Association analyses of NCR3 polymorphisms with P. falciparum mild malaria.
ABSTRACT Plasmodium falciparum malaria is a major cause of morbidity and mortality in many developing countries especially in sub-Saharan Africa. A susceptibility locus for mild malaria has been mapped to the MHC region, and TNF polymorphisms have been associated with mild malaria. The Natural Cytotoxicity-triggering Receptor 3 (NCR3) gene is located in the peak region of linkage, and is 15kb distal to TNF. In this study, we considered NCR3 as a candidate gene, and we genotyped ten NCR3 single nucleotide polymorphisms (SNPs). Here, we report evidence of an association between mild malaria and NCR3 -412G>C polymorphism located within the promoter. Population-based association analysis showed that NCR3 -412C carriers had more frequent mild malaria attacks than NCR3 -412GG individuals (P=0.001). Using the family-based association test (FBAT) program and its phenotype (PBAT) option, we further found that NCR3 -412C (P=0.0009) and a haplotype containing NCR3 -412C (P=0.008) were significantly associated with increased risk of mild malaria, and that the association was not due to the association of TNF with mild malaria. These observations suggest that there are at least two genes located on the central region of MHC involved in genetic control of human malaria. The association of NCR3 with malaria should provide new insights into the role of Natural Killer cells in this common disease.
- SourceAvailable from: Klaus Schughart[Show abstract] [Hide abstract]
ABSTRACT: The first scientific meeting of the newly established European SYSGENET network took place at the Helmholtz Centre for Infection Research (HZI) in Braunschweig, April 7-9, 2010. About 50 researchers working in the field of systems genetics using mouse genetic reference populations (GRP) participated in the meeting and exchanged their results, phenotyping approaches, and data analysis tools for studying systems genetics. In addition, the future of GRP resources and phenotyping in Europe was discussed.Mammalian Genome 08/2010; 21(7-8):331-6. · 2.42 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Natural killer (NK) cells influence innate and adaptive immune host defenses. Existing data indicate that manipulating the balance between inhibitory and activating NK receptor signals, the sensitivity of target cells to NK cell-mediated apoptosis, and NK cell cross-talk with dendritic cells might hold therapeutic promise. Efforts to modulate NK cell trafficking into inflamed tissues and/or lymph nodes, and to counteract NK cell suppressors, might also prove fruitful in the clinic. However, deeper investigation into the benefits of combination therapy, greater understanding of the functional distinctions between NK cell subsets, and design of new tools to monitor NK cell activity are needed to strengthen our ability to harness the power of NK cells for therapeutic aims.Nature Immunology 06/2008; 9(5):486-94. · 26.20 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Vaccine development faces major difficulties partly because of genetic variation in both infectious organisms and humans. This causes antigenic variation in infectious agents and a high interindividual variability in the human response to the vaccine. The exponential growth of genome sequence information has induced a shift from conventional culture-based to genome-based vaccinology, and allows the tackling of challenges in vaccine development due to pathogen genetic variability. Additionally, recent advances in immunogenetics and genomics should help in the understanding of the influence of genetic factors on the interindividual and interpopulation variations in immune responses to vaccines, and could be useful for developing new vaccine strategies. Accumulating results provide evidence for the existence of a number of genes involved in protective immune responses that are induced either by natural infections or vaccines. Variation in immune responses could be viewed as the result of a perturbation of gene networks; this should help in understanding how a particular polymorphism or a combination thereof could affect protective immune responses. Here we will present: i) the first genome-based vaccines that served as proof of concept, and that provided new critical insights into vaccine development strategies; ii) an overview of genetic predisposition in infectious diseases and genetic control in responses to vaccines; iii) population genetic differences that are a rationale behind group-targeted vaccines; iv) an outlook for genetic control in infectious diseases, with special emphasis on the concept of molecular networks that will provide a structure to the huge amount of genomic data.Brazilian journal of medical and biological research = Revista brasileira de pesquisas medicas e biologicas / Sociedade Brasileira de Biofisica ... [et al.] 10/2011; 45(5):376-85. · 1.08 Impact Factor
Association analyses of NCR3 polymorphisms with
P. falciparum mild malaria
Nicolas F. Delahaye, Mathieu Barbier, Francis Fumoux, Pascal Rihet*
Universite ´ de la Me ´diterrane ´e, IFR 48, Faculte ´ de Pharmacie, Laboratoire de Pharmacoge ´ne ´tique
des Maladies Parasitaires-EA 864, 27 Bd Jean Moulin 13385 Marseille, France
Received 25 July 2006; accepted 7 November 2006
Available online 5 December 2006
Plasmodium falciparum malaria is a major cause of morbidity and mortality in many developing countries especially in sub-Saharan Africa.
