Pf

HJa

b

c

d

e

f

a

ARR2AA

KPAIIM

1

blIarbeijet

c

0h

Acta Tropica 130 (2014) 80–87

Contents lists available at ScienceDirect

Acta Tropica

journa l homepage: www.e lsev ier .com/ locate /ac ta t ropica

attern of antibodies to the Duffy binding like domain of Plasmodiumalciparum antigen Pf332 in Senegalese individuals

alima A. Baloguna,∗, Nancy Awaha, Sandra Nilssonb, Christophe Rogierc,ean-Francois Traped, Qijun Chenb,f, Christian Roussilhone, Klavs Berzinsa

Department of Immunology, Wenner-Gren Institute, Stockholm University, SE-106 91 Stockholm, SwedenDepartment of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, SwedenInstitut Pasteur de Madagascar, Antananarivo, MadagascarInstitute de Recherches pour le Développement, Montpellier, FranceImmunology Unit, Pasteur Institut de Dakar, Senegal and Unité de Génétique fonctionnelle des maladies infectieuses, Institut Pasteur, Paris, FranceLaboratory of Parasitology, Institute of Pathogen Biology, Chinese Academy of Medical Sciences, Beijing 100730, China

r t i c l e i n f o

rticle history:eceived 5 December 2012eceived in revised form9 September 2013ccepted 25 October 2013vailable online 4 November 2013

eywords:f332-DBLntibody response

gG3 subclass

a b s t r a c t

Acquisition of antibodies against blood stage antigens is crucial in malaria immunity and the Plasmodiumfalciparum antigen Pf332, which is present in close association with the infected red blood cell membrane,is one such antigen. In this study, the antibody response to a Duffy binding like fragment of Pf332, termedPf332-DBL was investigated in sera from naturally exposed individuals living in Dielmo village, Senegal,with regard to immunoglobulin classes (IgG, IgM, IgE) and IgG subclasses (IgG1–4). While the levelsof IgM, IgG, IgG1 and IgG2 only displayed a moderate trend to increase with age, Pf332-DBL specificIgG3 levels increased significantly in the older villagers. In multivariate analysis, when controlling forconfounding factors, and in a linear model with a Poisson distribution, anti-Pf332-DBL IgG3 as well as theratio of cytophilic to non cytophilic anti-Pf332-DBL antibodies were found significantly associated witha reduced risk of malaria attack. This association was also present when the IgG3:IgG1 ratio was tested.

mmunityalaria attack

Finally, two subgroups of villagers with the same mean age, were delineated by IgG3 concentrations eitherlower or higher than the median value. A total of 45.2% of the individuals with low anti-Pf332-DBL-IgG3levels but only 21.4% of the villagers in the group with high levels of such antibodies had a clinical malariaattack during a period of 3 years of continuous follow-up after the blood sampling. In conclusion, Pf332-DBL induces naturally the acquisition of antibodies, and Pf332-DBL-specific IgG3 appears to be associated

alari

with protection against m

. Introduction

The impact of Plasmodium falciparum induced mortality is beingorne mainly by children residing in Sub-Saharan Africa, although

arge populations of the world are still at risk (Hay and Snow, 2006).n endemic areas, older children and adults develop a naturallycquired immunity to severe and life-threatening malaria, but stillemain susceptible to infection (Hayes et al., 1992). The asexuallood stage of the malaria parasite is responsible for the pathogen-sis and clinical manifestations encountered during P. falciparumnfection. Antigens present during this stage are solely, or in con-

unction with other factors, the main culprits involved in the wholevents occurring during the infection, resulting in clinical outcomeshat, during the initial encounter might be inimical to the host, or

∗ Corresponding author. Tel.: +46 736394040.E-mail addresses: balogunh@gmail.com, halima@imun.su.se (H.A. Balogun),

hristian.roussilhon@pasteur.fr (C. Roussilhon).

001-706X/$ – see front matter © 2013 Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.actatropica.2013.10.018

a in this endemic setting.© 2013 Published by Elsevier B.V.

might turn out to be favorable/protective for the host after severalencounters with the parasite.

It has been demonstrated that antibodies are important medi-ators of naturally acquired immunity, playing a major role inreducing the parasite load (Cohen et al., 1961; Sabchareon et al.,1991). During P. falciparum infection, antibodies generated afterseveral encounters with the parasite are of different immunoglob-ulin classes and subclasses, which have distinct biological functionsleading to clinical protection or pathogenesis. P. falciparum specificantibodies play a pivotal role in protection against malaria, andthis protection is mainly reflected by antibody responses directedto blood stage antigens of the parasite (Courtin et al., 2009).

P. falciparum blood-stage infections, which account for themajority of the morbidity and mortality, generate a protectiveresponse in most individuals exposed repeatedly to the para-

site (Garraud et al., 2003), and parasite derived surface antigenson infected erythrocytes are involved in such naturally acquiredimmunity (Bull and Marsh, 2002). Studies on protective immu-nity to malaria have involved monitoring subjects in endemic

ta Trop

cda

tpSbi2aaSatmuietiB

thf2doBtmlcfbn(tvait

bsdhBepiwccia

2

2

mt

of P. falciparum infection. Serum samples from Liberia were usedas positive control. The concentrations of anti-malarial antibodies

H.A. Balogun et al. / Ac

ommunities for variable durations of time, measuring the inci-ence of infection or clinical disease and their association withntibody levels to selected blood stage antigens.

