Bediako et al. BMC Medicine           (2019) 17:60 
https://doi.org/10.1186/s12916-019-1292-y
RESEARCH ARTICLE Open Access
Repeated clinical malaria episodes are
associated with modification of the
immune system in children
Yaw Bediako1†, Rhys Adams1†, Adam J. Reid2†, John Joseph Valletta3†, Francis M. Ndungu4†, Jan Sodenkamp1,7,
Jedidah Mwacharo4, Joyce Mwongeli Ngoi4,8, Domtila Kimani4, Oscar Kai4, Juliana Wambua4, George Nyangweso4,
Etienne P. de Villiers4,6, Mandy Sanders2, Magda Ewa Lotkowska2, Jing-Wen Lin1,9, Sarah Manni1, John W. G. Addy1,
Mario Recker3, Chris Newbold2,5, Matthew Berriman2, Philip Bejon4, Kevin Marsh6 and Jean Langhorne1*
Abstract
Background: There are over 200 million reported cases of malaria each year, and most children living in endemic
areas will experience multiple episodes of clinical disease before puberty. We set out to understand how frequent
clinical malaria, which elicits a strong inflammatory response, affects the immune system and whether these
modifications are observable in the absence of detectable parasitaemia.
Methods: We used a multi-dimensional approach comprising whole blood transcriptomic, cellular and plasma
cytokine analyses on a cohort of children living with endemic malaria, but uninfected at sampling, who had
been under active surveillance for malaria for 8 years. Children were categorised into two groups depending
on the cumulative number of episodes experienced: high (≥ 8) or low (< 5).
Results: We observe that multiple episodes of malaria are associated with modification of the immune system. Children
who had experienced a large number of episodes demonstrated upregulation of interferon-inducible genes, a clear
increase in circulating levels of the immunoregulatory cytokine IL-10 and enhanced activation of neutrophils, B cells and
CD8+ T cells.
Conclusion: Transcriptomic analysis together with cytokine and immune cell profiling of peripheral blood can robustly
detect immune differences between children with different numbers of prior malaria episodes. Multiple episodes of
malaria are associated with modification of the immune system in children. Such immune modifications may have
implications for the initiation of subsequent immune responses and the induction of vaccine-mediated protection.
Keywords: Malaria, Systems immunology, Immune activation
Background response in the host. Initial exposure often leads to dis-
Malaria is caused by infection with the protozoan parasite ease, but subsequent repeated exposures lead to the devel-
Plasmodium spp. and is responsible for approximately half opment of partially protective, non-sterile immunity [3–
a million deaths annually. Most of the mortality occurs 5]. There is mounting evidence that repeated clinical epi-
among children under 5 years of age [1], and progress in sodes of malaria result in substantial modification of the
control has recently stalled [2]. Malaria pathogenesis is host immune system. P. falciparum (Pf ) infection has
characterised by a complex interplay between an antigeni- been shown to stimulate T regulatory cells [6, 7] and to
cally diverse parasite and a constantly evolving immune significantly alter the phenotype and function of a number
of other immune cell populations including dendritic cells
* Correspondence: jean.langhorne@crick.ac.uk [8], conventional B [9, 10] and T lymphocytes [11, 12] and
Yaw Bediako, Rhys Adams, Adam J. Reid, John Joseph Valletta and Francis M. γδ T cells [13]. In line with this, some Pf proteins bind the
Ndungu are joint first authors inhibitory receptor LILRB1 found on NK and B cells [14].
1Francis Crick Institute, London, UK
Full list of author information is available at the end of the article
© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Bediako et al. BMC Medicine           (2019) 17:60 Page 2 of 14
The consequences of such immune modification have on their number of past clinical episodes. An additional
not been studied extensively; however, it is interesting to 27 age-matched children who had never had clinical mal-
note that a number of vaccine candidates have aria (naïve) were selected from Ngerenya (under active
demonstrated much-reduced efficacy when tested in surveillance since 1989), where malaria transmission has
malaria-endemic populations as compared to malaria-naïve remained very low since 2004. The low group consisted of
populations [15, 16]. Although the precise mechanism of children from Junju who had less than 5 recorded epi-
this is not fully understood, it suggests that complex inter- sodes of malaria, while the high group (also selected from
actions between malaria and the immune system affect the Junju) had between 8 and 18 recorded episodes of malaria.