A susceptibility locus for mild malaria has been mapped to the MHC region, and TNF polymorphisms have been associated with mild malaria.
The Natural Cytotoxicity-triggering Receptor 3 (NCR3) gene is located in the peak region of linkage, and is 15 kb distal to TNF. In this study, we
considered NCR3 as a candidate gene, and we genotyped ten NCR3 single nucleotide polymorphisms (SNPs). Here, we report evidence of an
association between mild malaria and NCR3 ?412G > C polymorphism located within the promoter. Population-based association analysis
showed that NCR3 ?412C carriers had more frequent mild malaria attacks than NCR3 ?412GG individuals (P ¼ 0.001). Using the family-
based association test (FBAT) program and its phenotype (PBAT) option, we further found that NCR3 ?412C (P ¼ 0.0009) and a haplotype
containing NCR3 ?412C (P ¼ 0.008) were significantly associated with increased risk of mild malaria, and that the association was not due
to the association of TNF with mild malaria. These observations suggest that there are at least two genes located on the central region of
MHC involved in genetic control of human malaria. The association of NCR3 with malaria should provide new insights into the role of Natural
Killer cells in this common disease.
? 2006 Elsevier Masson SAS. All rights reserved.
Keywords: Natural Killer Receptor; NKp30; Plasmodium falciparum; Mild malaria; Association; Genetic linkage
The outcome of malaria infection depends on parasite and
host genetic factors. In humans, twin studies and familial clus-
tering of the disease indicate that genetic factors control mild
malaria [1,2]. Candidate gene studies have evidenced associa-
tion of mild malaria with some genes, such as the beta-globin
gene (HBB) [3,4]. Recently, Mackinnon et al. only attributed a
small proportion of the variation of mild malaria incidence to
HbS and a-thalassemia, and pointed out that other genes are
involved in malaria resistance .
Linkage studies have mapped genes controlling mild
malaria on chromosome 6p21ep23 [5,6], with a peak close
to the TNF gene. In addition, several TNF polymorphisms
have been associated with mild malaria [6,7]. However, the
TNF allele frequencies were not high, suggesting that TNF
polymorphisms per se unlikely explain linkage of mild malaria
to the MHC region. This was further supported by linkage
analyses showing that linkage of mild malaria to the MHC
region remained significant when taking into account TNF
polymorphisms as covariates (Barbier et al., unpublished
data). This raised the possibility that other genes close to
TNF may be involved. Among the genes located within the
central region of MHC, the Natural Cytotoxicity-triggering
Receptor 3 (NCR3) gene that encodes the activating NK recep-
tor NKp30 could be considered as a candidate gene. Recently,
* Corresponding author. Tel./fax: þ33 4 9180 3674.
E-mail address: email@example.com (P. Rihet).
1286-4579/$ - see front matter ? 2006 Elsevier Masson SAS. All rights reserved.
Microbes and Infection 9 (2007) 160e166
it was shown that human NK cells directly interact with P. fal-
ciparum-infected erythrocytes (iRBC) and that it was neces-
sary for IFNg production [8,9]. In addition, NK cells were
shown to mediate cytolysis of iRBC [10,11]. It is thought
that neoantigens expressed on the surface of iRBC interact
with NK receptors, and that they provide an activating signal.
We therefore searched for NCR3 polymorphisms that might
be associated with mild malaria in an African population liv-
ing in a P. falciparum endemic area. We studied 34 families
comprising 193 individuals (53 parents and 140 sibs) with pre-
vious evidence of linkage .
We investigated the linkage disequilibrium (LD) patterns in
the central region of MHC, and we tested family-based asso-
ciation between NCR3 polymorphisms and phenotypes related
to mild malaria.
