidityand mortality. For these reasons, it will be important toevaluate how the predicted reduction in malaria transmis-sion affects the acquisition of immunity and establishedpremunition. In a series of planned studies in PNG, theassociation of altered immune parameters with changinginfection and morbidity rates will be used to define thresh-old markers of functional parasitological and clinicalimmunity in different age groups. Such markers may proveinformative in defining populations or groups at risk ofrebound morbidity.

Koch lacked a framework to observe loss of immunityupon cessation of transmission. Nonetheless, his observa-tion that malaria rates in the general population increaseupon influx of nonimmunes are likely to remain relevantto GFATM and Roll Back Malaria, in relation to thebroader issues of parasite population structure, loss ofacquired immunity, morbidity and transmission.

SUMMARY

Research into malaria immunology in PNG has had anillustrious history, starting with the seminal contributionsof Koch. With the advent of the molecular era, his funda-mental observations have been confirmed and extended bymultiple investigators and studies. A central focus for thepast 30 years has been studying immunology with a viewto the rational design of potential malaria vaccines. In2010, we stand on the brink of a new era in malaria con-trol, with high-profile calls for the global eradication ofboth P. falciparum and P. vivax. It remains unclear as towhether vaccines can actually contribute to the malariaeradication agenda. It remains unclear furthermore howchanges in malaria transmission will impact the acquisi-tion and maintenance of acquired immunity to the disease.Malaria immunology research in Papua New Guinea willcontinue to provide valuable insights into these and otherimportant issues.

ACKNOWLEDGEMENTS

This work was supported by the National Health and Medi-cal Research Council. L.S. is an International ResearchScholar of the Howard Hughes Medical Institute.

REFERENCES

1 Mueller I, Bockarie M, Alpers M & Smith T. The epidemiol-ogy of malaria in Papua New Guinea. Trends Parasitol 2003;19(6): 253–259.

2 Genton B, al-Yaman F, Beck HP, et al. The epidemiology ofmalaria in the Wosera area, East Sepik Province, Papua NewGuinea, in preparation for vaccine trials. I. Malariometricindices and immunity. Ann Trop Med Parasitol 1995; 89(4):359–376.

3 Koch R. Zweiter Bericht uber die Thatigkeit der Malaria-Expedition. Dtsch Med Wochenschr 1900; 26(5): 88–90. EnglishTranslation in Brit Med J Feb 10, 1900 325–327.

4 Koch R. Dritter Bericht uber die Thatigkeit der Malaria-Expe-dition. Dtsch Med Wochenschr 1900; 26(17): 281–284. EnglishTranslation in Brit. Med. J. May 12, 1900, 1183–1186.

5 Koch R. Vieter Bericht uber die Thatigkeit der Malaria-Expe-dition. Dtsch Med Wochenschr 1900; 26(25): 397–398. Englishtranslation in Brit. Med. J. June 30, 1900, 1597–1598.

6 Ewers W. Robert Koch, his work in New Guinea and hiscontribution to malariology. P N G Med J 1972; 15(2):117–124.

7 Schofield L & Mueller I. Clinical immunity to malaria. CurrMol Med 2006; 6: 205–221.

8 Clark IA & Schofield L. Pathogenesis of malaria. ParasitolToday 2000; 16(10): 451–454.

9 Baird JK, Jones T, Danudirgo E, et al. Age-dependentacquired protection against Plasmodium falciparum in peoplehaving two years exposure to hyperendemic malaria. Am JTrop Med Hyg 1991; 45(1): 65–76.

10 Baird JK. Age-dependent characteristics of protection v. sus-ceptibility to Plasmodium falciparum. Ann Trop Med Parasitol1998; 92(4): 367–390.

11 Doolan DL, Dobano C, Baird JK. Acquired immunity tomalaria. Clin Microbiol Rev 2009; 22(1): 13–36, Table of Con-tents.

12 Mehlotra RK, Lorry K, Kastens W, et al. Random distribu-tion of mixed species infections in Papua New Guinea. Am JTrop Med Hyg 2000; 62(2): 225–231.

13 Mehlotra RK, Kasehagen LJ, Baisor M, et al. Malaria infec-tions are randomly distributed in diverse holoendemic areas ofPapua New Guinea. Am J Trop Med Hyg 2002; 67(6): 555–562.

14 Mueller I, Widmer S, Michel D, et al. High sensitivity detec-tion of Plasmodium species reveals positive correlationsbetween infections of different species, shifts in age distributionand reduced local variation in Papua New Guinea. Malar J2009; 8: 41.

15 Smith T, Genton B, Baea K, et al. Prospective risk of morbid-ity in relation to malaria infection in an area of high endemi-city of multiple species of Plasmodium. Am J Trop Med Hyg2001; 64(5–6): 262–267.

16 Smith T, Hii JL, Genton B, et al. Associations of peak shiftsin age-prevalence for human malarias with bednet coverage.Trans R Soc Trop Med Hyg 2001; 95: 1–6.

17 Bruce MC, Donnelly CA, Packer M, et al. Age- and species-specific duration of infection in asymptomatic malaria infec-tions in Papua New Guinea. Parasitology 2000; 121(Pt 3): 247–256.