ability to elicit appropriate immune responses upon chal- A single blood sample was taken from each child and
lenge. Whether such immune modification persists in the processed as described below. All 69 children were geno-
absence of parasitaemia (steady state) is also not known. typed to confirm that none carried the sickle cell trait
Here, we examined healthy uninfected children living (haemoglobin AS genotype), a well-characterised poly-
in an endemic area who had been under active surveil- morphism associated with resistance to malaria infection
lance for clinical malaria for 8 years and had experienced [21]. All 69 children were also determined to be negative
either high or low numbers of clinical episodes (relative for Pf (microscopy and PCR) and had not had a clinical
to the population average). We took a multi-dimensional episode within the last 110 days prior to sampling.
approach, comprising whole blood transcriptomic, cellu-
lar and plasma cytokine analyses to describe the immune Sample collection
systems in these two groups of children, providing a Five millilitres of blood was drawn from each child by
comprehensive description of the effect of repeated epi- venesection in March 2015 prior to the start of the
sodes of clinical malaria on the steady-state immune sys- major malaria transmission season. One millilitre was
tem of children living in an endemic area. While immediately placed in a Tempus tube (Thermo Fisher
insufficient to establish the causal relationship between Scientific) and stored for downstream transcriptomic
malaria episodes and any immune modification (differ- analysis. The remaining blood was transported within 2
ences could reflect inherent immunological differences h of collection to the laboratory where 200 μL was ali-
that predispose certain individuals to increased numbers quoted for flow cytometry and 100 μL aliquoted for
of episodes), this study represents a necessary first step real-time PCR (to assess Pf status), and the remaining
in furthering our understanding of the complexity of sample was centrifuged to separate the plasma which
malaria immune responses. was stored at − 20 °C.
Materials and methods PCR analysis
Study population For PCR analysis, DNA was first extracted from 30 μL of
The participants for this study were drawn from two whole blood using QIAxtractor machine (QIAGEN,
previously described cohorts of children who had been Hilden, Germany). The DNA was eluted in 100 μL, from
under active weekly surveillance for 8 years [17, 18]. The which 5 μL of DNA were amplified by quantitative PCR.
Junju cohort is in an area of moderate malaria transmis- This was done using a TaqMan assay for the Pf
sion with a Pf prevalence of approximately 30% [15, 17] multicopy 18S ribosomal RNA genes, as described else-
during the rainy season, while the Ngerenya cohort is in where [22], except that we used a modified probe
an area where malaria transmission has fallen and (5′-FAM-AACAATTGGAGGGCAAG-NFQ-MGB-3′).
remained at almost zero since 2004 [18]. As described We used an Applied Biosystems 7500 Real-Time PCR
elsewhere [19, 20], children were visited every week by System with quantification by Applied Biosystems 7500
field workers (themselves living within the local commu- software v2.0.6. Samples were analysed in singlet wells.
nity) for the detection of malaria-associated fevers and Three negative control wells and 7 serial dilutions of
who were also available to assess any fevers occurring DNA extracted from in vitro parasite cultures were in-
between weekly visits. Any child with an axillary body cluded as standards on each plate in triplicate. Plates
temperature of greater than 37.5 °C was tested for Pf failing quality control standards were repeated. The
parasitaemia by rapid diagnostic test and confirmed by lower limit of accurate quantification of this method is
microscopic examination of thin and thick blood smears 10 parasites/mL within the PCR elute. By assessing 1/20
stained with 10% Giemsa. A clinical episode of malaria of 30 μL of blood with a gene target present on 3 chro-
was defined as body temperature above 37.5 °C with > mosomes, the method has a theoretical limitation of 4.5
2500 parasites per microlitre of blood. parasites/μL of whole blood, compared with a sensitivity
For our analysis, 42 children of similar age (7–10.5 of 50 parasites/μL for thick blood films. PCR standards
years) were selected belonging to 2 categories—“low” and were monitored through internal quality assurance and
“high” (under active surveillance since 2007) depending use of external quality control standards.