In the present study, antibodies reactive with one of such pro-ein, antigen Pf332, were analyzed. Pf332 is a megadalton protein,resent during the late stage of asexual development (Mattei andcherf, 1992). The protein appears in Maurer’s cleft in infected redlood cell (iRBC) cytoplasm, and later becomes associated with the

RBC membrane skeleton (Hinterberg et al., 1994; Waller et al.,010; Nilsson et al., 2012). The functional role of Pf332 in the par-site’s life cycle has not been fully elucidated, however it has beenssociated with different regions of the protein due to its large size.tudies have shown that it is involved in trafficking the PfEMP1ntigen (Haeggström et al., 2004; Hodder et al., 2009), investiga-ions with the N-terminal region of Pf332 revealed that the protein

ight be involved in re-invasion of red blood cells by facilitatingninfected red cell adhesion (Moll et al., 2007) as well as in the mod-

fication of host cell cytoskeleton (Glenister et al., 2009; Nilssont al., 2012). Another region, which is C-terminally located bindso actin (Waller et al., 2010), and is also suggested to be involvedn modulating the iRBC membrane skeleton (Waller et al., 2010;alogun et al., 2011).

This study focused on the N-terminal region of the pro-ein, which harbors a conserved Duffy binding-like (DBL) domainomologous to the DBL-domains of erythrocyte binding-like (EBL)

amily of invasion related proteins (Moll et al., 2007; Du et al.,010). Antibodies to other regions of Pf332, have previously beenemonstrated to be active in parasite neutralization, as well aspsonization in vitro (Ahlborg et al., 1996; Bolad and Berzins, 2000;alogun et al., 2009). Previous studies in the same study popula-ion analyzed antibodies to different regions of Pf332 in relation to

alaria. A study with the repeat fragment EB200 showed that highevels of IgG antibodies were predictive of fewer attacks of clini-al malaria (Ahlborg et al., 2002). In a study with the C-terminalragment Pf332-C231 (Balogun et al., 2009), the levels of anti-odies of all isotypes tended to be higher in individuals that didot experience any malaria attack during the follow-up periodIsraelsson et al., 2008). Recently, a study reported reactivity ofhe N-terminal region of Pf332 (Pf332-DBL) with plasma from indi-iduals living in malaria endemic regions of Uganda, Burkina Fasond Mali (Nilsson et al., 2011).These findings indicate that antibod-es to antigen Pf332 are induced after exposure to intense malariaransmission.

There are several reports on differences in anti-malarial anti-ody responses, both regarding immunoglobulin classes and IgGubclasses in relation to protection, which are indicated to beependent on various factors, such as the antigenic properties,ost age, cumulative exposure and genetics (Aucan et al., 2000;ouharoun-Tayoun and Druilhe, 1992; Ntoumi et al., 2005; Scopelt al., 2006; Tongren et al., 2005, 2006; Stanisic et al., 2009). In theresent epidemiological study, using the sera of malaria exposed

ndividuals from Dielmo, Senegal, the pattern of antibodies reactiveith recombinant Pf332-DBL, was analyzed. In addition, clini-

al data obtained between 6 and 36 months after blood sampleollection were used and further investigated if the immunolog-cal responses may possibly predict protection against malariattack.

. Materials and methods

.1. Study area, individuals tested and samples

The village of Dielmo is located in Senegal, West Africa, wherealaria is holoendemic, experiencing a high level of perennial

ransmission (Rogier et al., 1996, 1999). Sera were collected in

ica 130 (2014) 80–87 81

March 1994 through venous blood from 100 asymptomatic individ-uals who had detectible parasitaemia but no presentation of clinicalmalaria at the time of blood sampling (Rogier et al., 1996; Ahlborget al., 2002).

All individuals, with mean age 25.8 (4–87) years, were activelymonitored on a daily basis by a medical team for clinical attacksof malaria for a period of 36 months after sampling. The numberof infected bites per person per year during the study period was167 during the first year, 246 during the second and 234 duringthe third year of follow-up (Rogier et al., 1996, 1999). A malariaattack was defined in Dielmo as an episode of fever (with tem-perature >38.5 ◦C) associated with a parasite density exceeding anage-dependent pyrogenic threshold described and validated in thisvillage (Rogier et al., 1996). The parasite density threshold for eachage group was determined to be 20,000 at five years; 15,500 at 10years; 10,000 at 20 years; 7500 at 30 years of age and 5000 in adultsof more than 40 years of age (Rogier et al., 1996).

Sera from 10 Swedish donors not exposed to malaria served ascontrols. All samples were collected with informed consent and thestudy was approved by the Ministry of Health in Senegal (Rogieret al., 1996) and the National Ethics Committee in Sweden (Ahlborget al., 2002).

2.2. Antigen preparation

A His-tagged recombinant Pf332-DBL protein was generatedby cloning the gene fragment encoding N-terminal DBL-domainof Pf332 (amino acid residues 1–216) into a plasmid vectorpQE60 (Qiagen, Düsseldorf, Germany), expressed in Escherichia colibacteria strain SG13009 (Qiagen), and further purified using HisGraviTrapTM column (GE Healthcare, Uppsala, Sweden) (Moll et al.,2007; Nilsson et al., 2012).

2.3. Enzyme-linked immunosorbent assay (ELISA)

EIA/RIA plates (Costar, Corning, NY) were coated overnight at4 ◦C with 10 �g/ml of recombinant Pf332-DBL in sodium carbonatebuffer (pH 9.6). After blocking at 37 ◦C with carbonate buffer con-taining 1% BSA (w/v) for 2 h, the plates were washed with salinecontaining 0.05% Tween 20.

Plasma dilutions (1:1000 for IgG and IgM, 1:100 for IgE, 1:400for IgG1 and IgG3 and 1:20 for IgG2 and IgG4) were added towells in duplicates and incubated 1 h at 37 ◦C. IgG and IgM anti-bodies were detected using alkaline phosphatase conjugated toFc fragment specific goat anti-human IgG (Mabtech AB, Nacka,Sweden) and goat anti-human IgM (Jackson ImmunoResearch Lab-oratories, West Grove, PA), respectively. Antibodies of IgE andIgG1, IgG2, IgG3 and IgG4 subclasses were detected using biotin-conjugated goat or mouse anti-human monoclonal antibodies (IgE:Vector Laboratories, Burlington, Ca. IgG1 and IgG2: PharMingen,San Diego, CA. IgG3: Caltag Laboratories, Burlingame, CA. IgG4:Sigma, St. Louis, MO) and alkaline phosphatase conjugated strepta-vidine (Mabtech AB). The assay was developed with p-nitrophenylphosphate disodium salt (Sigma, St. Louis, MO) as substrate andthe optical densities were read at 405 nm in ELISA plate reader(VmaxTM Kinetic Microplate Reader, Menlo Park, CA).