18 Bruce MC, Donnelly CA, Alpers MP, et al. Cross-species inter-actions between malaria parasites in humans. Science 2000;287(5454): 845–848.

D. I. Stanisic et al. Parasite Immunology

630 � 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632

19 Mueller I, Genton B, Betuela I, et al. Vaccines against malaria:perspectives from Papua New Guinea. Hum Vaccin 2010; 6(1):17–20.

20 Michon P, Cole-Tobian J, Dabod E, et al. The risk of malarialinfections and disease in Papua New Guinean children. Am JTrop Med Hyg 2007; 76(6): 997–1008.

21 Lin E, Kiniboro B, Gray L, et al. Differential patterns ofinfection and disease with P. falciparum and P. vivax in youngPapua New Guinean children. PLoS ONE 2010; 5(2): e9047.

22 Salanti A, Staalsoe T, Jensen AT, et al. Selective upregulationof a single distinctly structured var gene in chondroitin sul-phate A-adhering Plasmodium falciparum involved in preg-nancy-associated malaria. Mol Microbiol 2003; 49(1): 179–191.

23 Jensen AT, Magistradi P, Sharp S, et al. Plasmodium falci-parum associated with severe childhood malaria preferentiallyexpresses PfEMP1 encoded by group A var genes. J Exp Med2004; 199(9): 1179–1190.

24 D’Ombrain M, Voss T, Maier A, et al. Plasmodium falciparumerythrocyte membrane protein-1 specifically suppresses earlyproduction of host interferon gamma. Cell Host Microbe 2007;2: 130–138.

25 Schofield L & Tachado SD. Regulation of host cell functionby glycosylphosphatidylinositols of the parasitic protozoa.Immunol Cell Biol 1996; 74(6): 555–563.

26 Tachado SD, Mazhari-Tabrizi R & Schofield L. Specificity insignal transduction among glycosylphosphatidylinositols ofPlasmodium falciparum, Trypanosoma brucei, Trypanosomacruzi and Leishmania spp. Parasite Immunol 1999; 21(12): 609–617.

27 Schofield L, Hewitt MC, Evans K, et al. Synthetic GPI as acandidate anti-toxic vaccine in a model of malaria. Nature2002; 418(6899): 785–789.

28 Naik RS, Branch OH, Woods AS, et al. Glycosylphosphatidyli-nositol anchors of Plasmodium falciparum: molecular charac-terization and naturally elicited antibody response that mayprovide immunity to malaria pathogenesis. J Exp Med 2000;192(11): 1563–1576.

29 Hudson Keenihan SN, Ratiwayanto S, Soebianto S, et al. Age-dependent impairment of IgG responses to glycosylphosphati-dylinositol with equal exposure to Plasmodium falciparumamong Javanese migrants to Papua, Indonesia. Am J TropMed Hyg 2003; 69(1): 36–41.

30 Boutlis CS, Gowda DC, Naik RS, et al. Antibodies to Plasmo-dium falciparum glycosylphosphatidylinositols: inverse associa-tion with tolerance of parasitemia in Papua New Guineanchildren and adults. Infect Immun 2002; 70(9): 5052–5057.

31 Luty AJ, Lell B, Schmidt-Ott R, et al. Interferon-gammaresponses are associated with resistance to reinfection withPlasmodium falciparum in young African children. J Infect Dis1999; 179(4): 980–988.

32 Al-Yaman F, Genton B, Taraika J, et al. Association betweencellular response (IL-4) to RESA ⁄ Pf155 and protection fromclinical malaria among Papua New Guinean children living ina malaria endemic area. Parasite Immunol 1997; 19: 249–254.

33 Al-Yaman F, Genton B, Taraika J, et al. Cellular immunity tomerozoite surface protein 2 (FC27 and 3D7) in Papua NewGuinean children. Temporal variation and relation to clinicaland parasitological status. Parasite Immunol 1997; 19: 207–214.

34 Baird JK & Snow RW. Acquired immunity in a holoendemicsetting of Plasmodium falciparum and P. vivax malaria. AmJ Trop Med Hyg 2007; 76(6): 995–996.

35 D’Ombrain M, Robinson L, Stanisic D, et al. Association ofearly interferon-gamma production with immunity to clinicalmalaria: a longitudinal study among Papua New Guinean chil-dren. Clin Infect Dis 2008; 47(11): 1380–1387.

36 Robinson L, D’Ombrain M, Stanisic D, et al. Cellular tumournecrosis factor, gamma interferon, and interleukin-6 responsesas correlates of immunity and risk of clinical Plasmodium falci-parum malariia in children from Papua New Guinea. InfectImmun 2009; 77(7): 3033–3043.

37 Xainli J, Baisor M, Kastens W, et al. Age-dependent cellularimmune responses to Plasmodium vivax Duffy binding proteinin humans. J Immunol 2002; 169: 3200–3207.