Bediako et al. BMC Medicine           (2019) 17:60 Page 3 of 14
Stool microscopy Plasma cytokine analysis
The formol-ether concentration method was used to One hundred microlitres of plasma from each partici-
prepare samples for the detection of helminths or their pant was submitted to Eve Technologies (Calgary,
eggs by microscopy. Canada) for analysis using the Human Cytokine/Chemo-
kine 65-plex Discovery Assay. This multiplex assay is
based on the Millipore MILLIPLEX cytokine array and
Flow cytometry is designed to detect and quantify the levels of the fol-
Two hundred microlitres of whole blood was mixed with lowing cytokines: EGF, eotaxin, FGF-2, Flt-3 ligand, frac-
a cocktail of monoclonal antibodies specific for human talkine, G-CSF, GM-CSF, GRO, IFN-α2, IFN-γ, IL-10,
immune cell surface markers. The cocktail consisted of IL-12 (p40), IL-12 (p70), IL-13, IL-15, IL-17A, IL-1ra,
antibodies against CD3, CD4, CD8, CD14, CD16, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9,
HLA-DR, CD11c, CD45RO, CD45RA, TCR γδ, CD56, IP-10, MCP-1, MCP-3, MDC (CCL22), MIP-1α, MIP-1β,
CD19 and CD303 as well as a live/dead stain (see Add- PDGF-AA, PDGF-AB/BB, RANTES, TGFα, TNF-α,
itional file 1: Table S1 for antibody conjugation informa- TNF-β, VEGF, sCD40L, Eotaxin-2, MCP-2, BCA-1,
tion). After staining for 30 min at 4 °C, erythrocytes were MCP-4, I-309, IL-16, TARC, 6CKine, eotaxin-3, LIF,
lysed using BD FACS Lysing Solution (BD Biosciences, TPO, SCF, TSLP, IL-33, IL-20, IL-21, IL-23, TRAIL,
San Jose, CA). Cells were washed and re-suspended in CTACK, SDF-1α+β, ENA-78, MIP-1d and IL-28A. Cyto-
200 μL of 1× PBS and analysed on a BD Fortessa flow kine levels were parameterised as log fluorescence and
cytometer (BD Biosciences, San Jose, CA) acquiring at tested using a three-way Kruskal-Wallis test between the
least 200,000 leukocyte events per sample. Given the size naive, low-episode and high-episode groups. Post-hoc
of the study and the need to limit time between sample Dunn’s tests were performed on cytokines with signifi-
collection and FACS analysis, sample collection and cant differences.
FACS were performed in batches over a number of days,
with appropriate single-colour controls acquired on each RNA isolation and library preparation
day. All FACS data were however analysed together once Tempus/blood mix (1 mL blood with 6 mL Tempus solu-
all the samples had been collected. Initial compensation tion) was thawed on ice for 1 h and transferred into a
and manual gating analysis were performed using FlowJo 50-mL Falcon tube. Next, 2 mL of ice cold 1×PBS was
(FlowJo LLC, Ashland, OR). added to the samples followed by the addition of 3 mL
chilled 100% ethanol. Samples were immediately vor-
texed for 30 s and then spun down at 15,000 rcf for
Unsupervised FACS analysis 60 min at 0 °C. After centrifugation, the supernatant was
Flow cytometry data was analysed using the integrated removed and the emptied tubes blotted on clean absorb-
analysis pipeline Cytofkit, available as an open-source ent paper to remove the remaining foam. No cell debris
R/Bioconductor package [23]. Briefly, fcs files contain- pellet was visible within the tube. Next, the cells were
ing all live gated, singlet events from each participant lysed by adding 200 μL of freshly prepared lysis/TCEP
were imported, the expression values of each marker solution (Perfect Pure kit, 5’PRIME) to the pellet and
extracted from each fcs file and the extracted data vortexed immediately for 1.5 min. RNA isolation was
transformed using “automatic logicle transformation”. performed using the Perfect Pure kit, following the
Expression matrices from all fcs files were then com- manufacturer’s instructions, and eluted in 40 μL of
bined into a single matrix, by sampling up to 10,000 nuclease-free water. Globin mRNA was depleted from
events from each fcs file. Dimensionality reduction was the total RNA using the GLOBINclear kit (Ambion).
performed using the Barnes-Hut variant of the t-SNE Indexed libraries were then generated using the KAPA
algorithm [24], and cellular subsets were identified Stranded mRNA-Seq Kit (Roche) on an automated plat-
using the clustering method proposed by Rodriguez form with 10 cycles of PCR amplification.
and Laio [25]. Individual clusters were then manually
annotated using a heatmap displaying the median in- RNA sequencing
tensity values per cluster for every marker. This heat- Seventy-five samples, comprising 6 replicates of a single
map was used to identify each cluster’s defining European sample (batch controls), 27 samples from naive
markers and designate each cluster as a previously de- children and 42 samples from exposed children, were se-
scribed population or unknown population. For each quenced in a single multiplexed pool using 5 lanes (75 bp
cellular population, we performed a Kruskal-Wallis test PE) of a HiSeq 2500 (Illumina). The reads were combined
between the three groups of children. For significant across lanes for each sample but not across runs and
cell types, we performed a post-hoc Dunn’s test be- mapped using Kallisto v0.42.3 [26]. As a reference, we
tween each group. used all cDNA sequences from the GRCh38 human
Bediako et al. BMC Medicine           (2019) 17:60 Page 4 of 14
genome. Read counts per gene were calculated by sum- expression of gene i in high and low malaria episode
ming over their transcripts. Genes with fewer than 10 read children. We then performed a Mann-Whitney test of
counts in at least 2 samples were removed. Sequence data response rates for each module between high and low
has been deposited in the European Genome-phenome malaria episode children.