Seropositivity was defined as the mean antibody level +2SD of 10non-malaria-exposed Swedish donors not known to have history

were calculated from standard curves obtained in a sandwich ELISAwith six dilutions of myeloma protein of IgG1–4 isotypes (Serotec,Oxford, UK) or with highly purified IgG, IgM (Jackson ImmunoRe-search Labs) and IgE (NIBSC, Hertfordshire, UK), respectively.

82 H.A. Balogun et al. / Acta Tropica 130 (2014) 80–87

* **

0

.1

.2

.3

.4

.5

.6

OD

valu

es

IgM OD IgG OD IgE OD0

.2

.4

.6

4-9 y 10-14 y 15-19 y 20-87 y

*

*

*IgG

OD

valu

es

IgM

0

.1

.2

.3

.4

.5

.6

.7

OD

valu

es

4-9 y 10-14 y 15-19 y 20-87 y

**

*

0

.005

.01

.015

.02

.025

.03

.035

.04

4-9 y 10-14 y 15-19 y 20-87 y

IgE

OD

valu

es

p= 0.07 *

F the lei = 18),p spond

2

(aim

wotavciptpt(sviwaa

ig. 1. Anti-Pf332-DBL reactivity pattern in malaria-exposed individuals, depictingn different age groups of 4–9 years (n = 15), 10–14 years (n = 17), 15–19 years (nercentiles and the line indicates the median (50th percentile). The whiskers corre

.4. Analysis

Statistical analyzes were performed with JMP® and Statviewversion 5.0; SAS Institute) softwares. Significance was consideredt a probability value of p ≤ 0.05. In univariate analysis, differencesn antibody levels were tested using Mann–Whitney test and Spear-

an’s rank correlation coefficient (�).Multiple regression was used to identify factors of importance

ith regard to the number of clinical malaria attacks during periodsf time. The Akaike Information criterion was used when choosinghe optimal number of parameters to include in the model aimedt explaining the incidence of malaria attacks identified in Dielmoillage from the date of blood sampling and up to 3 years after theollection of plasma samples used in ELISA studies. In all cases, thenitial selection of explanatory variables included age, the exacteriod of time spent in the village by each of the individual tested,he number of infected mosquito bites determined during the studyeriod, the level of blood parasitemia found at the time of sampling,he presence or absence of G6PD deficit and the presence of Hb ASversus that of Hb AA). Variables were then excluded according to atepwise procedure. A test of the null hypothesis was based on theariance ratio designated F and the significant variables were tested

n a model fitted with a Poisson distribution. Significant variables

ere identified by checking the difference between the full modelnd the model without the term tested. Significance was consideredt a probability value of p ≤ 0.05.

vels of Immunoglobulin G (IgG), IgM and IgE reactivity to recombinant Pf332-DBL20–87 years (n = 50). The boxes represent values between the 25th and the 75thto minimum and maximum values. *p < 0.05, **p < 0.0001.

Correlation analyses with other Pf332 fragments; EB200 andPf332-C231 were carried out with data from previous studies inwhich the same Senegalese samples were used (Ahlborg et al.,2002; Israelsson et al., 2008).

3. Results

3.1. Anti-Pf332-DBL reactivity pattern in malaria-exposedindividuals

Sera from 100 asymptomatic malaria exposed adults from Sene-gal were assayed for anti-Pf332-DBL IgG, IgM and IgE activities.The IgG response to Pf332-DBL was the highest, followed by IgMand IgE. (Fig. 1). When individuals were age stratified the pat-tern of immunological response observed in 4 different age groups(4–9, 10–14, 15–19, 20–87 years) differed with the Ig isotypes. TheIgG response displayed a moderate tendency to increase with age(� = 0.23 p = 0.02), with an increase up to age group 15–19, whileresponse in the oldest age group was similar to age group 10–14years (Fig. 1). There was no increase with age in IgM (� = 0.17p = 0.09), where similar levels across the older age groups were

observed, except for age group 4–9 years, which had lower levels,and a significant difference (p < 0.05) when compared to the olderage groups (Fig. 1). Regarding IgE, the highest level of response wasseen in age group 10–14 years (Fig. 1).

H.A. Balogun et al. / Acta Tropica 130 (2014) 80–87 83

p= 0.01

OD

val

ues

IgG1

0

.1

.2

.3

.4

.5

4-9 y 10-14 y 15-19 y 20-87 y

p= 0.003 IgG2

0

.1

.2

.3

.4

.5

.6

.7

.8

.9

4-9 y 10-14 y 15-19 y 20-87 y

OD

val

ues

IgG3

0

.05

.1

.15

.2

.25

.3

.35

.4

.45

4-9 y 10-14 y 15-19 y 20-87 y

OD

val

ues

p= 0.022

p= 0.04

p= 0.002

OD

val

ues

0

.01

.02

.03

.04

.05

.06

.07

.08

.09

.1

4-9 y 10-14 y 15-19 y 20-87 y

IgG4

p= 0.003

p= 0.03

p= 0.05

Fig. 2. IgG subclass pattern of anti-Pf332-DBL-specific reactivity in different age-groups of Dielmo villagers. The level of reactivity to the antigen is represented by OD values.E dicate

wblppa(

3o

Poas

D(pa

anti-Pf332-DBL IgM antibodies and those reactive with Pf332-C231(� = 0.711, p = <0.0001) (Fig. 3). The corresponding correlation foranti-Pf332-EB200 antibodies could not be determined, as the IgMdata for this fragment are not available.

Table 1Correlations between antibody classes and subclasses to Pf332-DBL.

