BioMed CentralMalaria Journal

ss

Open AcceResearchCohort study of the association of antibody levels to AMA1 MSP119 MSP3 and GLURP with protection from clinical malaria in Ghanaian childrenDaniel Dodoo1 Anastasia Aikins1 Kwadwo Asamoah Kusi12 Helena Lamptey1 Ed Remarque2 Paul Milligan3 Samuel Bosomprah4 Roma Chilengi5 Yaa Difie Osei6 Bartholomew Dicky Akanmori1 and Michael Theisen7

Address 1Noguchi Memorial Institute for Medical Research University of Ghana Legon Ghana 2Department of Parasitology Biomedical Primate Research Centre Lange Kleiweg 139 2288 GJ Rijswijk The Netherlands 3Department of Infectious and Tropical Diseases London School of Hygiene and Tropical Medicine Keppel Street London WC1E 7HT UK 4Ministry of Health PO Box M44 Accra Ghana 5The African Malaria Network Trust Tanzania Commission for Science and Technology Building PO Box 33207 Dar es Salaam Tanzania 6Department of Biochemistry University of Ghana Legon Ghana and 7Department of Infectious Disease Immunology State Serum Institute Copenhagen Denmark

Email Daniel Dodoo - ddodoonoguchimimcomorg Anastasia Aikins - aocrannoguchimimcomorg Kwadwo Asamoah Kusi - kusibprcnl Helena Lamptey - hnarteynoguchimimorg Ed Remarque - Remarquebprcnl Paul Milligan - PaulMilliganlshtmacuk Samuel Bosomprah - sbosomprahyahoocom Roma Chilengi - chilengiamanet-trustorg Yaa Difie Osei - yaaoseiugedugh Bartholomew Dicky Akanmori - BAkanmorinoguchimimcomorg Michael Theisen - mthssidk

Corresponding author

AbstractBackground Antigen-specific antibody-mediated immune responses play an important role innatural protection against clinical malaria but conflicting estimates of this association have emergedfrom immuno-epidemiological studies in different geographical settings This study was aimed atassessing in a standardized manner the relationship between the antibody responses to four malariavaccine candidate antigens and protection from clinical malaria in a cohort of Ghanaian children

Methods Standardized ELISA protocols were used to measure isotype and IgG subclass levels toApical Membrane Antigen 1 (AMA1) Merozoite Surface Protein 1ndash19 (MSP119) Merozoite SurfaceProtein 3 (MSP3) and Glutamate Rich Protein (GLURP) antigens in plasma samples from 352Ghanaian children aged three to 10 years with subsequent malaria surveillance for nine monthsThis is one of a series of studies in different epidemiological settings using the same standardizedELISA protocols to permit comparisons of results from different laboratories

Results The incidence rate of malaria was 035 episodes per child per year Isotype and IgGsubclasses for all antigens investigated increased with age while the risk of malaria decreased withage After adjusting for age higher levels of IgG to GLURP MSP119 MSP3 and IgM to MSP119 MSP3and AMA1 were associated with decreased malaria incidence Of the IgG subclasses only IgG1 toMSP119 was associated with reduced incidence of clinical malaria A previous study in the samelocation failed to find an association of antibodies to MSP119 with clinical malaria The disagreementmay be due to differences in reagents ELISA and analytical procedures used in the two studies

Published 29 July 2008

Malaria Journal 2008 7142 doi1011861475-2875-7-142

Received 12 March 2008Accepted 29 July 2008

This article is available from httpwwwmalariajournalcomcontent71142

copy 2008 Dodoo et al licensee BioMed Central Ltd This is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby20) which permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Page 1 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

When IgG IgM and IgG subclass levels for all four antigens were included in a combined modelonly IgG1 [(080 (067ndash097) p = 0018)] and IgM [(048 (032ndash072) p lt 0001)] to MSP119 wereindependently associated with protection from malaria

Conclusion Using standardized procedures the study has confirmed the importance ofantibodies to MSP119 in reducing the risk of clinical malaria in Ghanaian children thus substantiatingits potential as a malaria vaccine candidate

BackgroundMalaria remains one of the most important causes of mor-bidity and mortality in the world Current methods ofcontrol are only partially effective and therefore thedevelopment of a vaccine which can provide a high degreeof protection is a priority Antibody-mediated immuneresponses to malaria antigens are known to be involved inprotecting against disease [1-4] but the antigens thatinduce protective antibodies have not been conclusivelyidentified Immuno-epidemiological studies from differ-ent laboratories have sometimes yielded conflictingresults [5-8] This may be partly due to differences inmalaria endemicity and the use of different study designsreagents assay protocols and statistical methodologies Inan attempt to make such studies more comparable theAfro-Immuno Assay (AIA) network project was initiatedThe network includes six African Institutions in GabonGhana Burkina Faso Senegal Tanzania and Zimbabweand three European Institutions from Denmark TheNetherlands and France The Afro-Immuno Assay networkhas developed standardized enzyme immuno assays [9-11] that ensure the use of the same reagents protocols andstatistical methods to assess the association betweenacquisition of malaria specific antibody responses to fourpotential malaria vaccine candidate antigens and possibleprotection from clinical malaria Samples for the AIAmulti-center project were retrospectively obtained fromcohort studies in six different geographical and epidemio-logical settings comprising low endemic to holoendemicareas These antigens include the Glutamate Rich Protein(GLURP) the Merozoite Surface Protein 3 (MSP3) [12]the 19-kilo Dalton region of the Merozoite Surface Pro-tein 1 (MSP119) [13] and the Apical Membrane Antigen 1(AMA1) [14] which are all thought to induce protectiveantibody responses through various mechanisms [15-18]Vaccines incorporating these antigens are currently in clin-ical trials and are described in detail elsewhere [719-26]It is likely that a future malaria vaccine will comprise mul-tiple rather than single antigens and it is therefore usefulto study natural immune responses to multiple malariaantigens in relation to incidence of malaria in a morestandardized way In this study the standardized AIAELISA procedures [9-11] were used to assess the relation-ship between incidence of clinical malaria and naturallyacquired isotype and IgG subclass responses to these fourleading malaria vaccine candidate antigens AMA1

MSP119 MSP3 and GLURP in Ghanaian children fromthree to 10 years of age

Materials and methodsStudy area study population and morbidity surveillanceSamples used in this study were obtained in March 2002from a longitudinal study conducted in Dodowa inwhich 352 children aged three to 10 years (in the activephase of acquiring immunity to malaria) were enrolled ina study whose original aim had been to assess the role ofcytokine regulation and immunity to malaria Dodowa isa semi-rural town in the Dangme West District of theGreater Accra Region of Ghana about 50 km from thecapital Accra and is an area of moderate and stablemalaria transmission with a seasonal peak Bed net cover-age in this area was low about 10 [27] The study wasapproved by the Noguchi Memorial Institutes EthicalReview Board After obtaining consent from parentsblood was obtained from each child and plasma stored at-20degC until use Malaria episodes were detected usingboth active and passive surveillance implemented over aperiod of nine-months spanning the entire malaria trans-mission season Clinical and parasitological informationwas captured using a standard questionnaire Each childwas visited once a week and the child and the parents orguardians were asked about symptoms of malaria sincethe last visit and whether shehe had received any anti-malarial treatment The child was then given a physicalexamination and the body temperature measured Chil-dren with a history of fever within 48 hours andor axil-lary temperature equal to or above 375degC had a rapid testfor malaria parasitaemia using OptiMALtrade (DiaMedFLOW Inc Portland Oregon) and then thick and thinblood films were prepared for microscopy for estimatingthe parasite density which was used later in the more spe-cific malaria case definition employed for data analysisOnly children with measured or reported fever and witha positive rapid test were treated with chloroquine (inaccordance with the then prevailing national malariatreatment policy) and in the case of severe symptoms thechild was referred to the hospital At monthly intervalsblood smears from finger pricks were obtained from allchildren irrespective of symptoms to estimate the preva-lence of asymptomatic malaria infections The parents ofthe children were instructed to report to the field team ifthe child had any symptoms of disease at any time The

Page 2 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

qualifying case definition for malaria in the data analysiswas reported fever andor a measured temperature equalto or above 375degC with parasitaemia ge 5000 parasitesμl of blood this case definition has been found to have90 sensitivity and specificity in the study area [28-30]Positive and negative control plasma used in ELISA meas-urements were obtained from adult Liberians and Danesrespectively

Malaria antigensAMA1 was from the Pichia pastoris expressed ectodomainof Plasmodium falciparum FVO strain comprising aminoacids 25ndash545 [31] (Donated by A Thomas BiomedicalPrimate Research Centre) GLURP was an Escherichia colirecombinant protein containing the conserved non-repeat N-terminal region (amino acids 25ndash514) called R0[32] (Donated by M Theisen Statens Serum Institut)MSP119 was a Baculovirus antigen of the C-terminal regionof the merozoite protein surface 1 produced in insect cellsinfected with a recombinant Baculovirus containing a syn-thetic G-C enriched PfMSP1 gene (Palo Alto allele) cod-ing for 43 N-terminal MSP1 precursor residues and 16amino acid residues upstream of the classical MSP-119(NIS---FCS) [33] (Donated by S Longacre Institut Pas-teur) The MSP3 antigen used in this study was a long syn-thetic peptide called LR55 (amino acids 181 ndash 276) of themerozoite surface protein 3 [34] (Donated by M TheisenStatens Serum Institut) All the antigens were providedthrough the AIA Project