2. Materials and methods
The study population consisted of 193 individuals (53 par-
ents and 140 sibs) belonging to 34 families living in Burkina
Faso. Linkage of mild malaria to chromosome 6p21ep23 was
reported elsewhere in the study population . The mean age
of sibs was 12.1 ? 6.2 years (range 1e34 years). The study
population and the area of parasite exposure have been de-
scribed elsewhere . The Medical Authority of Burkina
Faso approved the study protocol.
2.2. Clinical data and phenotype determination
Clinical diagnosis and phenotype determination have been
described elsewhere . Briefly, according to the WHO, a di-
agnosis of malaria attack was based on P. falciparum parasite-
mia fever (axillary temperature more than 37.5?C) and the
classical symptoms (headache, aching, vomiting or diarrhea
in the children). According to the recommendation of the Cen-
tre National de Lutte Contre le Paludisme (CNLP) of Burkina
Faso, each episode of illness was treated with 25 mg/kg chlo-
roquine during three days or until recovery. Parasitemia was
checked at the end of the treatment. In all, 62 of the 193 family
members (57 sibs and 5 parents) presented at least one uncom-
plicated malaria attack during the survey. After treatment,
these individuals had neither parasite nor fever. They were
considered in the analysis as affected individuals. Sibs from
5 families were unaffected. Eleven, 13, 5, and 1 families
contained 1, 2, 3, and 5 affected sibs, respectively. The first
phenotype was thus a binary trait (P1).
To take into account the influence of known covariates on
the phenotype, we performed logistic regression using the
statistical SPSS software (SPSS, Boulogne, France). Since
age and hemoglobin genotype are known to influence the de-
velopment of malaria attacks [4,13], we adjusted the logit of
the probability P of malaria attack for age and hemoglobin ge-
notype . The residual z of the logistic regression model
was the second phenotype used in linkage and association
analyses (P2), as previously described [6,7]. All the sibs
were included in linkage and association analyses.
Phenotype P3 (maximum parasitemia) was based on a loga-
rithmic transformation of the highest parasitemia in each indi-
vidual. Determination of parasitemia was described in our
previous study . Briefly, each family was visited 20 times
during the 24 months of the study, and parasitemia was mea-
sured. In addition, parasitemia was measured during febrile
episodes. The mean number of parasitemia measurements
per subject was 15.3 ? 5.2 (range 1e24). Fingerprint periph-
eral blood samples were taken from all family members pres-
ent and thick and thin blood films were stained with Giemsa.
Only P. falciparum asexual forms were retained to determine
parasitemia. Parasitemia was defined as the number of parasit-
ized erythrocytes observed per ml in thin blood films. Maxi-
mum parasitemia was based on the highest parasitemia in
each individual during the study. Multiple polynomial regres-
sion was done with age and the number of measurements using
the statistical SPSS software (SPSS, Boulogne, France). The
explicative variables were treated as continuous variables.
The analysis revealed that age and the number of measure-
ments had an effect on maximum parasitemia (P ¼ 0.0001).
The residual of the regression model, which took into account
age and the number of measurements, was the third phenotype
used in association and linkage analyses. All the sibs were in-
cluded in linkage and association analyses.