IgG IgM IgE IgG 1 IgG2 IgG 3 IgG 4

IgGIgM 0.7IgE 0.37 0.47IgG 1 0.89 0.80 0.42IgG 2 0.14* 0.13* 0.16* 0.12*

IgG 3 0.78 0.77 0.55 0.83 0.16*

IgG 4 0.30 0.46 0.51 0.42 0.04** 0.42

The values used for analysis are the optical density ELISA values. Correlation

ach box plot represents values between the 25% to the 75% quartile and the line in

Regarding IgG subclass reactivities, the patterns were in lineith most anti-malarial responses, median antibody responses

eing predominantly IgG1 and IgG3, followed by comparativelyower but detectable levels of IgG2 and IgG4. IgG3 (� = 0.4= <0.0001) responses increased markedly with age. IgG1 (� = 0.24;= 0.02) and IgG4 (� = 0.23 p = 0.03) responses increased moder-tely with age while IgG2 (� = 0.06; p = 0.6) responses did notFig. 2).

.2. Correlations between anti-Pf332-DBL antibodies, and withther Pf332-fragments

Correlation analyses were performed between the levels allf332-DBL antibodies (IgG, IgM, IgE, and IgG1 to IgG4 subclasses)btained by ELISA. All immunoglobulin responses were positivelyssociated, except for IgG2, which displayed consistent low non-ignificant associations (Table 1).

In addition, correlation analyses were carried out with Pf332-

BL antibodies and previously studied Pf332-fragments; EB200

Ahlborg et al., 2002) and Pf332-C231 (Israelsson et al., 2008). Aositive correlation between the levels of Pf332-DBL reactive IgGntibodies and those with Pf332-C231 specificity was detected

s the median. The whiskers correspond to minimum and maximum values.

(� = 0.706, p = <0.0001), while a lower degree of correlation (� = 0.4,p = 0.0002) was seen with the IgG antibodies to Pf332-EB200.Similarly, there was a positive correlation between the levels of

coefficients were determined using Spearman rank correlation using all samples(n = 100). All correlations are significant at a p value of <0.001, unless otherwisestated.

* p < 0.2.** p = 0.7.

84 H.A. Balogun et al. / Acta Tropica 130 (2014) 80–87

IgMIgG

-,2

0

,2

,4

,6

,8

1

1,2

1,4

1,6

0 ,2 ,4 ,6 ,8 1 1,2 1,4

OD DBL

OD

C23

1

-,5

0

,5

1

1,5

2

2,5

3

-,25 0 ,25 ,5 ,75 1 1,25 1,5 1,75 2 2,25

OD

C23

1

OD DBL

bodies

3n

3dwwbsmi

atIfmDvaowetabl(

dhtaaa

TD

Fr

Fig. 3. Correlation of optical density ELISA values of IgG and IgM anti

.3. Association between Pf332-DBL-specific antibodies and theumber of malaria attacks

The number of clinical malaria attacks identified during up to6 months of active follow-up after sampling and the number ofays actually spent in Dielmo village by each of the 100 inhabitantsere determined. The number of attacks decreased significantlyith the age of the villagers after 2 years of follow-up (Table 2),

ut increased significantly in association with the level of para-itemia determined at the time of blood sampling (not shown), andarginally also with the time spent in the Dielmo village by each

ndividual (Table 2).A potential relationship between Pf332-DBL reactive antibodies

nd the number of malaria attacks detected in Dielmo was inves-igated by multivariate analysis. When either total anti-Pf332-DBLgG, IgM or IgE responses were tested, none of these variables wasound significantly associated with or predictive of the number of

alaria attacks (data not shown). In contrast, when anti-Pf332-BL IgG1 to IgG4 responses were tested as continuous explanatoryariables, IgG1 were found positively (p = 0.0258) and IgG3 neg-tively (p = 0.0004) associated with the number of malaria attacksbserved after 2 years of survey. When anti-Pf332-DBL IgG1 or IgG3ere tested separately and when controlling for other potential

xplanatory variables (including age, the time spent in the village,he number of infected bites during the study period, the blood par-sitemia level, G6PD deficit, HbAA or AS), the positive associationetween the number of malaria attacks and IgG1 responses was

ost while the negative association with IgG3 remained significantp = 0.0049).

When tested in a model fitted to the data with a Poissonistribution, the anti-Pf332-DBL IgG3 responses were found toave a significant effect on the number of malaria attacks iden-

ified in the villagers. The difference between the model includingnti-Pf332-DBL IgG3, and the model without this anti-Pf332-DBLntibody response was marked after 1 year (�2 = 4.42; p = 0.0355),s well as after 2 years (�2 = 18.56; p = 0.000016) of clinical

able 2istribution of malaria attack detected by age groups after various periods of individual d

Mean age ± 1SD 7.4 ±No. of individuals N = 17Attacks/6 months (person-days at risk) incidence rate per person per year 11 (30Attacks/1 year (person-days at risk) incidence rate per person per year 22 (57Attacks/2 years (person-days at risk) incidence rate per person per year 36 (11Attacks/3 years (person-days at risk) incidence rate per person per year 48 (16

or each age group and each period of study, the total number of clinical malaria attacksisk is shown (in brackets) and the incidence rate of malaria attacks is indicated per perso

against DBL and C231 fragments of Pf332 in Senegalese individuals.

survey and remained consistent by year three of the study(�2 = 10.72; p = 0.00106). Thus, results strongly suggest a potentialrelationship between the presence of anti-Pf332-DBL IgG3 and areduced number of malaria attacks in some Dielmo villagers.

3.4. Ratio of anti-Pf332-DBL cytophilic to non-cytophilicantibodies in relation to the number of malaria attacks

Since a potential opposing effect was seen earlier regardingthe number of malaria attacks and IgG3 or IgG1 responses(Ahlborg et al., 2002; Israelsson et al., 2008), the anti-Pf332-DBLcytophilic:non-cytophilic antibody ratio and the anti-Pf332-DBLIgG3:IgG1 ratio levels were also tested. When estimating the num-ber of malaria attacks detected after 1 year and more of follow-up,as a function of the ratios of cytophilic to non cytophilic antibodyresponses (�2 = 8.16; p = 0.00428), or as a function of the IgG3:IgG1ratios (�2 = 31.46; p = 0.0002 × 10−5), a marked influence of thesevariables was observed and in the model fitted to the data with aPoisson distribution.