Enzyme-linked immunosorbent assay (ELISA)Specific isotype and IgG subclass levels against GLURPMSP119 MSP3 and AMA1 were measured using indirectELISA according to the AIA standard ELISA protocols [9-11] All antigens tested were optimized and shown to bestable for at least three weeks when antigen-coated platesand serumplasma dilutions are refrigerated The subclassspecific reagents used were selected on the basis of lowcross reactivities among themselves To control for inter-assay and day-to-day variations in the standardized ELISAprocedure three-fold serial dilutions of reference stand-ard reagents (IgG IgM and IgG1 to IgG4) were directlycoated on each ELISA plate (Maxisorp Nunc Denmark) ata start concentration of 1000 ngml (100 μlwell) ODvalues for the test samples were converted into antibodyunits with the standard reference curves generated foreach ELISA plate using a four parameter curve-fit Micro-soft Excel-based application Samples were re-tested if thecoefficient of variation between duplicate absorbance val-ues were higher than 15 and plates were also re-tested ifthe R-square value of the standard curve was less than97 The reference standards PBS buffer blank positiveand negative control plasma pools that were included ineach ELISA test plate allow for the determination of failed

assay runs The AIA ELISA procedure used in this study isdescribed in detail elsewhere [10]

Statistical analysisClinical data were double entered using Microsoft Fox Proand immunological data using Excel STATA version 92(Statcorp Texas) was used for statistical analysis Childrenwere considered to have a clinical malaria episode if theyhad parasitaemia of ge 5000 parasitesμl with a measuredtemperature ge 375degC or a history of fever in the last 48hours [63536]

For each antigen Poisson regression was used to investi-gate the association between the levels of antibody meas-ured at baseline and the incidence rate of the first (oronly) episode of clinical malaria The level of total IgGIgM and each IgG subclass were analysed for each antigenin turn Antibody values were transformed to log base 2so that the rate ratio represents the ratio of malaria inci-dence corresponding to a doubling of antibody level Toinvestigate whether the relationship between malaria inci-dence and antibody level was nonlinear a likelihood ratiotest was used to compare the fit of the model when anti-body level was included as a categorical or a continuousvariable When there were zero antibody values indicat-ing levels below the detection limit the zero (left cen-sored) values were assigned a nominal value equal to halfthe smallest measured value for that variable If the pro-portion of zero values was large the variable was treatedas categorical with the reference category containing thezero values and the positive values divided into threeequal groups A likelihood ratio test was used to deter-mine the P-value for the association with malaria inci-dence Age at enrolment was considered to be animportant potential confounder and was included in theregressions as a factor with categories defined by quintilesTo model seasonality in malaria incidence the calendarmonth of surveillance was included in the models as a fac-tor To construct a parsimonious model using all theimmunological variables firstly a model was producedfor each antigen in this model each IgG subclass total IgGand IgM were candidates for inclusion provided the P-value for association with malaria incidence was 01 orless when considered individually Variables were thenremoved from the model if the P-value for the likelihoodratio test was more than 01 provided removal did notchange coefficients of variables in the model by more than10 In a second stage the variables included in thesemodels were candidates for inclusion in a final modelderived in a similar way Baseline parasitaemia was notconsidered as a potential confounder but the interactionbetween each immunological variable and the presence ofparasitaemia at baseline in their effects on malaria inci-dence was examined A consequence of using a more spe-cific case definition in the analysis than was used to decide

Page 3 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

treatment during the study is that children could havereceived anti-malarial treatments during the period theyare considered at risk in the analysis To explore theimpact of these drug treatments a time dependent varia-ble was defined to allow for a reduced risk of malaria fora period of 28 days after each drug treatment which wasincluded as a covariate in the Poisson regression modelLOESS smoothing was applied in R software to plot anti-body levels in relation to age Spearmans rank correlationtest was used to assess associations between antibody lev-els and age

ResultsPattern of P falciparum infections and malaria in the study cohortOf the 352 children recruited for the original study eightwere lost to follow-up immediately after the baselineblood sampling Of the 344 children followed up sixtyfour (19) had at least one episode of malaria (53 chil-dren had one episode nine had two and two had threeepisodes) The incidence rate of malaria in the studycohort was 035 attacks per child per year (Table 1) Therisk of clinical malaria decreased with increasing age [37]Sixty-six percent of the children had asymptomatic parasi-taemia at baseline Parasites were predominantly P falci-parum (95) and the prevalence of parasitaemiameasured each month was roughly constant (rangingfrom 50 to 65) The incidence of clinical malariahowever varied during the survey period rising graduallyfrom March to May peaking in July and then decreasinguntil November (Figure 1)

Relationship between age and antibody levelsSeveral studies in malaria endemic regions have shownincreasing antibody levels with age This pattern is morepronounced the greater the intensity and duration ofmalaria transmission In this study the relationshipbetween antibody levels and age were assessed for anti-bodies against MSP119 AMA1 GLURP and MSP3 Thelevels of IgG and IgM to MSP119 MSP3 AMA1 andGLURP increased with age (Spearman correlation coeffi-cient (rs) 021 ndash 045 p lt 0001 Figure 2) IgG levels toAMA1 however gradually increased until six years of age

and then leveled off For the four antigens tested IgG1IgG2 and IgG3 significantly increased with age (rs 012 ndash036 p lt 003) with the exception of IgG2 to AMA1 Therewas no evidence that the level of IgG4 was associated withage for any of the antigens Like IgG to AMA1 IgG3 toMSP3 steadily increased with age until seven years of agethen leveled off (Figure 3)

The levels of both IgG and IgG1 were highest for AMA1and lowest for MSP3 whereas IgG levels to MSP119 andGLURP were comparable (Figures 2 and 3) The levels ofIgG3 were higher in AMA1 than the comparable levels inGLURP MSP119 and MSP3 (Figure 3) In general the lev-els of cytophilic IgG1 and IgG3 were higher than those ofnon-cytophilic IgG2 and IgG4 IgG4 levels being the low-est

Antibody levels in relation to protection from clinical malariaTotal IgG to MSP3 MSP119 and GLURP and IgM to allfour antigens tested (MSP3 MSP119 GLURP and AMA1)were associated with reduced malaria incidence in crudeanalyses (Table 2) The incidence of both clinical malariaand antibody levels were associated with age age is there-fore a potential confounder and it is important to adjustfor its effects After adjusting for the effect of age there wasevidence of a significant association between total IgG toMSP3 MSP119 GLURP and reduced risk of malaria (IgGto MSP3 rate ratio 069 (95CI 053 090) P = 001 IgGto MSP119 075 (061 092) P = 001 IgG to GLURP 079(064 098) P = 004) and IgM levels to AMA1 MSP3MSP119 were also significantly associated with reducedrisk of malaria (Table 2)

A large proportion of the measurements of IgG4 to MSP3and AMA1 and IgG2 to AMA1 were zero (left-censored)values and so these variables were treated as categoricalvariables (Table 3) In the crude analysis IgG1 IgG2 andIgG3 to MSP3 MSP119 GLURP and also IgG4 to MSP3were associated with a reduced risk of clinical malaria(Table 3 4) but after adjustment for age only one ofthese variables IgG1 to MSP119 remained significantly

Table 1 Malaria incidence by age group

Age in yrs No of children

Cum incidence of malaria

no episodes of malaria (years at risk)

Incidence rate per child year (95CI)

3 32 1232 (38) 15 (1952) 079 (048ndash131)4ndash5 86 2486 (28) 28 (5475) 051 (035ndash074)6ndash7 79 1179 (14) 13 (5138) 025 (015ndash044)8 48 648 (13) 7 (3111) 022 (011ndash047)

9ndash10 99 1199 (11) 14 (6122) 023 (014ndash039)

Total 344 64344 (19) 77 (21748) 035 (028ndash044)

Page 4 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

associated with malaria incidence (rate ratio 089 (95CI080 099) P = 004 Table 4)

When all immunological variables were consideredsimultaneously only two variables were independentlyassociated with reduced malaria incidence IgG1 [(080(066ndash096) p = 0018)] to MSP119 and IgM [(048(032ndash072) p lt 0001)] to MSP119 (Table 5)

The estimated rate ratio for the effect of antimalarial treat-ments given to children with parasitaemia lt 5000uL was025 (95CI 006ndash105) indicating these children had asubstantially reduced risk of being found positive withparasitaemia gt= 5000ul during the 28 days following thetreatment But the estimates of rate ratios for other varia-bles in the model were unchanged when this variable wasincluded in the model suggesting treatment effects did not

cause a bias in the estimation of the effects of immunolog-ical variables

DiscussionThis study in Ghanaian children is one of a series of stud-ies designed to assess using standardized methods theassociation of antibody levels to four leading asexualblood-stage malaria antigens (MSP119 MSP3 AMA1 andGLURP) with the incidence of clinical malaria in differentepidemiological settings Previous results have been diffi-cult to interpret due to different study protocols and ana-lytical methods having been used [5-79142938-40] Inthis study the prevalence of asymptomatic malaria para-sitaemia was relatively high and stable while incidence ofclinical malaria fluctuated in parallel with the intensity oftransmission and seasonal rainfall pattern These patternsare typical of this area and have been reported in previous