38 King C, Michon P, Shakri A, et al. Naturally acquired Duffy-binding protein-specific binding inhibitory antibodies conferprotection from blood-stage Plasmodium vivax infection. ProcNatl Scad Sci U S A 2008; 105(24): 8363–8368.

39 Cohen S, McGregor IA, Carrington S, et al. Gamma-globulinand acquired immunity to human malaria. Nature 1961; 192:733–737.

40 Genton B, Anders RF, Alpers MP, et al. The malaria vaccinedevelopment program in Papua New Guinea. Trends Parasitol2003; 19(6): 264–270.

41 Al-Yaman F, Genton B, Anders RF, et al. Relationshipbetween humoral response to Plasmodium falciparum merozoitesurface antigen-2 and malaria morbidity in a highly endemicarea of Papua New Guinea. Am J Trop Med Hyg 1994; 51(5):593–602.

42 Al-Yaman F, Genton B, Falk M, et al. Humoral response toPlasmodium falciparum ring-infected erythrocyte surface anti-gen in a highly endemic area of Papua New Guinea. Am JTrop Med Hyg 1995; 52(1): 66–71.

43 Al-Yaman F, Genton B, Kramer KJ, et al. Acquired antibodylevels to Plasmodium falciparum merozoite surface antigen 1 inresidents of a highly endemic area of Papua New Guinea.Trans R Soc Trop Med Hyg 1995; 89(5): 555–559.

44 Beck HP, Felger I, Genton B, et al. Humoral and cell-mediatedimmunity to the Plasmodium falciparum ring-infected erythro-cyte surface antigen in an adult population exposed to highlyendemic malaria. Infect Immun 1995; 63: 596–600.

45 Beck HP, Felger I, Kabintik S, et al. Assessment of the humoraland cell-mediated immunity against the Plasmodium falciparumvaccine candidates circumsporozoite protein and SPf66 inadults living in highly endemic malarious areas of Papua NewGuinea. Am J Trop Med Hyg 1994; 51(3): 356–364.

46 Connelly M, King CL, Bucci K, et al. T-cell immunity to pep-tide epitopes of liver-stage antigen 1 in an area of Papua NewGuinea in which malaria is holoendemic. Infect Immun 1997;65(12): 5082–5087.

47 Al-Yaman F, Genton B, Kramer K, et al. Assessment of therole of naturally acquired antibody levels to Plasmodium falci-parum merozoite surface protein-1 in protecting Papua NewGuinean children from malaria morbidity. Am J Trop MedHyg 1996; 54(5): 443–448.

48 Al-Yaman F, Genton B, Anders R, et al. Assessment of therole of the humoral response to Plasmodium falciparum MSP2compared to RESA and SPf66 in protecting Papua new Gui-nean children from clinical malaria. Parasite Immunol 1995;17: 493–501.

49 Stanisic DI, Richards J, McCallum F, et al. IgG subclass-speci-fic responses against Plasmodium falciparum merozoite antigensare associated with control of parasitemia and protection fromsymptomatic illness. Infect Immun 2009; 77(3): 1165–1174.

Volume 32, Number 8, August 2010 Malaria Immunology Research in PNG

� 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632 631

50 Malkin E, Diemert D, McArthur J, et al. Phase 1 clinical trialof apical membrane antigen 1: an asexual blood-stage vaccinefor Plasmodium falciparum malaria. Infect Immun 2005; 73(6):3677–3685.

51 Forsyth KP, Philip G, Smith T, et al. Diversity of antigensexpressed on the surface of erythrocytes infected with maturePlasmodium falciparum parasites in Papua New Guinea. Am JTrop Med Hyg 1989; 41(3): 259–265.

52 Cole-Tobian J, Michon P, Baisor M, et al. Strain-specific Duffybinding protein antibodies correlate with protection againstinfection with homologous compared to heterologous Plasmo-dium vivax strains in Papua New Guinean children. InfectImmun, 2009; 77(9): 4009–4017.

53 Grimberg BT, Udomsangpetch R, Xainli J, et al. Plasmodiumvivax invasion of human erythrocytes inhibited by antibodiesdirected against the Duffy binding protein. PLoS Med 2007;4(12): e337.

54 Genton B, Betuela I, Felger I, et al. A recombinant blood-stage malaria vaccine reduces Plasmodum falciparum densityand exerts selective pressure on parasite populations in a phase1-2b trial in Papua New Guinea. J Infect Dis 2002; 185: 820–827.

55 Genton B, Al-Yaman F, Betuela I, et al. Safety and immuno-genicity of a three-component blood-stage malaria vaccine(MSP1, MSP2, RESA) against Plasmodium falciparum inPapua New Guinean children. Vaccine 2003; 22(1): 30–41.

56 Fluck C, Smith T, Beck H, et al. Strain-specific humoralresponse to a polymorphic malaria vaccine. Infect Immun 2004;72(11): 6300–6305.

57 Hommel M, Elliott SR, Soma V, et al. Evaluating the anti-genic diversity of placental binding Plasmodium falciparum var-iants and the antibody repertoire among pregnant women.Infect Immun 2010; 78(5): 1963–1978.

D. I. Stanisic et al. Parasite Immunology

632 � 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 623–632