3.5. Relationship between anti-Pf332-DBL IgG3 levels andmalaria attacks

Individuals with anti-Pf332-DBL IgG3 levels above the medianvalue were found in the subgroup of 49 Dielmo villagers with thelowest number of malaria attack detected during the 3 years ofactive follow-up (Table 3). In addition, between 6 months and 3years of active follow-up, the percentage of individuals who pre-sented with a malaria attack increased by 19.6% in the subgroupwith anti-Pf332-DBL IgG3 levels under the median and by 10.2% inthe subgroup with anti-Pf332-DBL IgG3 levels above the median(Table 3).

By selecting 84 villagers of less than 45 years of age, the meanages of low or high anti-Pf332-DBL IgG1 or IgG3 responders wassimilar (19.8 ± 11.6 and 21.2 ± 10.4, respectively). After 3 years ofstudy, the malaria attack incidence rate was 0.673 per person per

aily active follow-up of Dielmo villagers.

1.6 14.9 ± 3.3 28.9 ± 5.8 53.2 ± 13.1

N = 32 N = 30 N = 2143) 1.320 10 (5600) 0.652 2 (4869) 0.150 1 (3455) 0.10697) 1.386 16 (11104) 0.527 2 (9261) 0.079 1 (6907) 0.053339) 1.159 30 (21392) 0.512 2 (18212) 0.040 2 (13736) 0.053847) 1.041 41 (31379) 0.477 7 (27480) 0.093 2 (20816) 0.035

determined per period of survey is indicated. The total number of person-days atn per year in italics.

H.A. Balogun et al. / Acta Tropica 130 (2014) 80–87 85

Table 3Number of individuals exposed to (and % prevalence of) malaria attacks over time, in relation to levels of anti-Pf332-DBL IgG3.

Anti-Pf332-DBLIgG3 < median (N = 51)

Anti-Pf332-DBLIgG3 > median (N = 49)

Attacks (+) after 6 months 13 2Attacks (−) after 6 months 38 47 �2: 8.98; p = 0.0027Prevalence of attacks 25.49% 4.08% OR = 8.04 [1.71 -37.84]; P = 0.0066Attacks (+) after 1 year 16 2Attacks (−) after 1 year 35 47 �2: 12.61 p = 0.0004Prevalence of attacks 31.37% 4.08% OR = 10.74 [2.32 -49.80]; p = 0.001Attacks (+) after 2 years 20 4Attacks (−) after 2 years 31 45 �2: 13.21 p = 0.0003Prevalence of attacks 39.22% 8.16% OR = 7.26 [2.26 -23.31]; p = 0.0006Attacks (+) after 3 years 23 7Attacks (−) after 3 years 28 42 �2: 11.30 p = 0.0008Prevalence of attacks 45.1% 14.29% OR = 4.93 [1.86–13.02]; p = 0.00167

A dian va delinv e bloo

ysfus8(toocc

4

iitloptat

Dabiioplri

TD

Bwa

nti-Pf332-DBL IgG3 levels were tested as a dichotomous variable by using the mebove the median value of 4.0 �g, and as a result, two subgroups of responders wereariable and logistic regression analyses were carried out at different times after th

ear (in low anti-Pf332-DBL IgG3 responders) and 0.211 per per-on per year (in high anti-Pf332-DBL IgG3 responders) and this 3.19old difference was significant (Wilcoxon test, p = 0.0099). Of note,p to 45.2% (19/42) of the individuals with anti-Pf332-DBL IgG3erum concentrations below the median value determined for the4 villagers selected had a malaria attack. In contrast, only 21.4%9/42) of the villagers with anti-Pf332-DBL IgG3 serum concentra-ions above the median value presented with a clinical episodef malaria (�2: 5.36 p = 0.0206). By comparison, the distributionf villagers with or without malaria attack was not found signifi-ant when anti-Pf332-DBL IgG1 responses were tested in similaronditions (Table 4).

. Discussion

Understanding the immune responses elicited by P. falciparumnfection in human beings living in malaria endemic areas, is anmportant prerequisite for understanding immunity. It is thoughthat acquired immunity targeting blood stage antigens acts byimiting parasite replication, thereby preventing the developmentf high density parasitemia, but is less effective at protecting fromarasite infection per se (Langhorne et al., 2008). Nevertheless, pro-ection from malaria is only relative, depending on multiple factors,nd so far there exists no absolute marker to accurately characterizehis status (Marsh and Kinyanjui, 2006).

In the present study, the pattern of antibody reactivity to Pf332-BL (Moll et al., 2007), a recombinant fragment of the blood-stagentigen Pf332 (Mattei and Scherf, 1992), was analyzed. This recom-inant fragment was shown to be recognized by antibodies from

ndividuals living in a malaria endemic area of Senegal. Our find-ngs suggest a somewhat singular pattern of antibody recognitionf Pf332-DBL as compared to the other fragments of Pf332 studied

reviously (Ahlborg et al., 2002; Israelsson et al., 2008). Neverthe-

ess, the results still support earlier findings seen regarding immuneesponses to Pf332 exhibited by the very same individuals stud-ed in Dielmo village (Kulane et al., 1999; Perraut et al., 2000;

able 4ielmo villagers with different levels of anti-Pf332-DBL IgG1 or IgG3 responses displayed

Period of study IgG1 responses Noattack

Yes attack (%) Chisquare a(p value)

0 to 6 months IgG1 ≤ 6.8 �g/ml 32 10 (23.8%) 0.86 (0.35)IgG1 ≥ 6.8 �g/ml 38 4 (9.5%)

0 to 3 years IgG1 ≤ 6.8 �g/ml 26 16 (38.1%) 3.09 (0.079)IgG1 ≥ 6.8 �g/ml 30 12 (28.6%)

y excluding individuals of more than 45 years of age, anti-Pf332-DBL IgG1 or IgG3 respere used to delineate subgroups of low and high antibody responders with the same me

fter 6 months and up to 3 years of active, continuous and daily active follow-up of the D

alue of IgG3 concentrations as a cut off level. (IgG3 responses were either under oreated). The presence or absence of malaria attacks was then used as a dichotomousd sampling.