The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malariaFigure 1The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malaria The point prevalence of asymptomatic parasitaemia for each month is represented as a bar graph and the pattern of clinical malaria (incident rate) is shown as a line graph

Page 5 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

studies [641-43] The variation in the incidence of clinicalmalaria during the study period may be due to the intro-duction of new parasites with different antigenic presenta-tion into the population leading to clinical malaria insusceptible individuals The risk of malaria decreased withage while isotype and IgG subclass levels to the four anti-gens generally increased with age This is consistent withthe hypothesis that immunity to malaria is largely effectedthrough antibody-mediated mechanisms and that protec-tive antibody levels to relevant antigens increase with age-related exposure to the parasites [44] Increasing IgG andIgM levels with age may reflect greater cumulative expo-sure of older children but may also be due to older chil-dren having a more mature immune system [45] Theassociation of IgM responses with reduced malaria inci-dence indicates a possible role in immunity in Ghanaianchildren Although much emphasis has been placed onIgG as the important isotype in immunity against malariaIgM which has lower affinity but is multivalent mayafford protection via other mechanisms such as the block-ing of merozoite invasion of erythrocytes complementactivation agglutination of merozoites [46]

The association with malaria incidence of IgG responsesto MSP3 MSP119 and GLURP is consistent with data fromseveral other immuno-epidemiological studies [57-91229303847-49] indicating that these antigens maybe targets of protective antibodies [124] There was noevidence that IgG levels to AMA1 were associated with

malaria incidence and there was no evidence of an inter-action with baseline parasitaemia in contrast with similarstudies conducted in Kenya [14] As shown in other stud-ies the cytophilic antibody levels to the four antigenstested in this study were higher than the non-cytophilicones emphasizing their importance in anti-malariaimmunity [929385051] IgG and cytophilic antibodylevels were highest for AMA1 while the levels were rela-tively low for MSP3 These differences in specific antibodylevels may be related to the number of immunogenic B-Cell epitopes exposed to the immune system and couldalso be related to the structure location and function ofthe particular antigen(s) With the exception of IgG2 lev-els to AMA1 IgG2 levels to GLURP MSP119 and MSP3increased with age which may suggest that IgG2 isinvolved in immunity against malaria In recent studiesmalaria antigen specific IgG2 have been shown to bindwith high affinity to mutant Fcγ RII H131-gtR receptors[52] on monocytes granulocytes and B cells thus afford-ing protection against malaria through monocytes and orneutrophil mediated mechanisms in subjects expressingthe mutant CD32 form [53] There was however no evi-dence found that IgG2 was associated with malaria inci-dence for any of the four antigens tested

The IgG subclasses IgG1 IgG2 and IgG3 for MSP3MSP119 GLURP and IgG4 to MSP3 were associated with areduced risk of malaria in un-adjusted analysis but ofthese only IgG1 to MSP119 was independently associated

Baseline IgG and IgM levels in relation to ageFigure 2Baseline IgG and IgM levels in relation to age The top and bottom panels represents total IgG or IgM levels to AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 6 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

qualifying case definition for malaria in the data analysiswas reported fever andor a measured temperature equalto or above 375degC with parasitaemia ge 5000 parasitesμl of blood this case definition has been found to have90 sensitivity and specificity in the study area [28-30]Positive and negative control plasma used in ELISA meas-urements were obtained from adult Liberians and Danesrespectively

Malaria antigensAMA1 was from the Pichia pastoris expressed ectodomainof Plasmodium falciparum FVO strain comprising aminoacids 25ndash545 [31] (Donated by A Thomas BiomedicalPrimate Research Centre) GLURP was an Escherichia colirecombinant protein containing the conserved non-repeat N-terminal region (amino acids 25ndash514) called R0[32] (Donated by M Theisen Statens Serum Institut)MSP119 was a Baculovirus antigen of the C-terminal regionof the merozoite protein surface 1 produced in insect cellsinfected with a recombinant Baculovirus containing a syn-thetic G-C enriched PfMSP1 gene (Palo Alto allele) cod-ing for 43 N-terminal MSP1 precursor residues and 16amino acid residues upstream of the classical MSP-119(NIS---FCS) [33] (Donated by S Longacre Institut Pas-teur) The MSP3 antigen used in this study was a long syn-thetic peptide called LR55 (amino acids 181 ndash 276) of themerozoite surface protein 3 [34] (Donated by M TheisenStatens Serum Institut) All the antigens were providedthrough the AIA Project

Enzyme-linked immunosorbent assay (ELISA)Specific isotype and IgG subclass levels against GLURPMSP119 MSP3 and AMA1 were measured using indirectELISA according to the AIA standard ELISA protocols [9-11] All antigens tested were optimized and shown to bestable for at least three weeks when antigen-coated platesand serumplasma dilutions are refrigerated The subclassspecific reagents used were selected on the basis of lowcross reactivities among themselves To control for inter-assay and day-to-day variations in the standardized ELISAprocedure three-fold serial dilutions of reference stand-ard reagents (IgG IgM and IgG1 to IgG4) were directlycoated on each ELISA plate (Maxisorp Nunc Denmark) ata start concentration of 1000 ngml (100 μlwell) ODvalues for the test samples were converted into antibodyunits with the standard reference curves generated foreach ELISA plate using a four parameter curve-fit Micro-soft Excel-based application Samples were re-tested if thecoefficient of variation between duplicate absorbance val-ues were higher than 15 and plates were also re-tested ifthe R-square value of the standard curve was less than97 The reference standards PBS buffer blank positiveand negative control plasma pools that were included ineach ELISA test plate allow for the determination of failed

assay runs The AIA ELISA procedure used in this study isdescribed in detail elsewhere [10]

Statistical analysisClinical data were double entered using Microsoft Fox Proand immunological data using Excel STATA version 92(Statcorp Texas) was used for statistical analysis Childrenwere considered to have a clinical malaria episode if theyhad parasitaemia of ge 5000 parasitesμl with a measuredtemperature ge 375degC or a history of fever in the last 48hours [63536]

For each antigen Poisson regression was used to investi-gate the association between the levels of antibody meas-ured at baseline and the incidence rate of the first (oronly) episode of clinical malaria The level of total IgGIgM and each IgG subclass were analysed for each antigenin turn Antibody values were transformed to log base 2so that the rate ratio represents the ratio of malaria inci-dence corresponding to a doubling of antibody level Toinvestigate whether the relationship between malaria inci-dence and antibody level was nonlinear a likelihood ratiotest was used to compare the fit of the model when anti-body level was included as a categorical or a continuousvariable When there were zero antibody values indicat-ing levels below the detection limit the zero (left cen-sored) values were assigned a nominal value equal to halfthe smallest measured value for that variable If the pro-portion of zero values was large the variable was treatedas categorical with the reference category containing thezero values and the positive values divided into threeequal groups A likelihood ratio test was used to deter-mine the P-value for the association with malaria inci-dence Age at enrolment was considered to be animportant potential confounder and was included in theregressions as a factor with categories defined by quintilesTo model seasonality in malaria incidence the calendarmonth of surveillance was included in the models as a fac-tor To construct a parsimonious model using all theimmunological variables firstly a model was producedfor each antigen in this model each IgG subclass total IgGand IgM were candidates for inclusion provided the P-value for association with malaria incidence was 01 orless when considered individually Variables were thenremoved from the model if the P-value for the likelihoodratio test was more than 01 provided removal did notchange coefficients of variables in the model by more than10 In a second stage the variables included in thesemodels were candidates for inclusion in a final modelderived in a similar way Baseline parasitaemia was notconsidered as a potential confounder but the interactionbetween each immunological variable and the presence ofparasitaemia at baseline in their effects on malaria inci-dence was examined A consequence of using a more spe-cific case definition in the analysis than was used to decide

Page 3 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

treatment during the study is that children could havereceived anti-malarial treatments during the period theyare considered at risk in the analysis To explore theimpact of these drug treatments a time dependent varia-ble was defined to allow for a reduced risk of malaria fora period of 28 days after each drug treatment which wasincluded as a covariate in the Poisson regression modelLOESS smoothing was applied in R software to plot anti-body levels in relation to age Spearmans rank correlationtest was used to assess associations between antibody lev-els and age

ResultsPattern of P falciparum infections and malaria in the study cohortOf the 352 children recruited for the original study eightwere lost to follow-up immediately after the baselineblood sampling Of the 344 children followed up sixtyfour (19) had at least one episode of malaria (53 chil-dren had one episode nine had two and two had threeepisodes) The incidence rate of malaria in the studycohort was 035 attacks per child per year (Table 1) Therisk of clinical malaria decreased with increasing age [37]Sixty-six percent of the children had asymptomatic parasi-taemia at baseline Parasites were predominantly P falci-parum (95) and the prevalence of parasitaemiameasured each month was roughly constant (rangingfrom 50 to 65) The incidence of clinical malariahowever varied during the survey period rising graduallyfrom March to May peaking in July and then decreasinguntil November (Figure 1)