Archive (EGA)—accession number EGAS00001003167.
Cellular deconvolution
Differential expression analysis We performed cellular deconvolution to identify
Differential expression analysis was performed using cell-specific gene expression profiles. We learned the
DESeq2 [27] version 1.16.1. The raw RNA-seq counts are gene expression profiles from the LM22 set of genes
modelled as a negative binomial distribution while expli- previously used to deconvolve cell populations from
citly normalising for library size. p values were adjusted microarray data [35]. To prepare the data for deconvolu-
for multiple comparisons using the Benjamini-Hochberg tion, we manually gated cell populations to mirror those
correction (false discovery rate (FDR)). used to generate the LM22 gene set. Gene expression
was performed on transcripts per million (TPM) as has
Modular analysis been previously advocated for RNA-seq measurements
We applied modular analysis [28, 29] to our RNA-seq [36]. For each gene, we performed deconvolution over
data to ask whether any patterns would distinguish the seven cell types determined manually as illustrated in
low- and high-episode groups of children (see study Additional file 2: Figure S1 (NK, neutrophil, B cell,
population above). We used previously described clus- CD8+ T cell, CD4+ T cell, γδ T cell, monocytes) and
ters (modules) of genes that were co-regulated across three child categories (universal (all samples), not naïve
nine different transcriptomic data sets obtained from pa- (high+low), high). Since deconvolving small populations
tients with a variety of immune conditions [28–32]. could be more error-prone, we limited our analysis to
Analogous to previously described methods [33], we cal- the seven cell categories that were present in a signifi-
culated modular over/underexpression (s) as: cant proportion of the children.
X   For RNA expression of each gene as measured by
¼ 1sM 100 j j D gi a; gi b TPM, y, we fit a profile (t) to the fraction of sub-cellM ; ;i∈M types (F) measured in children. The sub-cell types are
where separated into three distinct categories: universal (U), ð Þ ð Þ not naive (N) and high episodes (H), and arranged into asign μ
(g , g i;a
−μi;b if p gi;a; gi;b < 0:05
i, a i, b)= matrix as F = [F(U), F(N), F(H)]. The universal fraction0 otherwise
F(U), is the fraction of cells measured for each child.
For each gene i within a module M, we performed a The sum fraction of cells for a child was less than 1,
Mann-Whitney test and calculated the p value (p) be- since not all cell events were categorised as a recognis-
tween child groups a and b. Here, M is the set of genes able immune cell. The subsequent terms F(N) and F(H)
in a module, and |M| is the number genes in that mod- are variations of the universal fraction, defined as:
ule. Child categories include naive, low number of epi- 
sodes and high number of episodes. If the test yielded a F ðUÞ if c∈P
p value < 0.05, then the sign of the differences in median Fc;iðPÞ ¼
c;i
0 otherwise
rlog values (μ) were added to sM (sign is a function that
returns − 1 for negative numbers, 0 for 0, and + 1 for where P is a set of children in a category, and c is an in-
positive numbers). The list of genes in the modules was dividual child. We modelled the gene expression as the
obtained from a previously published report [29]. linear set of equations Ft ¼ ŷ . For each gene, we fit a
In a recent study, a modular transcriptional repertoire profile with lasso penalty as:
analysis was used to find markers for malarial immunity
T 1
following an RTS,S study [34]. In contrast to modular argmintðFt−yÞ ðFt−yÞ−λjtj
expression, which describes changes over entire categor- We chose the lasso penalty (λ) that maximised the
ies of children, we also defined modular response (r) for tenfold cross-validated coefficient of determination (i.e.
individuals as: R2) to find non-zero cell-specific profiles. This was im-
1 X  ¼ plemented in Python using scikit-learn [37]. For thisrc;M 100 j sign g −μM j i;c i lasso penalty, we then performed a Bayesian lasso fit toi∈M
obtain z-scores for the non-zero cell-specific profiles.