Ahlborg et al., 2002; Israelsson et al., 2008), as well as recent find-ings showing the seroprevalence of anti-Pf332-DBL antibodies indistinct malaria endemic regions (Nilsson et al., 2011). All antibodyclasses studied (IgG, IgM, IgE) were detected in all age groups, withIgG responses being the highest, followed by IgM and IgE. The effectof age on the level of antibodies detected was mostly observed forIgG, IgG3, IgG1 and IgG4. In an earlier study, anti-C231-IgE antibodylevels were found to be associated with a reduced risk of malariaattacks (Israelsson et al., 2008), while no such association was seenwith anti-DBL-IgE antibodies in the present study.

Several malaria antigens are known to induce IgG1 and IgG3responses, and this appears to be dependent on factors such as anti-genic composition and age of the infected individual (Tongren et al.,2006; Matondo Maya et al., 2006). Regarding the anti-Pf332-DBLIgG subclasses, IgG1 and IgG3 were the predominant subclassesdetected, but low amounts of IgG2 and IgG4 were also found. How-ever, the youngest age group (of 4–9 years) displayed relativelyconsistent amounts of IgG2 when compared to older age groups.The levels of most classes and subclasses of antibodies showed sig-nificant positive associations. One major exception was the IgG2antibodies, which consistently displayed only low associations withthe other antibodies. This finding is in contrast with the anti-Pf332-C231 antibody responses, were IgG2 antibodies were detected atappreciable amounts (Israelsson et al., 2008; Nasr et al., 2007). Ofnote, in another epidemiological setting, low levels of IgG2 anti-bodies reactive with Pf332-C231 were associated with a lower riskof severe malaria (Nasr et al., 2007).

The different patterns of IgG subclasses observed in the antibodyresponses against Pf332-DBL and the previously studied Pf332-EB200 and Pf332-C231 fragments is not surprising, as these Pf332fragments, differ considerably in their amino acid compositions.Pf332-EB200 is derived from the central region of Pf332, which is

dominated by glutamic acid rich repeats (Mattei and Scherf, 1992).Pf332-C231 is derived from the C-terminal region of Pf332, withhigher complexity of amino acids, and almost lacking glutamic acidrepeats (Balogun et al., 2009). Pf332-DBL represents the N-terminal

different susceptibility to the occurrence of clinical episodes of malaria.

nd IgG3 responses Noattack

Yes attack (%) Chisquare and(p value)

IgG3 ≤ 3.62 �g/ml 31 11 (26.2%) 5.36 (0.0206)IgG3 ≥ 3.62 �g/ml 38 4 (9.5%)IgG3 ≤ 3.62 �g/ml 23 19 (45.2%) 3.98 (0.0461)IgG3 ≥ 3.62 �g/ml 33 9 (21.4%)

onses above or under the median values determined for the 84 selected villagersan age. They were tested with regard to the occurrence of malaria attacks observedielmo villagers.

8 ta Trop

r(ag(B

otwtpaiwmp

paavhoioadt

cfripepatrtim

A

t

Rtt5E

R

A

A

A

B

6 H.A. Balogun et al. / Ac

egion of the antigen, which contains a Duffy binding-like domainMoll et al., 2007). Thus, these Pf332 fragments possess differentntigenic capabilities, and as earlier postulated, a megadalton anti-en like Pf332 may engage different regions in various functionsHodder et al., 2009; Glenister et al., 2009; Waller et al., 2010;alogun et al., 2011).

It is of interest to relate the human immune response, in termsf antibodies generated, to protection, as this may provide a bet-er understanding on mechanisms involved in parasite control, asell as establish the basis for a better understanding of immunity

o human malaria in exposed individuals. In the present study theossible association between anti-Pf332-DBL antibody responsesnd the number of malaria attacks experienced by the donors, wasnvestigated. When either total anti-Pf332-DBL IgG, or IgM or IgE

ere added in the model selected to explain the occurrence ofalaria attacks, none of these variables were associated with or

redictive of the number of malaria attacks.Knowledge on antibody IgG subclass responses associated with

rotection against malaria vary, especially with regard to specificntigens (Stanisic et al., 2009). In the present study, high levels ofnti-Pf332-DBL IgG3 were found in the blood samples of Dielmoillagers with a reduced prospective risk of malaria attack. Thisumoral response might correspond either to a surrogate markerf protection or reflect an involvement of cytophilic IgG3 antibod-es in an anti-parasite immune mechanism involved in the controlf parasite multiplication (Baird, 1995), but their contribution inn anti-disease immunity can not be excluded. Indeed, associationoes not really mean causation, but without doubts our observa-ions are original and stimulate further investigations.

Within the context of the present study, there appears to be aertain level of anti-Pf332-DBL IgG3 antibodies, which is frequentlyound as contemporary of protection in villagers with a reducedisk of malaria attack in Dielmo, Senegal, controlling for confound-ng effect of age, malaria transmission intensity and several geneticrotective factors. This observation might be of potential inter-st, urging further studies of antibodies to the Pf332-DBL antigen,referably in different malaria endemic settings. The present studylso strengthens the interest of antigen Pf332, which contains epi-opes able to induce potentially parasite neutralizing antibodyesponses, apparently associated with a certain level of protec-ion against malaria attacks. This antigen certainly deserves furthernvestigations aiming at in detail to characterize its potential as a

alaria vaccine candidate.

cknowledgement

We wish to thank Dr. Nnaemeka Iriemenam for his initial con-ribution to this work.