Relationship between age and antibody levelsSeveral studies in malaria endemic regions have shownincreasing antibody levels with age This pattern is morepronounced the greater the intensity and duration ofmalaria transmission In this study the relationshipbetween antibody levels and age were assessed for anti-bodies against MSP119 AMA1 GLURP and MSP3 Thelevels of IgG and IgM to MSP119 MSP3 AMA1 andGLURP increased with age (Spearman correlation coeffi-cient (rs) 021 ndash 045 p lt 0001 Figure 2) IgG levels toAMA1 however gradually increased until six years of age

and then leveled off For the four antigens tested IgG1IgG2 and IgG3 significantly increased with age (rs 012 ndash036 p lt 003) with the exception of IgG2 to AMA1 Therewas no evidence that the level of IgG4 was associated withage for any of the antigens Like IgG to AMA1 IgG3 toMSP3 steadily increased with age until seven years of agethen leveled off (Figure 3)

The levels of both IgG and IgG1 were highest for AMA1and lowest for MSP3 whereas IgG levels to MSP119 andGLURP were comparable (Figures 2 and 3) The levels ofIgG3 were higher in AMA1 than the comparable levels inGLURP MSP119 and MSP3 (Figure 3) In general the lev-els of cytophilic IgG1 and IgG3 were higher than those ofnon-cytophilic IgG2 and IgG4 IgG4 levels being the low-est

Antibody levels in relation to protection from clinical malariaTotal IgG to MSP3 MSP119 and GLURP and IgM to allfour antigens tested (MSP3 MSP119 GLURP and AMA1)were associated with reduced malaria incidence in crudeanalyses (Table 2) The incidence of both clinical malariaand antibody levels were associated with age age is there-fore a potential confounder and it is important to adjustfor its effects After adjusting for the effect of age there wasevidence of a significant association between total IgG toMSP3 MSP119 GLURP and reduced risk of malaria (IgGto MSP3 rate ratio 069 (95CI 053 090) P = 001 IgGto MSP119 075 (061 092) P = 001 IgG to GLURP 079(064 098) P = 004) and IgM levels to AMA1 MSP3MSP119 were also significantly associated with reducedrisk of malaria (Table 2)

A large proportion of the measurements of IgG4 to MSP3and AMA1 and IgG2 to AMA1 were zero (left-censored)values and so these variables were treated as categoricalvariables (Table 3) In the crude analysis IgG1 IgG2 andIgG3 to MSP3 MSP119 GLURP and also IgG4 to MSP3were associated with a reduced risk of clinical malaria(Table 3 4) but after adjustment for age only one ofthese variables IgG1 to MSP119 remained significantly

Table 1 Malaria incidence by age group

Age in yrs No of children

Cum incidence of malaria

no episodes of malaria (years at risk)

Incidence rate per child year (95CI)

3 32 1232 (38) 15 (1952) 079 (048ndash131)4ndash5 86 2486 (28) 28 (5475) 051 (035ndash074)6ndash7 79 1179 (14) 13 (5138) 025 (015ndash044)8 48 648 (13) 7 (3111) 022 (011ndash047)

9ndash10 99 1199 (11) 14 (6122) 023 (014ndash039)

Total 344 64344 (19) 77 (21748) 035 (028ndash044)

Page 4 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

associated with malaria incidence (rate ratio 089 (95CI080 099) P = 004 Table 4)

When all immunological variables were consideredsimultaneously only two variables were independentlyassociated with reduced malaria incidence IgG1 [(080(066ndash096) p = 0018)] to MSP119 and IgM [(048(032ndash072) p lt 0001)] to MSP119 (Table 5)

The estimated rate ratio for the effect of antimalarial treat-ments given to children with parasitaemia lt 5000uL was025 (95CI 006ndash105) indicating these children had asubstantially reduced risk of being found positive withparasitaemia gt= 5000ul during the 28 days following thetreatment But the estimates of rate ratios for other varia-bles in the model were unchanged when this variable wasincluded in the model suggesting treatment effects did not

cause a bias in the estimation of the effects of immunolog-ical variables

DiscussionThis study in Ghanaian children is one of a series of stud-ies designed to assess using standardized methods theassociation of antibody levels to four leading asexualblood-stage malaria antigens (MSP119 MSP3 AMA1 andGLURP) with the incidence of clinical malaria in differentepidemiological settings Previous results have been diffi-cult to interpret due to different study protocols and ana-lytical methods having been used [5-79142938-40] Inthis study the prevalence of asymptomatic malaria para-sitaemia was relatively high and stable while incidence ofclinical malaria fluctuated in parallel with the intensity oftransmission and seasonal rainfall pattern These patternsare typical of this area and have been reported in previous

The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malariaFigure 1The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malaria The point prevalence of asymptomatic parasitaemia for each month is represented as a bar graph and the pattern of clinical malaria (incident rate) is shown as a line graph

Page 5 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

studies [641-43] The variation in the incidence of clinicalmalaria during the study period may be due to the intro-duction of new parasites with different antigenic presenta-tion into the population leading to clinical malaria insusceptible individuals The risk of malaria decreased withage while isotype and IgG subclass levels to the four anti-gens generally increased with age This is consistent withthe hypothesis that immunity to malaria is largely effectedthrough antibody-mediated mechanisms and that protec-tive antibody levels to relevant antigens increase with age-related exposure to the parasites [44] Increasing IgG andIgM levels with age may reflect greater cumulative expo-sure of older children but may also be due to older chil-dren having a more mature immune system [45] Theassociation of IgM responses with reduced malaria inci-dence indicates a possible role in immunity in Ghanaianchildren Although much emphasis has been placed onIgG as the important isotype in immunity against malariaIgM which has lower affinity but is multivalent mayafford protection via other mechanisms such as the block-ing of merozoite invasion of erythrocytes complementactivation agglutination of merozoites [46]

The association with malaria incidence of IgG responsesto MSP3 MSP119 and GLURP is consistent with data fromseveral other immuno-epidemiological studies [57-91229303847-49] indicating that these antigens maybe targets of protective antibodies [124] There was noevidence that IgG levels to AMA1 were associated with

malaria incidence and there was no evidence of an inter-action with baseline parasitaemia in contrast with similarstudies conducted in Kenya [14] As shown in other stud-ies the cytophilic antibody levels to the four antigenstested in this study were higher than the non-cytophilicones emphasizing their importance in anti-malariaimmunity [929385051] IgG and cytophilic antibodylevels were highest for AMA1 while the levels were rela-tively low for MSP3 These differences in specific antibodylevels may be related to the number of immunogenic B-Cell epitopes exposed to the immune system and couldalso be related to the structure location and function ofthe particular antigen(s) With the exception of IgG2 lev-els to AMA1 IgG2 levels to GLURP MSP119 and MSP3increased with age which may suggest that IgG2 isinvolved in immunity against malaria In recent studiesmalaria antigen specific IgG2 have been shown to bindwith high affinity to mutant Fcγ RII H131-gtR receptors[52] on monocytes granulocytes and B cells thus afford-ing protection against malaria through monocytes and orneutrophil mediated mechanisms in subjects expressingthe mutant CD32 form [53] There was however no evi-dence found that IgG2 was associated with malaria inci-dence for any of the four antigens tested

The IgG subclasses IgG1 IgG2 and IgG3 for MSP3MSP119 GLURP and IgG4 to MSP3 were associated with areduced risk of malaria in un-adjusted analysis but ofthese only IgG1 to MSP119 was independently associated

Baseline IgG and IgM levels in relation to ageFigure 2Baseline IgG and IgM levels in relation to age The top and bottom panels represents total IgG or IgM levels to AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 6 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

treatment during the study is that children could havereceived anti-malarial treatments during the period theyare considered at risk in the analysis To explore theimpact of these drug treatments a time dependent varia-ble was defined to allow for a reduced risk of malaria fora period of 28 days after each drug treatment which wasincluded as a covariate in the Poisson regression modelLOESS smoothing was applied in R software to plot anti-body levels in relation to age Spearmans rank correlationtest was used to assess associations between antibody lev-els and age

ResultsPattern of P falciparum infections and malaria in the study cohortOf the 352 children recruited for the original study eightwere lost to follow-up immediately after the baselineblood sampling Of the 344 children followed up sixtyfour (19) had at least one episode of malaria (53 chil-dren had one episode nine had two and two had threeepisodes) The incidence rate of malaria in the studycohort was 035 attacks per child per year (Table 1) Therisk of clinical malaria decreased with increasing age [37]Sixty-six percent of the children had asymptomatic parasi-taemia at baseline Parasites were predominantly P falci-parum (95) and the prevalence of parasitaemiameasured each month was roughly constant (rangingfrom 50 to 65) The incidence of clinical malariahowever varied during the survey period rising graduallyfrom March to May peaking in July and then decreasinguntil November (Figure 1)

Relationship between age and antibody levelsSeveral studies in malaria endemic regions have shownincreasing antibody levels with age This pattern is morepronounced the greater the intensity and duration ofmalaria transmission In this study the relationshipbetween antibody levels and age were assessed for anti-bodies against MSP119 AMA1 GLURP and MSP3 Thelevels of IgG and IgM to MSP119 MSP3 AMA1 andGLURP increased with age (Spearman correlation coeffi-cient (rs) 021 ndash 045 p lt 0001 Figure 2) IgG levels toAMA1 however gradually increased until six years of age

and then leveled off For the four antigens tested IgG1IgG2 and IgG3 significantly increased with age (rs 012 ndash036 p lt 003) with the exception of IgG2 to AMA1 Therewas no evidence that the level of IgG4 was associated withage for any of the antigens Like IgG to AMA1 IgG3 toMSP3 steadily increased with age until seven years of agethen leveled off (Figure 3)