DNA was extracted from mononuclear cells separated by
Ficoll-Hypaque density gradient as described . DNA was
preamplified with the Primer Extension Preamplification
method . Since exon 1 does not contain known SNP, and
since exon 3 contains only a synonymous SNP, we focused
our analyses on the promoter, and exons 2 and 4. We checked,
nevertheless, that exon 3 did not contain new SNPs in 30 un-
related individuals (data not shown). To identify NCR3 muta-
tions in the promoter region, exon 2 and exon 4, we performed
sequencing analysis of three defined PCR products. Three
primer pairs were designed with the PRIMER program
(NCR3Frag1: 50-GATGGGTCTGGGTACTGGTG-30and 50-
GGGATCTGAGCAGTGAGGTC-30; NCR3Frag2: 50-ATCC
GCTGA-30; NCR3Frag3: 50-CTGAACTTTCCCTTCCACC
A-30and 50-GGTCCAGCCAGTAAAAACCA-30). PCR ampli-
fication was carried out with AmpliTaq (PE Biosystems) in
50 ml reactions, using a Hybaid apparatus. Thermocycling
started with a single denaturation step for 5 min at 95?C,
following 40 cycles of denaturation for 40 sec at 95?C,
annealing for 30 sec (64?C for NCR3Frag1 primers, 60?C
for NCR3Frag2 primers, and 64?C for NCR3Frag3 primers)
and extension at 72?C for 10 sec. One final extension step
was added for 10 min at 72?C. Before starting the sequencing
reaction, the PCR products were purified with the Qiagen
QIAquick PCR purification kit and quantified by 2% agarose
gel electrophoresis. Sequencing reaction was performed with
the CEQ 8000 kit and a CEQ 8000 automated fluorescent
N.F. Delahaye et al. / Microbes and Infection 9 (2007) 160e166
sequencer (Beckman Coulter). LTA 252 mutation was detected
by PCR-RFLP after NcoI digestion of PCR products as de-
scribed  (50-CCCGTGCTTCGTGCTTTGGACTA-30and
2.4. Allele frequencies, haplotype
reconstruction and LD analysis
All genotypes passed a Mendelian check with the program
FBAT . Using MERLIN program, we further searched for
improbable recombination events from SNP maps to detect
genotyping errors . The detectable genotype errors in the
sample were less than 0.1%. Allele frequencies were calcu-
lated by gene-counting. No deviation from HardyeWeinberg
equilibrium was detected by using a c2with 1 df.
We generated haplotypes on the basis of family genotypic
data with GENEHUNTER program . Pair-wise LD was
calculated with GOLD program . We tested LD between
pairs of biallelic markers by the r2statistic .
2.5. Statistical analyses
Genotype-binary trait (mild malaria attack) correlations
were first assessed using contingency-table analysis. This
was performed for the two categories (affected versus unaf-
fected) of mild malaria attacks, with a focus on the presence
versus the absence of mutation. The c2test and Fisher’s exact
test were used. Since age and hemoglobin genotype are known
to influence the development of malaria attacks [4,13], step-
wise multivariate logistic analysis that took into account age
and hemoglobin genotype was then applied on all variables.
We started with all SNPs as covariates, and we eliminated
non-significant covariates through the likelihood ratio crite-
rion. The goodness-of-fit of the model was tested by the Hos-
mereLemeshow statistic: a significant test indicated that the
model poorly fitted the data. The c2test, Fisher’s exact test,
and logistic regression were performed with the SPSS soft-
ware. Only terms significant at the 5% level were retained.
Association in the presence of linkage was assessed using
family-based association tests (FBATs) , which avoid
biases due to population stratification, population heterogene-
ity, or population admixture. PBAT program (the phenotype
option of FBAT) was used to identify the most relevant regres-
sion model that described the phenotypes as a function of co-
variates . We performed either 100,000 permutations to
calculate empirical P values or data analysis allowing pheno-
typic non-normal distribution using FBAT or PBAT programs.
By use of the Bonferroni correction, the significance level
of the FBAT statistic was adjusted for the number of FBATs
computed (12 tests corresponding to 6 markers and 2 pheno-
types analysed), that is, an adjusted significance level of
P ¼ 0.004. Since the Bonferroni correction procedure does
not take into account allelic association, we also used
a FBAT multi-marker test estimating the covariance between
markers and circumventing multiple testing problems. In this
case, the null hypothesis was no linkage or association be-
tween any marker and mild malaria. This procedure does not
require resolving phase. We also performed haplotype
association analysis that requires resolving phase. To tackle
multi-testing in haplotype association analysis, we used the
multi-allelic procedure computed by FBAT program, and we
calculated empirical P values.
3.1. Identification of polymorphisms in NCR3
We focused the studies on three polymorphisms within the
promoter, five within the exons, one within the introns, and
one within the 30untranslated region (UTR) (Fig. 1). All the
variants were previously reported in dbSNP database (http://
www.ncbi.nlm.nih.gov./SNP/) (Table 1). The NCR3 2859
and NCR3 3571 polymorphisms were the two known non-syn-
onymous mutations reported in the NCR3 gene. We did not
detect the NCR3 2859, NCR3 3008 and NCR3 3649
polymorphisms in the study population. The NCR3 ?172
polymorphism was identified in one family. The rare allele
had a frequency <1%, and the power of association analysis
was too small for this polymorphism. Therefore, this polymor-
phism was excluded from further analyses.