This work was supported by grants from the Swedish Agency foresearch Development with Developing Countries (SIDA, SAREC),he Swedish Medical research Council (VR) and a grant withinhe BiomalPar European Network of Excellence (LSMP-CT-2004-03578). KB is an affiliated member of the EVIMalaR Network ofxcellence.

eferences

hlborg, N., Haddad, D., Siddique, A.B., Roussilhon, C., Rogier, C., Trape, J.F., Troye-Blomberg, M., Berzins, K., 2002. Antibody responses to the repetitive Plasmodiumfalciparum antigen Pf332 in humans naturally primed to the parasite. Clin. Exp.Immunol. 129, 318–325.

hlborg, N., Iqbal, J., Bjork, L., Ståhl, S., Perlmann, P., Berzins, K., 1996. Plasmodiumfalciparum: differential parasite growth inhibition mediated by antibodies to theantigens Pf332 and Pf155/RESA. Exp. Parasitol. 82, 155–163.

ucan, C., Traore, Y., Tall, F., Nacro, B., Traore-Leroux, T., Fumoux, F., Rihet, P., 2000.High immunoglobulin G2 (IgG2) and low IgG4 levels are associated with humanresistance to Plasmodium falciparum malaria. Infect. Immun. 68, 1252–1258.

aird, J.K., 1995. Host age as a determinant of naturally acquired immunity to Plas-modium falciparum. Parasitol. Today 11, 105–111.

ica 130 (2014) 80–87

Balogun, H.A., Awah, N.W., Farouk, S.E., Berzins, K., 2011. Pf332-C231-reactiveantibodies affect growth and development of intra-erythrocytic Plasmodiumfalciparum parasites. Vaccine 30, 21–28.

Balogun, H.A., Vasconcelos, N.M., Lindberg, R., Haeggström, M., Moll, K., Chen, Q.,Wahlgren, M., Berzins, K., 2009. Immunogenicity and antigenic properties ofPf332-C231, a fragment of a non-repeat region of the Plasmodium falciparumantigen Pf332. Vaccine 28, 90–97.

Bolad, A., Berzins, K., 2000. Antigenic diversity of Plasmodium falciparum andantibody-mediated parasite neutralization. Scand. J. Immunol. 52, 233–239.

Bouharoun-Tayoun, H., Druilhe, P., 1992. Plasmodium falciparum malaria: evidencefor an isotype imbalance which may be responsible for delayed acquisition ofprotective immunity. Infect. Immun. 60, 1473–1481.

Bull, P.C., Marsh, K., 2002. The role of antibodies to Plasmodium falciparum-infected-erythrocyte surface antigens in naturally acquired immunity to malaria. TrendsMicrobiol. 10, 55–58.

Cohen, S., McGregor, I.A., Carrington, S., 1961. Gamma-globulin and acquired immu-nity to human malaria. Nature 192, 733–737.

Courtin, D., Oesterholt, M., Huismans, H., Kusi, K., Milet, J., Badaut, C., Gaye, O., Roef-fen, W., Remarque, E.J., Sauerwein, R., Garcia, A., Luty, A.J., 2009. The quantityand quality of African children’s IgG responses to merozoite surface antigensreflect protection against Plasmodium falciparum malaria. PLoS One 4, e7590.

Du, C., Nilsson, S., Lu, H., Yin, J., Jiang, N., Wahlgren, M., Chen, Q., 2010. Immuno-genicity of the Plasmodium falciparum Pf332-DBL domain in combination withdifferent adjuvants. Vaccine 28, 4977–4983.

Garraud, O., Perraut, R., Riveau, G., Nutman, T.B., 2003. Class and subclass selectionin parasite-specific antibody responses. Trends Parasitol. 19, 300–304.

Glenister, F.K., Fernandez, K.M., Kats, L.M., Hanssen, E., Mohandas, N., Coppel, R.L.,Cooke, B.M., 2009. Functional alteration of red blood cells by a mega-Daltonprotein of Plasmodium falciparum. Blood 113, 919–928.

Haeggström, M., Kironde, F., Berzins, K., Chen, Q., Wahlgren, M., Fernandez, V., 2004.Common trafficking pathway for variant antigens destined for the surface of thePlasmodium falciparum-infected erythrocyte. Mol Biochem Parasitol 33, 1–14.

Hay, S.I., Snow, R.W., 2006. The malaria atlas project: developing global maps ofmalaria risk. PLoS Med. 3, e473.

Hayes, R.J., Marsh, K., Snow, R.W., 1992. Case-control studies of severe malaria. J.Trop. Med. Hyg. 95, 157–166.

Hinterberg, K., Scherf, A., Gysin, J., Toyoshima, T., Aikawa, M., Mazie, J.C., Da Silva,L.J., Mattei, D., 1994. Plasmodium falciparum: the Pf332 antigen is secreted fromthe parasite by a brefeldin A-dependent pathway and is translocated to theerythrocyte membrane via the Maurer’s clefts. Exp. Parasitol. 79, 279–291.

Hodder, A.N., Maier, A.G., Rug, M., Brown, M., Hommel, M., Pantic, I., Puig-de-Morales-Marinkovic, M., Smith, B., Triglia, T., Beeson, J., Cowman, A.F., 2009.Analysis of structure and function of the giant protein Pf332 in Plasmodiumfalciparum. Mol. Microbiol. 71, 48–65.

Israelsson, E., Balogun, H., Vasconcelos, N.M., Beser, J., Roussilhon, C., Rogier, C.,Trape, J.F., Berzins, K., 2008. Antibody responses to a C-terminal fragment ofthe Plasmodium falciparum blood-stage antigen Pf332 in Senegalese individualsnaturally primed to the parasite. Clin. Exp. Immunol. 152, 64–71.

Kulane, A., Siddique, A.B., Sarthou, J.L., Tall, A., Dieye, A., Perlmann, H., Perlmann,P., Troye-Blomberg, M., Ahlborg, N., 1999. Human immune responses to thehighly repetitive Plasmodium falciparum antigen Pf332. Am. J. Trop. Med. Hyg.61, 141–148.