The levels of both IgG and IgG1 were highest for AMA1and lowest for MSP3 whereas IgG levels to MSP119 andGLURP were comparable (Figures 2 and 3) The levels ofIgG3 were higher in AMA1 than the comparable levels inGLURP MSP119 and MSP3 (Figure 3) In general the lev-els of cytophilic IgG1 and IgG3 were higher than those ofnon-cytophilic IgG2 and IgG4 IgG4 levels being the low-est

Antibody levels in relation to protection from clinical malariaTotal IgG to MSP3 MSP119 and GLURP and IgM to allfour antigens tested (MSP3 MSP119 GLURP and AMA1)were associated with reduced malaria incidence in crudeanalyses (Table 2) The incidence of both clinical malariaand antibody levels were associated with age age is there-fore a potential confounder and it is important to adjustfor its effects After adjusting for the effect of age there wasevidence of a significant association between total IgG toMSP3 MSP119 GLURP and reduced risk of malaria (IgGto MSP3 rate ratio 069 (95CI 053 090) P = 001 IgGto MSP119 075 (061 092) P = 001 IgG to GLURP 079(064 098) P = 004) and IgM levels to AMA1 MSP3MSP119 were also significantly associated with reducedrisk of malaria (Table 2)

A large proportion of the measurements of IgG4 to MSP3and AMA1 and IgG2 to AMA1 were zero (left-censored)values and so these variables were treated as categoricalvariables (Table 3) In the crude analysis IgG1 IgG2 andIgG3 to MSP3 MSP119 GLURP and also IgG4 to MSP3were associated with a reduced risk of clinical malaria(Table 3 4) but after adjustment for age only one ofthese variables IgG1 to MSP119 remained significantly

Table 1 Malaria incidence by age group

Age in yrs No of children

Cum incidence of malaria

no episodes of malaria (years at risk)

Incidence rate per child year (95CI)

3 32 1232 (38) 15 (1952) 079 (048ndash131)4ndash5 86 2486 (28) 28 (5475) 051 (035ndash074)6ndash7 79 1179 (14) 13 (5138) 025 (015ndash044)8 48 648 (13) 7 (3111) 022 (011ndash047)

9ndash10 99 1199 (11) 14 (6122) 023 (014ndash039)

Total 344 64344 (19) 77 (21748) 035 (028ndash044)

Page 4 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

associated with malaria incidence (rate ratio 089 (95CI080 099) P = 004 Table 4)

When all immunological variables were consideredsimultaneously only two variables were independentlyassociated with reduced malaria incidence IgG1 [(080(066ndash096) p = 0018)] to MSP119 and IgM [(048(032ndash072) p lt 0001)] to MSP119 (Table 5)

The estimated rate ratio for the effect of antimalarial treat-ments given to children with parasitaemia lt 5000uL was025 (95CI 006ndash105) indicating these children had asubstantially reduced risk of being found positive withparasitaemia gt= 5000ul during the 28 days following thetreatment But the estimates of rate ratios for other varia-bles in the model were unchanged when this variable wasincluded in the model suggesting treatment effects did not

cause a bias in the estimation of the effects of immunolog-ical variables

DiscussionThis study in Ghanaian children is one of a series of stud-ies designed to assess using standardized methods theassociation of antibody levels to four leading asexualblood-stage malaria antigens (MSP119 MSP3 AMA1 andGLURP) with the incidence of clinical malaria in differentepidemiological settings Previous results have been diffi-cult to interpret due to different study protocols and ana-lytical methods having been used [5-79142938-40] Inthis study the prevalence of asymptomatic malaria para-sitaemia was relatively high and stable while incidence ofclinical malaria fluctuated in parallel with the intensity oftransmission and seasonal rainfall pattern These patternsare typical of this area and have been reported in previous

The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malariaFigure 1The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malaria The point prevalence of asymptomatic parasitaemia for each month is represented as a bar graph and the pattern of clinical malaria (incident rate) is shown as a line graph

Page 5 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

studies [641-43] The variation in the incidence of clinicalmalaria during the study period may be due to the intro-duction of new parasites with different antigenic presenta-tion into the population leading to clinical malaria insusceptible individuals The risk of malaria decreased withage while isotype and IgG subclass levels to the four anti-gens generally increased with age This is consistent withthe hypothesis that immunity to malaria is largely effectedthrough antibody-mediated mechanisms and that protec-tive antibody levels to relevant antigens increase with age-related exposure to the parasites [44] Increasing IgG andIgM levels with age may reflect greater cumulative expo-sure of older children but may also be due to older chil-dren having a more mature immune system [45] Theassociation of IgM responses with reduced malaria inci-dence indicates a possible role in immunity in Ghanaianchildren Although much emphasis has been placed onIgG as the important isotype in immunity against malariaIgM which has lower affinity but is multivalent mayafford protection via other mechanisms such as the block-ing of merozoite invasion of erythrocytes complementactivation agglutination of merozoites [46]

The association with malaria incidence of IgG responsesto MSP3 MSP119 and GLURP is consistent with data fromseveral other immuno-epidemiological studies [57-91229303847-49] indicating that these antigens maybe targets of protective antibodies [124] There was noevidence that IgG levels to AMA1 were associated with

malaria incidence and there was no evidence of an inter-action with baseline parasitaemia in contrast with similarstudies conducted in Kenya [14] As shown in other stud-ies the cytophilic antibody levels to the four antigenstested in this study were higher than the non-cytophilicones emphasizing their importance in anti-malariaimmunity [929385051] IgG and cytophilic antibodylevels were highest for AMA1 while the levels were rela-tively low for MSP3 These differences in specific antibodylevels may be related to the number of immunogenic B-Cell epitopes exposed to the immune system and couldalso be related to the structure location and function ofthe particular antigen(s) With the exception of IgG2 lev-els to AMA1 IgG2 levels to GLURP MSP119 and MSP3increased with age which may suggest that IgG2 isinvolved in immunity against malaria In recent studiesmalaria antigen specific IgG2 have been shown to bindwith high affinity to mutant Fcγ RII H131-gtR receptors[52] on monocytes granulocytes and B cells thus afford-ing protection against malaria through monocytes and orneutrophil mediated mechanisms in subjects expressingthe mutant CD32 form [53] There was however no evi-dence found that IgG2 was associated with malaria inci-dence for any of the four antigens tested

The IgG subclasses IgG1 IgG2 and IgG3 for MSP3MSP119 GLURP and IgG4 to MSP3 were associated with areduced risk of malaria in un-adjusted analysis but ofthese only IgG1 to MSP119 was independently associated

Baseline IgG and IgM levels in relation to ageFigure 2Baseline IgG and IgM levels in relation to age The top and bottom panels represents total IgG or IgM levels to AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 6 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

associated with malaria incidence (rate ratio 089 (95CI080 099) P = 004 Table 4)

When all immunological variables were consideredsimultaneously only two variables were independentlyassociated with reduced malaria incidence IgG1 [(080(066ndash096) p = 0018)] to MSP119 and IgM [(048(032ndash072) p lt 0001)] to MSP119 (Table 5)

The estimated rate ratio for the effect of antimalarial treat-ments given to children with parasitaemia lt 5000uL was025 (95CI 006ndash105) indicating these children had asubstantially reduced risk of being found positive withparasitaemia gt= 5000ul during the 28 days following thetreatment But the estimates of rate ratios for other varia-bles in the model were unchanged when this variable wasincluded in the model suggesting treatment effects did not

cause a bias in the estimation of the effects of immunolog-ical variables

DiscussionThis study in Ghanaian children is one of a series of stud-ies designed to assess using standardized methods theassociation of antibody levels to four leading asexualblood-stage malaria antigens (MSP119 MSP3 AMA1 andGLURP) with the incidence of clinical malaria in differentepidemiological settings Previous results have been diffi-cult to interpret due to different study protocols and ana-lytical methods having been used [5-79142938-40] Inthis study the prevalence of asymptomatic malaria para-sitaemia was relatively high and stable while incidence ofclinical malaria fluctuated in parallel with the intensity oftransmission and seasonal rainfall pattern These patternsare typical of this area and have been reported in previous

The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malariaFigure 1The monthly point prevalence of asymptomatic malaria parasitaemia and incident rate of clinical malaria The point prevalence of asymptomatic parasitaemia for each month is represented as a bar graph and the pattern of clinical malaria (incident rate) is shown as a line graph

Page 5 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

studies [641-43] The variation in the incidence of clinicalmalaria during the study period may be due to the intro-duction of new parasites with different antigenic presenta-tion into the population leading to clinical malaria insusceptible individuals The risk of malaria decreased withage while isotype and IgG subclass levels to the four anti-gens generally increased with age This is consistent withthe hypothesis that immunity to malaria is largely effectedthrough antibody-mediated mechanisms and that protec-tive antibody levels to relevant antigens increase with age-related exposure to the parasites [44] Increasing IgG andIgM levels with age may reflect greater cumulative expo-sure of older children but may also be due to older chil-dren having a more mature immune system [45] Theassociation of IgM responses with reduced malaria inci-dence indicates a possible role in immunity in Ghanaianchildren Although much emphasis has been placed onIgG as the important isotype in immunity against malariaIgM which has lower affinity but is multivalent mayafford protection via other mechanisms such as the block-ing of merozoite invasion of erythrocytes complementactivation agglutination of merozoites [46]