3.2. Population-based association of NCR3 ?412
(rs2736191) with mild malaria
To assess the association of polymorphisms with mild
malaria, we first conducted statistical calculations for each
SNP by c2tests with 1 df. According to the distribution of
NCR3 ?412 genotypes GG, GC and CC in affected sibs,
31% of GG sibs, 51% of GC sibs and 57% of CC sibs
presented at least one malaria attack (P1) during the study (Ta-
ble 2). NCR3 ?412C carriage was associated with a higher
risk of mild malaria attacks (c2¼ 6.46, df ¼ 1, P ¼ 0.011).
No association was detected with the other SNPs. Since age
and hemoglobin genotype are known to influence the develop-
ment of malaria attacks [4,13], we performed logistic regres-
sion analysis to re-evaluate the association between NCR3
?412 polymorphism and mild malaria attack. The logistic
regression analysis that took into account the influence of
age and hemoglobin genotype confirmed this association
(c2¼ 10.7, df ¼ 1, P ¼ 0.001). The result of the Hosmere
Lemeshow test indicated that the model fitted the data. The
odds of malaria attack between NCR3 ?412GG and NCR3
?412GC/CC was 2.93 (95% confidence interval 1.49e5.76).
3.3. Family-based association of NCR3 ?412
with mild malaria
Family-based association tests were performed for each
SNP. NCR3 ?412 polymorphism demonstrated nominal asso-
ciation, using the standard quantitative FBAT statistic ,
with the binary trait P1 (P ¼ 0.024) and the quantitative trait
P2 related to mild malaria attacks (P ¼ 0.00092) (Table 3).
In contrast, the NCR3 ?412 polymorphism was not associated
with maximum parasitemia (P3), while NCR3 2708T was
N.F. Delahaye et al. / Microbes and Infection 9 (2007) 160e166
negatively associated with P3 (P ¼ 0.032). NCR3 ?412C was
positively associated with P1 and P2. These two associations
between the NCR3 ?412 polymorphism and the phenotypes
P1 and P2 were confirmed with the PBAT program , using
the GFBAT statistic which is adjusted for environmental
correlation within families  (Table 3). Nevertheless, the
association of the binary trait P1 with the NCR3 ?412
polymorphism was not significant when it was corrected for
multiple testing. In contrast, the association of the quantitative
trait P2 with the NCR3 ?412 polymorphism was robust even
with a conservative Bonferroni correction of the significance
level, and with a multi-marker correction procedure imple-
mented in FBAT.
3.4. LD analysis, haplotype analysis, and family-based
association model including TNF polymorphisms as
covariates: the association of NCR3 ?412 with mild
malaria was not due to the association of
TNF with mild malaria
To test whether the association of NCR3 polymorphisms
with traits related to mild malaria could be due to the associ-
ation of neighbouring polymorphisms with the same traits, we
generated haplotypes using NCR3, TNF and LTA genotypes,
and we calculated pair-wise LD coefficients (Table 4). TNF
1304 and TNF ?238 that were associated with P3  were
in LD with NCR3 2708. TNF 267 and TNF ?244 that were
not associated with mild malaria  were in slight LD with
NCR3 ?412. No TNF variant associated with mild malaria
 was in LD with NCR3 ?412. In particular, NCR3 ?412
was not in LD with TNF 1304 and TNF ?308. Besides,
NCR3 ?412 was in strong LD with LTA 252 (r2¼ 0.366,
P < 0.0001), which was not found to be associated with
mild malaria (data not shown). The common haplotype
characterized by NCR3 ?412C and LTA 252C had a frequency
The haplotype association results were consistent with our
results based on SNP-by-SNP association. An haplotype
containing NCR3 2708T was negatively associated with the
quantitative trait P3 (P ¼ 0.033). In addition, we found a pos-
itive association of the quantitative trait P2 with the haplotype
represented by NCR3 ?412C and LTA 252C (P ¼ 0.008).