Langhorne, J., Ndungu, F.M., Sponaas, A.M., Marsh, K., 2008. Immunity to malaria:more questions than answers. Nat. Immunol. 9, 725–732.

Marsh, K., Kinyanjui, S., 2006. Immune effector mechanisms in malaria. ParasiteImmunol. 28, 51–60.

Matondo Maya, D.W., Mavoungou, E., Deloron, P., Theisen, M., Ntoumi, F., 2006.Distribution of IgG subclass antibodies specific for Plasmodium falciparumglutamate-rich-protein molecule in sickle cell trait children with asymptomaticinfections. Exp. Parasitol. 112, 92–98.

Mattei, D., Scherf, A., 1992. The Pf332 gene codes for a megadalton protein of Plas-modium falciparum asexual blood stages. Mem. Inst. Oswaldo Cruz 87 (Suppl 3),163–168.

Moll, K., Chene, A., Ribacke, U., Kaneko, O., Nilsson, S., Winter, G., Haeggström, M.,Pan, W., Berzins, K., Wahlgren, M., Chen, Q., 2007. A novel DBL-domain of the P.falciparum 332 molecule possibly involved in erythrocyte adhesion. PLoS One 2,e477.

Nasr, A., Iriemenam, N.C., Troye-Blomberg, M., Giha, H.A., Balogun, H.A., Osman,O.F., Montgomery, S.M., ElGhazali, G., Berzins, K., 2007. Fc gamma receptor IIa(CD32) polymorphism and antibody responses to asexual blood-stage antigensof Plasmodium falciparum malaria in Sudanese patients. Scand. J. Immunol. 66,87–96.

Nilsson, S., Moll, K., Angeletti, D., Albrecht, L., Kursula, I., Jiang, N., Sun, X., Berzins, K.,Wahlgren, M., Chen, Q., 2011. Characterization of the Duffy-binding like domainof Plasmodium falciparum blood stage antigen 332. Malar. Res. Treat., 671439,http://dx.doi.org/10.4061/2011/.

Nilsson, S., Angeletti, D., Chen, Q., Wahlgren, M., Moll, K., 2012. Plasmodium falcipa-rum antigen 332 is a resident peripheral membrane protein of Maurer’s clefts.PLoS One 7, e46980.

Ntoumi, F., Flori, L., Mayengue, P.I., Matondo Maya, D.W., Issifou, S., Deloron, P., Lell,B., Kremsner, P.G., Rihet, P., 2005. Influence of carriage of hemoglobin AS and the

Fc gamma receptor IIa-R131 allele on levels of immunoglobulin G2 antibodiesto Plasmodium falciparum merozoite antigens in Gabonese children. J. Infect. Dis.192, 1975–1980.

Perraut, R., Mercereau-Puijalon, O., Diouf, B., Tall, A., Guillotte, M., Le Scanf, C., Trape,J.F., Spiegel, A., Garraud, O., 2000. Seasonal fluctuation of antibody levels to

ta Trop

R

R

S

S

H.A. Balogun et al. / Ac

Plasmodium falciparum parasitized red blood cell-associated antigens in twoSenegalese villages with different transmission conditions. Am. J. Trop. Med.Hyg. 62, 746–751.

ogier, C., Commenges, D., Trape, J.F., 1996. Evidence for an age dependent pyrogenicthreshold of Plasmodium falciparum parasitemia in highly endemic population.Am. J. Trop. Med. Hyg. 54, 613–619.

ogier, C., Tall, A., Diagne, N., Fontenille, D., Spiegel, A., Trape, J.F., 1999. Plasmo-dium falciparum clinical malaria lessons from longitudinal studies in Senegal.Parassitologia 41, 255–259.

abchareon, A., Burnouf, T., Ouattara, D., Attanath, P., Bouharoun-Tayoun, H., Chan-tavanich, P., Foucault, C., Chongsuphajaisiddhi, T., Druilhe, P., 1991. Parasitologicand clinical human response to immunoglobulin administration in falciparum

malaria. Am J Trop Med Hyg 45, 297–308.

copel, K.K., Fontes, C.J., Ferreira, M.U., Braga, E.M., 2006. Factors associated withimmunoglobulin G subclass polarization in naturally acquired antibodies toPlasmodium falciparum merozoite surface proteins: a cross-sectional survey inBrazilian Amazonia. Clin. Vac. Immunol. 13, 810–813.

ica 130 (2014) 80–87 87

Stanisic, D.I., Richards, J.S., McCallum, F.J., Michon, P., King, C.L., Schoepflin, S., Gilson,P.R., Murphy, V.J., Anders, R.F., Mueller, I., Beeson, J.G., 2009. Immunoglobulin Gsubclass-specific responses against Plasmodium falciparum merozoite antigensare associated with control of parasitemia and protection from symptomaticillness. Infect. Immun. 77, 1165–1174.

Tongren, J.E., Corran, P.H., Jarra, W., Langhorne, J., Riley, E.M., 2005. Epitope-specificregulation of immunoglobulin class switching in mice immunized with malarialmerozoite surface proteins. Infect. Immun. 73, 8119–8129.

Tongren, J.E., Drakeley, C.J., McDonald, S.L., Reyburn, H.G., Manjurano, A., Nkya,W.M., Lemnge, M.M., Gowda, C.D., Todd, J.E., Corran, P.H., Riley, E.M., 2006.Target antigen, age, and duration of antigen exposure independently reg-ulate immunoglobulin G subclass switching in malaria. Infect. Immun. 74,

257–264.

Waller, K.L., Stubberfield, L.M., Dubljevic, V., Buckingham, D.W., Mohandas, N.,Coppel, R.L., Cooke, B.M., 2010. Interaction of the exported malaria proteinPf332 with the red blood cell membrane skeleton. Biochim. Biophys. Acta 1798,861–871.