The association with malaria incidence of IgG responsesto MSP3 MSP119 and GLURP is consistent with data fromseveral other immuno-epidemiological studies [57-91229303847-49] indicating that these antigens maybe targets of protective antibodies [124] There was noevidence that IgG levels to AMA1 were associated with

malaria incidence and there was no evidence of an inter-action with baseline parasitaemia in contrast with similarstudies conducted in Kenya [14] As shown in other stud-ies the cytophilic antibody levels to the four antigenstested in this study were higher than the non-cytophilicones emphasizing their importance in anti-malariaimmunity [929385051] IgG and cytophilic antibodylevels were highest for AMA1 while the levels were rela-tively low for MSP3 These differences in specific antibodylevels may be related to the number of immunogenic B-Cell epitopes exposed to the immune system and couldalso be related to the structure location and function ofthe particular antigen(s) With the exception of IgG2 lev-els to AMA1 IgG2 levels to GLURP MSP119 and MSP3increased with age which may suggest that IgG2 isinvolved in immunity against malaria In recent studiesmalaria antigen specific IgG2 have been shown to bindwith high affinity to mutant Fcγ RII H131-gtR receptors[52] on monocytes granulocytes and B cells thus afford-ing protection against malaria through monocytes and orneutrophil mediated mechanisms in subjects expressingthe mutant CD32 form [53] There was however no evi-dence found that IgG2 was associated with malaria inci-dence for any of the four antigens tested

The IgG subclasses IgG1 IgG2 and IgG3 for MSP3MSP119 GLURP and IgG4 to MSP3 were associated with areduced risk of malaria in un-adjusted analysis but ofthese only IgG1 to MSP119 was independently associated

Baseline IgG and IgM levels in relation to ageFigure 2Baseline IgG and IgM levels in relation to age The top and bottom panels represents total IgG or IgM levels to AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 6 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

studies [641-43] The variation in the incidence of clinicalmalaria during the study period may be due to the intro-duction of new parasites with different antigenic presenta-tion into the population leading to clinical malaria insusceptible individuals The risk of malaria decreased withage while isotype and IgG subclass levels to the four anti-gens generally increased with age This is consistent withthe hypothesis that immunity to malaria is largely effectedthrough antibody-mediated mechanisms and that protec-tive antibody levels to relevant antigens increase with age-related exposure to the parasites [44] Increasing IgG andIgM levels with age may reflect greater cumulative expo-sure of older children but may also be due to older chil-dren having a more mature immune system [45] Theassociation of IgM responses with reduced malaria inci-dence indicates a possible role in immunity in Ghanaianchildren Although much emphasis has been placed onIgG as the important isotype in immunity against malariaIgM which has lower affinity but is multivalent mayafford protection via other mechanisms such as the block-ing of merozoite invasion of erythrocytes complementactivation agglutination of merozoites [46]

The association with malaria incidence of IgG responsesto MSP3 MSP119 and GLURP is consistent with data fromseveral other immuno-epidemiological studies [57-91229303847-49] indicating that these antigens maybe targets of protective antibodies [124] There was noevidence that IgG levels to AMA1 were associated with

malaria incidence and there was no evidence of an inter-action with baseline parasitaemia in contrast with similarstudies conducted in Kenya [14] As shown in other stud-ies the cytophilic antibody levels to the four antigenstested in this study were higher than the non-cytophilicones emphasizing their importance in anti-malariaimmunity [929385051] IgG and cytophilic antibodylevels were highest for AMA1 while the levels were rela-tively low for MSP3 These differences in specific antibodylevels may be related to the number of immunogenic B-Cell epitopes exposed to the immune system and couldalso be related to the structure location and function ofthe particular antigen(s) With the exception of IgG2 lev-els to AMA1 IgG2 levels to GLURP MSP119 and MSP3increased with age which may suggest that IgG2 isinvolved in immunity against malaria In recent studiesmalaria antigen specific IgG2 have been shown to bindwith high affinity to mutant Fcγ RII H131-gtR receptors[52] on monocytes granulocytes and B cells thus afford-ing protection against malaria through monocytes and orneutrophil mediated mechanisms in subjects expressingthe mutant CD32 form [53] There was however no evi-dence found that IgG2 was associated with malaria inci-dence for any of the four antigens tested

The IgG subclasses IgG1 IgG2 and IgG3 for MSP3MSP119 GLURP and IgG4 to MSP3 were associated with areduced risk of malaria in un-adjusted analysis but ofthese only IgG1 to MSP119 was independently associated

Baseline IgG and IgM levels in relation to ageFigure 2Baseline IgG and IgM levels in relation to age The top and bottom panels represents total IgG or IgM levels to AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 6 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

with malaria incidence after adjustment for age Otherstudies have shown the importance of IgG1 in clearingparasitaemia in children [54754] In a previous cohortstudy conducted in the same area there was no associa-tion between antibody levels to MSP119 and malaria inci-dence [6] This may be due to differences in antigen andantibody reagents used in the two studies the MSP119used in this study was a Baculovirus product that includeda synthetic G-C enriched PfMSP1 gene that coded for the43 N-terminal MSP1 precursor residues and 16 amino

acid residues upstream of the classical MSP-119 (NIS---FCS) [33] compared to the one produced in E coli whichhad been used in the previous study Although antibodiesto MSP119 have been shown to be associated with bothexposure and protection from disease the fine specifici-ties of such responses may contribute to protection [40]The antigen used in this study may have assessed antibod-ies of fine specificities that are protective [405556]whereas the antigen used in the previous study did not Itmay therefore be important to assess in a standardized

Baseline IgG subclass levels in relation to ageFigure 3Baseline IgG subclass levels in relation to age The top 2nd 3rd and 4th panels represent IgG subclasses 1 to 4 against AMA1 GLURP MSP119 and MSP3 in relation to age of Ghanaian children respectively The line shows the LOESS smoothed estimate of the geometric mean

Page 7 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Page 8 of 11(page number not for citation purposes)

Table 2 Association of total IgG and IgM with malaria incidence

IgG Crude IRR (95CI) P-value IRR adjusted for age (95CI) P-value

MSP3 059 (045 076) lt 00001 069 (053 090) 001MSP119 074 (063 088) lt 00001 075 (061 092) 001GLURP 068 (055 082) lt 00001 079 (064 098) 004AMA1 096 (084 109) 050 103 (089 120) 067IgM

MSP3 062 (049 080) lt 00001 076 (059 097) 003MSP1 059 (047 075) lt 00001 068 (053 088) lt 001GLURP 071 (058 087) lt 001 084 (068 103) 009AMA1 048 (034 067) lt 00001 063 (044 091) 001

Malaria incidence rate ratios (IRR) with 95 confidence interval corresponding to a doubling of baseline antibody level before and after adjustment for the effect of age

Table 3 Association of IgG2 and IgG4 with malaria incidence

Antigen Antibody Crude IRR (95CI) P-value Age-adjusted IRR P-value

MSP3

IgG2 (coded as log(2)) 074 (057 098) 004 085 (063 115) 029IgG4 lt detection limit 1 001 1 020

lt median 045 (021 095) 056 (026 119)ge median 047 (028 079) 068 (040 115)

MSP119IgG2 (coded as log(2)) 062 (045 086) lt 001 0836 (0606 1157) 0275IgG4 (coded as log(2)) 107 (095 121) 025 0838 (0608 1159) 0278

GLURPIgG2 (coded as log(2)) 065 (049 086) lt 001 079 (058 108) 013IgG4 (coded as log(2)) 094 (081 108) 037 101 (087 116) 094

AMA1IgG2 lt detection limit 1 018 1 054

lt median 180 (086 378)ge median 178 (090 351)

IgG4 lt detection limit 1 019 1 054lt median 096 (047 193)ge median 062 (036 107)

For IgG 2 to AMA1 and for IgG4 to AMA1 and MSP3 antibody level was categorizedFor IgG2 to MSP3 MSP119 and GLURP the association is expressed as the incidence rate ratio (IRR) corresponding to a doubling of baseline antibody level

Table 4 Association of IgG1 and IgG3 with malaria incidence

IgG1 Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value

MSP3 082 (072 094) lt 001 088 (076 103) 011MSP119 086 (077 097) 001 089 (080 099) 004GLURP 085 (074 098) 002 093 (081 108) 035AMA1 095 (086 106) 036 101 (090 114) 083

IgG3

MSP3 090 (082 098) 001 094 (086 103) 017MSP119 090 (082 099) 003 093 (084 103) 018GLURP 083 (074 093) lt 001 091 (080103) 014AMA1 097 (089 106) 048 098 (090 107) 073

The incidence rate ratio (IRR) corresponds to a doubling of baseline antibody level before and after adjustment for effects of age