Multi-allelic procedure that simultaneously analyzes the asso-
ciation of all the haplotypes with P2 also yielded significant
results (P ¼ 0.0005).
To further dissect the pattern of association between NCR3
polymorphisms and traits related to mild malaria, we utilized
conditional power calculation using PBAT . On this basis,
we identified the most statistically significant set of covariates
in the conditional mean model. According to the PBAT screen-
ing procedure, the most relevant set of covariates for the
association analysis with NCR3 ?412 were TNF 1304 and
TNF ?308 polymorphisms associated with mild malaria ,
and the LTA 252 polymorphism which was in strong LD
with NCR3 ?412. When TNF 1304, TNF ?308 and LTA
252 were simultaneously included as covariates in the model,
NCR3 ?412 was strongly associated with mild malaria attack
(P ¼ 0.00008)
significant after Bonferroni correction for multiple testing. In
contrast, when TNF ?238 and TNF 1304 were taken into ac-
count as covariates, we found that the association of NCR3
2708 with P3 was no longer significant (data not shown).
PromoterExon 1Exon 3Exon 2Exon 43’ UTR region
Fig. 1. Location of the NCR3 mutations studied by sequencing.anon-synonymous SNP, UTR untranslated.
Overview of the NCR3 polymorphisms genotyped in the African population
NCBI numberPosition in the geneAllele frequencyd
G > C
T > C
G > A
C > T
G > A
G > A
G > T
G > A
A > G
T > C
aNCR3Frag: NCR3 fragment.
bPosition relative to transcription start site (þ1).
cWild allele > variant allele. The allele with the higher prevalence was considered the wild one.
dThe frequency of the variant allele is shown.
N.F. Delahaye et al. / Microbes and Infection 9 (2007) 160e166
We report here the first genetic association of a NK cell
receptor with malaria. Our results show that NCR3 ?412 poly-
morphism was associated with the risk of developing mild
malaria attack. NCR3 ?412 polymorphism was not, however,
associated with maximum parasitemia, which was based on
the highest parasitemia in each individual during the study.
Similarly, we previously reported that TNF ?308 polymor-
phism was associated with mild malaria, and that it was not
associated with maximum parasitemia . Conversely, we
found that TNF ?238 polymorphism was associated with
maximum parasitemia, and that it was not associated with
mild malaria. Although we also described TNF variants asso-
ciated with both mild malaria and maximum parasitemia, our
data suggest that the mechanisms involved in the control of
malarial infection and disease may partly differ. It is conceiv-
able that some variants may affect mild malaria by influencing
the cytokine production of effector cells, and that these
variants may only weakly influence parasitemia.
NCR3 ?412 polymorphism may influence the expression
of NKp30 on the surface of NK cells and may alter the re-
sponse of NK cells to P. falciparum-infected erythrocytes. In
this way, the GGTCCT sequence containing NCR3 ?412
polymorphic site is a RREB1/LZ321 binding motif .
This cis-acting element stimulates transcription independent
of the direction in which it is inserted. In addition, it is thought
to be involved in the Ras-Raf signalling pathway , which
participates in the regulation of NK cell activation . It
should be stressed, however, that the biological role of
NKp30 and NCR3 ?412 polymorphism remains to be sup-
ported by direct functional evidence. We cannot exclude that
NCR3 ?412 polymorphism may be in LD with a causative
polymorphism. In particular, a predisposing allele at an
Clinical data of sibs for NCR3 ?412 polymorphism
aSibs with at least one malaria attack during the 24 months of the study.
Family-based association tests for NCR3 ?412 polymorphism
Malaria attack (P1)
Risk of developing
malaria attacks (P2)
aFBAT is the family-based association test.
bGFBAT is the FBAT statistic which is adjusted for environmental
correlation within families.
cSNPs in covariates: TNF ?308, TNF 1304 and LTA 252.
Table 4 LD coefficients for the locus NCR3-TNF-LTA
*, Significant result, P < 0.05. **, Significant result, P < 0.0001.
N.F. Delahaye et al. / Microbes and Infection 9 (2007) 160e166