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

way the various MSP119 and other antigens that are pro-duced in different expression systems in order to selectthe most appropriate antigenexpression system formalaria vaccine development In another study in thisseries in Burkina Faso the same antigens reagents ELISAprocedure and analytical methods were used none of theisotypes and subclasses to MSP119 was associated with theincidence of clinical malaria Since the same laboratorymethods were used the different outcomes of these twostudies may be attributed to differences in malaria trans-mission or to the age of the children [44] In Burkina Fasothe malaria transmission season is much shorter whichmay influence the induction of differing antibody typesfor controlling malaria as shown in recent studies con-ducted in areas with different malaria endemicities in Tan-zania [944] Although total IgG to GLURP and MSP3were associated with the risk of malaria none of the con-stituent subclasses was identified to be associated withprotection When the effects of all the immunological var-iables were considered simultaneously only IgG1 andIgM to MSP119 were independently associated with theincidence of clinical malaria which may indicate theimportance of MSP119 in malaria vaccine developmentParasite growth inhibition assays would be required toconfirm if this association reflects a functional role ofMSP119 in immunity

ConclusionIn conclusion using standardized AIA ELISA anti-MSP19antibodies (IgG1 and IgM) have been shown to be themost strongly correlated with reduced risk of clinicalmalaria among the four malaria vaccine candidates testedThe standardized AIA ELISA developed for this projectcould be used to validate malaria vaccine candidate anti-gens provide useful baseline information for clinical tri-als and contribute to quality assured laboratory capacityin Africa

Authors contributionsDD carried out field studies developed assays and draftedthe manuscript AA performed the ELISA compiled dataand assisted in the manuscript writing KAK and HLassisted with the ELISA MT assisted with the assay devel-opment and manuscript writing while PM SB and ERwrote the analysis plan performed the data analysis and

together with RC BDA and YDO contributed to the writ-ing of the manuscript All authors read and approved thefinal manuscript

AcknowledgementsThe authors are grateful to the children in Dodowa community for partic-ipation in this study Dr Soren Jepsen is acknowledged for supporting and encouraging the Network Dr Ramadhani Abdalla Noor is also acknowl-edged for the support and management contribution

This investigation received support of AMANET and Netherlands Ministry of Foreign Affairs (DGIS) The Wellcome Trust (grant 057736) is also acknowledged for funding the prospective study that generated plasma samples and morbidity survey data

References1 Cohen S McGREGOR IA CARRINGTON S Gamma-globulin

and acquired immunity to human malaria Nature 1961192733-737 [http]

2 Cohen S Butcher GA Serum antibody in acquired malarialimmunity Trans R Soc Trop Med Hyg 1971 65125-135

3 Good MF Towards the development of the ideal malaria vac-cine A decade of progress in a difficult field Med J Aust 1991154284-289

4 Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tay-oun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe PParasitologic and clinical human response to immunoglobu-lin administration in falciparum malaria Am J Trop Med Hyg1991 45297-308

5 Braga EM Barros RM Reis TA Fontes CJ Morais CG Martins MSKrettli AU Association of the IgG response to Plasmodium fal-ciparum merozoite protein (C-terminal 19 kD) with clinicalimmunity to malaria in the Brazilian Amazon region Am JTrop Med Hyg 2002 66461-466

6 Dodoo D Theander TG Kurtzhals JA Koram K Riley E AkanmoriBD Nkrumah FK Hviid L Levels of antibody to conserved partsof Plasmodium falciparum merozoite surface protein 1 inGhanaian children are not associated with protection fromclinical malaria Infect Immun 1999 672131-2137

7 Egan AF Morris J Barnish G Allen S Greenwood BM Kaslow DCHolder AA Riley EM Clinical immunity to Plasmodium falci-parum malaria is associated with serum antibodies to the 19-kDa C-terminal fragment of the merozoite surface antigenPfMSP-1 J Infect Dis 1996 173765-769

8 Soe S Theisen M Roussilhon C Aye KS Druilhe P Associationbetween protection against clinical malaria and antibodiesto merozoite surface antigens in an area of hyperendemicityin Myanmar complementarity between responses to mero-zoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72247-252

9 Lusingu JP Vestergaard LS Alifrangis M Mmbando BP Theisen MKitua AY Lemnge MM Theander TG Cytophilic antibodies toPlasmodium falciparum glutamate rich protein are associatedwith malaria protection in an area of holoendemic transmis-sion Malar J 2005 448

10 Nebie I Diarra A Ouedraogo A Soulama I Bougouma EC Tiono ABKonate AT Chilengi R Theisen M Dodoo D Remarque E Bosomp-rah S Milligan P Sirima SB Humoral responses to Plasmodium

Table 5 Adjusted rate ratios for immunological variables independently associated with malaria risk in the final model

Crude IRR (95CI)

P-value IRR adjusted for age (95CI)

P-value IRR adjusted for effects of age and treatment (95CI)

IgM to MSP119 044 (030ndash064) P lt 0001 049 (033ndash073) P lt 0001 048 (032ndash072)IgG1 to MSP119 078 (065ndash093) P = 0005 080 (066ndash096) P = 0018 080 (067ndash097)Treatment effect 028 (007ndash115) P = 0076 - - 025 (006ndash105)

Page 9 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

falciparum blood-stage antigens and association with inci-dence of clinical malaria in children living in an area of sea-sonal malaria transmission in Burkina Faso West AfricaInfect Immun 2008 76759-766

11 Sirima SB Nebie I Ouedraogo A Tiono AB Konate AT Gansane ADerme AI Diarra A Ouedraogo A Soulama I Cuzzin-Ouattara NCousens S Leroy O Safety and immunogenicity of the Plasmo-dium falciparum merozoite surface protein-3 long syntheticpeptide (MSP3-LSP) malaria vaccine in healthy semi-immune adult males in Burkina Faso West Africa Vaccine2007 252723-2732

12 Meraldi V Nebie I Tiono AB Diallo D Sanogo E Theisen M DruilheP Corradin G Moret R Sirima BS Natural antibody response toPlasmodium falciparum Exp-1 MSP-3 and GLURP long syn-thetic peptides and association with protection Parasite Immu-nol 2004 26265-272

13 Hui G Hashimoto C Plasmodium falciparum anti-MSP1-19 anti-bodies induced by MSP1-42 and MSP1-19 based vaccines dif-fered in specificity and parasite growth inhibition in terms ofrecognition of conserved versus variant epitopes Vaccine2007 25948-956

14 Polley SD Mwangi T Kocken CH Thomas AW Dutta S Lanar DERemarque E Ross A Williams TN Mwambingu G Lowe B ConwayDJ Marsh K Human antibodies to recombinant protein con-structs of Plasmodium falciparum Apical Membrane Antigen1 (AMA1) and their associations with protection frommalaria Vaccine 2004 23718-728

15 Bouharoun-Tayoun H Oeuvray C Lunel F Druilhe P Mechanismsunderlying the monocyte-mediated antibody-dependent kill-ing of Plasmodium falciparum asexual blood stages J Exp Med1995 182409-418

16 Hermsen CC Verhage DF Telgt DS Teelen K Bousema JT Roesten-berg M Bolad A Berzins K Corradin G Leroy O Theisen M Sauer-wein RW Glutamate-rich protein (GLURP) inducesantibodies that inhibit in vitro growth of Plasmodium falci-parum in a phase 1 malaria vaccine trial Vaccine 2007252930-2940

17 Miura K Zhou H Muratova OV Orcutt AC Giersing B Miller LHLong CA In immunization with Plasmodium falciparum apicalmembrane antigen 1 the specificity of antibodies dependson the species immunized Infect Immun 2007 755827-5836

18 Theisen M Soe S Oeuvray C Thomas AW Vuust J Danielsen SJepsen S Druilhe P The glutamate-rich protein (GLURP) ofPlasmodium falciparum is a target for antibody-dependentmonocyte-mediated inhibition of parasite growth in vitroInfect Immun 1998 6611-17

19 Blackman MJ Holder AA Secondary processing of the Plasmo-dium falciparum merozoite surface protein-1 (MSP1) by a cal-cium-dependent membrane-bound serine proteaseshedding of MSP133 as a noncovalently associated complexwith other fragments of the MSP1 Mol Biochem Parasitol 199250307-315

20 Borre MB Dziegiel M Hoslashgh B Petersen E Rieneck K Riley EMeis JF Aikawa M Nakamura K Harada M etc Primary struc-ture and localization of a conserved immunogenic Plasmo-dium falciparum glutamate rich protein (GLURP) expressedin both the preerythrocytic and erythrocytic stages of thevertebrate lifecycle Mol Biochem Parasitol 1991 49119-132

21 Holder AA Blackman MJ Burghaus PA Chappel JA Ling IT McCal-lum-Deighton N Shai S A malaria merozoite surface protein(MSP1)-structure processing and function Mem Inst OswaldoCruz 1992 87 Suppl 337-42

22 Lunel F Druilhe P Effector cells involved in nonspecific andantibody-dependent mechanisms directed against Plasmo-dium falciparum blood stages in vitro Infect Immun 1989572043-2049

23 Mitchell GH Thomas AW Margos G Dluzewski AR Bannister LHApical membrane antigen 1 a major malaria vaccine candi-date mediates the close attachment of invasive merozoitesto host red blood cells Infect Immun 2004 72154-158

24 Narum DL Thomas AW Differential localization of full-lengthand processed forms of PF83AMA-1 an apical membraneantigen of Plasmodium falciparum merozoites Mol Biochem Par-asitol 1994 6759-68

25 ODonnell RA Saul A Cowman AF Crabb BS Functional conser-vation of the malaria vaccine antigen MSP-119across dis-tantly related Plasmodium species Nat Med 2000 691-95

26 Oeuvray C Bouharoun-Tayoun H Gras-Masse H Bottius E KaidohT Aikawa M Filgueira MC Tartar A Druilhe P Merozoite surfaceprotein-3 a malaria protein inducing antibodies that pro-mote Plasmodium falciparum killing by cooperation withblood monocytes Blood 1994 841594-1602

27 Ahorlu CK Dunyo SK Afari EA Koram KA Nkrumah FK Malaria-related beliefs and behaviour in southern Ghana implica-tions for treatment prevention and control Trop Med IntHealth 1997 2488-499

28 Cavanagh DR Dodoo D Hviid L Kurtzhals JA Theander TG Akan-mori BD Polley S Conway DJ Koram K McBride JS Antibodies tothe N-terminal block 2 of Plasmodium falciparum merozoitesurface protein 1 are associated with protection against clin-ical malaria Infect Immun 2004 726492-6502

29 Dodoo D Theisen M Kurtzhals JA Akanmori BD Koram KA JepsenS Nkrumah FK Theander TG Hviid L Naturally acquired anti-bodies to the glutamate-rich protein are associated withprotection against Plasmodium falciparum malaria J Infect Dis2000 1811202-1205

30 Theisen M Dodoo D Toure-Balde A Soe S Corradin G Koram KKKurtzhals JA Hviid L Theander T Akanmori B Ndiaye M Druilhe PSelection of glutamate-rich protein long synthetic peptidesfor vaccine development antigenicity and relationship withclinical protection and immunogenicity Infect Immun 2001695223-5229

31 Kocken CH Withers-Martinez C Dubbeld MA van der WA HackettF Valderrama A Blackman MJ Thomas AW High-level expressionof the malaria blood-stage vaccine candidate Plasmodium fal-ciparum apical membrane antigen 1 and induction of anti-bodies that inhibit erythrocyte invasion Infect Immun 2002704471-4476

32 Theisen M Vuust J Gottschau A Jepsen S Hogh B Antigenicityand immunogenicity of recombinant glutamate-rich proteinof Plasmodium falciparum expressed in Escherichia coli ClinDiagn Lab Immunol 1995 230-34

33 Bonnet S Petres S Holm I Fontaine T Rosario S Roth C LongacreS Soluble and glyco-lipid modified baculovirus Plasmodiumfalciparum C-terminal merozoite surface protein 1 twoforms of a leading malaria vaccine candidate Vaccine 2006245997-6008

34 Audran R Cachat M Lurati F Soe S Leroy O Corradin G Druilhe PSpertini F Phase I malaria vaccine trial with a long syntheticpeptide derived from the merozoite surface protein 3 anti-gen Infect Immun 2005 738017-8026

35 Schellenberg JR Smith T Alonso PL Hayes RJ What is clinicalmalaria Finding case definitions for field research in highlyendemic areas Parasitol Today 1994 10439-442

36 Smith T Schellenberg JA Hayes R Attributable fraction esti-mates and case definitions for malaria in endemic areas StatMed 1994 132345-2358

37 Cavanagh DR Elhassan IM Roper C Robinson VJ Giha H HolderAA Hviid L Theander TG Arnot DE McBride JS A longitudinalstudy of type-specific antibody responses to Plasmodium fal-ciparum merozoite surface protein-1 in an area of unstablemalaria in Sudan J Immunol 1998 161347-359

38 Oeuvray C Theisen M Rogier C Trape JF Jepsen S Druilhe PCytophilic immunoglobulin responses to Plasmodium falci-parum glutamate-rich protein are correlated with protectionagainst clinical malaria in Dielmo Senegal Infect Immun 2000682617-2620

39 Okech B Mujuzi G Ogwal A Shirai H Horii T Egwang TG High tit-ers of IgG antibodies against Plasmodium falciparum serinerepeat antigen 5 (SERA5) are associated with protectionagainst severe malaria in Ugandan children Am J Trop Med Hyg2006 74191-197

40 Okech BA Corran PH Todd J Joynson-Hicks A Uthaipibull CEgwang TG Holder AA Riley EM Fine specificity of serum anti-bodies to Plasmodium falciparum merozoite surface proteinPfMSP-1(19) predicts protection from malaria infection andhigh-density parasitemia Infect Immun 2004 721557-1567

41 Afari EA Nakano T Binka F Owusu-Agyei S Asigbee J Seasonalcharacteristics of malaria infection in under-five children of

Page 10 of 11(page number not for citation purposes)

Malaria Journal 2008 7142 httpwwwmalariajournalcomcontent71142

Publish with BioMed Central and every scientist can read your work free of charge

BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime

Sir Paul Nurse Cancer Research UK

Your research papers will be

available free of charge to the entire biomedical community

peer reviewed and published immediately upon acceptance

cited in PubMed and archived on PubMed Central

yours mdash you keep the copyright

Submit your manuscript herehttpwwwbiomedcentralcominfopublishing_advasp

BioMedcentral

a rural community in southern Ghana West Afr J Med 19931239-42

42 Afari EA Appawu M Dunyo S Baffoe-Wilmot A Nkrumah FKMalaria infection morbidity and transmission in two ecolog-ical zones Southern Ghana Afr J Health Sci 1995 2312-315

43 Dodoo D Staalsoe T Giha H Kurtzhals JA Akanmori BD Koram KDunyo S Nkrumah FK Hviid L Theander TG Antibodies to vari-ant antigens on the surfaces of infected erythrocytes areassociated with protection from malaria in Ghanaian chil-dren Infect Immun 2001 693713-3718

44 Tongren JE Drakeley CJ McDonald SL Reyburn HG Manjurano ANkya WM Lemnge MM Gowda CD Todd JE Corran PH Riley EMTarget antigen age and duration of antigen exposure inde-pendently regulate immunoglobulin G subclass switching inmalaria Infect Immun 2006 74257-264

45 Baird JK Host age as a determinant of naturally acquiredimmunity to Plasmodium falciparum Parasitol Today 199511105-111

46 Boudin C Chumpitazi B Dziegiel M Peyron F Picot S Hogh B Ambr-oise-Thomas P Possible role of specific immunoglobulin Mantibodies to Plasmodium falciparum antigens in immunopro-tection of humans living in a hyperendemic area BurkinaFaso J Clin Microbiol 1993 31636-641

47 Branch OH Udhayakumar V Hightower AW Oloo AJ Hawley WANahlen BL Bloland PB Kaslow DC Lal AA A longitudinal investi-gation of IgG and IgM antibody responses to the merozoitesurface protein-1 19-kiloDalton domain of Plasmodium falci-parum in pregnant women and infants associations withfebrile illness parasitemia and anemia Am J Trop Med Hyg1998 58211-219

48 Branch OH Oloo AJ Nahlen BL Kaslow D Lal AA Anti-mero-zoite surface protein-1 19-kDa IgG in mother-infant pairsnaturally exposed to Plasmodium falciparum subclass analysiswith age exposure to asexual parasitemia and protectionagainst malaria V The Asembo Bay Cohort Project J InfectDis 2000 1811746-1752

49 Morais CG Soares IS Carvalho LH Fontes CJ Krettli AU Braga EMIgG isotype to C-terminal 19 kDa of Plasmodium vivax mero-zoite surface protein 1 among subjects with different levelsof exposure to malaria in Brazil Parasitol Res 2005 95420-426

50 Bouharoun-Tayoun H Druilhe P Plasmodium falciparum malariaevidence for an isotype imbalance which may be responsiblefor delayed acquisition of protective immunity Infect Immun1992 601473-1481

51 Bouharoun-Tayoun H Druilhe P Antibodies in falciparummalaria what matters most quantity or quality Mem InstOswaldo Cruz 1992 87 Suppl 3229-234

52 Nasr A Iriemenam NC Troye-Blomberg M Giha HA Balogun HAOsman OF Montgomery SM ElGhazali G Berzins K Fc gammareceptor IIa (CD32) polymorphism and antibody responsesto asexual blood-stage antigens of Plasmodium falciparummalaria in Sudanese patients Scand J Immunol 2007 6687-96

53 Braga EM Scopel KKG Komatsu NT daSilva-Nunes M Ferreira MUPolymorphism of the Fc gamma receptor IIA and malariamorbidity Journal of Molecular and Genetic Medicine 2005 15-10

54 Aubouy A Migot-Nabias F Deloron P Correlations betweentreatment outcome and both anti-MSP119 antibodyresponse and erythrocyte-related genetic factors in Plasmo-dium falciparum malaria Infect Genet Evol 2007 7147-154

55 Corran PH ODonnell RA Todd J Uthaipibull C Holder AA CrabbBS Riley EM The fine specificity but not the invasion inhibi-tory activity of 19-kilodalton merozoite surface protein 1-specific antibodies is associated with resistance to malarialparasitemia in a cross-sectional survey in The Gambia InfectImmun 2004 726185-6189

56 Uthaipibull C Aufiero B Syed SE Hansen B Guevara Patino JAAngov E Ling IT Fegeding K Morgan WD Ockenhouse C Birdsall BFeeney J Lyon JA Holder AA Inhibitory and blocking mono-clonal antibody epitopes on merozoite surface protein 1 ofthe malaria parasite Plasmodium falciparum J Mol Biol 20013071381-1394

Page 11 of 11(page number not for citation purposes)