where rc,M is the response of child c in module M, |M| The model’s parameters were inferred using MCMC
is the number of genes in module M, gi,c is the rlog gene [38]. As further controls, we performed this deconvolu-
expression of gene i in child c, and μi is the median gene tion on simulated data. In one set, child RNA-seq
Bediako et al. BMC Medicine           (2019) 17:60 Page 5 of 14
measurements were scrambled. The resulting number of transmission area of Junju but who had only experienced
positive results was used to estimate false discovery a low number of cumulative clinical episodes. DESeq2 was
rates. used to estimate the group effect size and false discovery
rate (FDR) for all genes. Only small effect sizes were in-
Gene set enrichment analysis ferred, and most had high FDR values (Fig. 1a). Hierarch-
Gene set enrichment analysis [39] was performed on the ical clustering of individuals based on genes with an FDR
list of genes identified as altered in cell-specific signa- < 0.2 also did not separate individuals into distinct epi-
tures following deconvolution using the Molecular Sig- demiological groups (Fig. 1b). This shows that we are un-
natures Database (MSigDB) [40] and queried Gene able to differentiate between naïve and low-episode
Ontology terms, Reactome [41, 42] and KEGG [43]. individuals on the basis of their transcriptome.
Results Differential gene expression analysis differentiates
Characteristics of study population between high and low malaria episode groups
Study participants were drawn from two cohorts of chil- Next, we compared children from the same moderate
dren (Junju and Ngerenya) who had been under active transmission area but who had experienced low or high
surveillance (see the “Materials and methods” section) numbers of episodes over the 8-year study period. Fig-
for 8 years and were selected to fall into 3 categories. ure 1c shows the estimated effect size and FDR for all
Children in the naive group (n = 27) had never had clin- genes. We detected more differentially expressed genes
ical malaria, those in the low group (n = 21) had experi- than in the previous analysis of naïve versus low-episode
enced less than 5 clinical episodes (median = 2) over the individuals, and while we observe modest effect sizes,
8-year period and those in the high group (n = 21) had a there are a small number of genes with low FDR and
history of 8–18 clinical episodes of malaria (median = known immunologically relevant function.
12) (Table 1). There was no discernible difference be- Importantly, hierarchical clustering based on genes
tween the groups in terms of parasitaemia or severity of with an FDR < 0.2 sorted children almost perfectly into
fever during a clinical episode (Additional file 3: Figure their episode categories. We used k-means clustering to
S2). None of the children recruited had experienced a identify four gene expression clusters (Fig. 1d; Add-
clinical episode within the last 110 days (although unsur- itional file 4: Table S2). Separation into more than four
prisingly the groups differed in time to the last episode clusters resulted in additional clusters indistinguishable
and exposure index), and none were parasitaemic at from one of the first four.
time of sampling. In general, there was quite a high level of heterogeneity
within the three groups of children (Additional file 5: Fig-
Differential gene expression analysis cannot differentiate ure S3); however, subtle but detectable differences were
between naïve and low malaria episode groups observed between the low- and high-episode groups.
We compared individuals from the malaria-naïve cohort High-episode individuals appeared to be characterised by
(Ngerenya) to the individuals who lived in the moderate a transcriptional signature suggestive of greater immune
Table 1 Baseline characteristics of the three epidemiological groups
p value
Town Ngerenya Junju
Group Malaria-naïve Low High
n 27 21 21
Number of clinical episodes (median [IQR]) 0 [0, 0] 2 [1, 2] 12 [9, 14] < 0.001
Age (mean (sd)) 8.8 (1.1) 8.8 (0.3) 8.9 (0.3) 0.9
Sex = M (%) 12 (44.4) 14 (66.7) 9 (42.9) 0.2
Exposure index (median [IQR]) 0 [0, 0.02] 0.43 [0.15, 0.57] 0.77 [0.54, 0.81] < 0.001
Days since last episode (median [IQR]) n/a 490 [380, 1254] 167 [127, 259] < 0.001
Stool microscopy (%) 0.4
Ascaris lumbricoides 0 (0.0) 0 (0.0) 2 (9.5)
Hook worm 1 (3.7) 1 (4.8) 0 (0.0)
Trichuris trichiura 1 (3.7) 0 (0.0) 0 (0.0)
No parasites detected 24 (88.9) 20 (95.2) 19 (90.5)
Bediako et al. BMC Medicine           (2019) 17:60 Page 6 of 14
a c
b d
Fig. 1 Differential gene expression analysis distinguishes blocks of genes separating high- and low-episode groups. a DESeq2 was used to compare
the gene expression profiles between naïve and low-episode children. Only small effect sizes were inferred, and most had high FDR values. b
Hierarchical clustering of individuals based on differentially expressed genes also did not separate individuals into distinct epidemiological groups. c
Differential gene expression analysis between high- and low-episode children reveals subtle but detectable differences including a number of genes
with low FDR and known immunological relevant function (highlighted). d We selected 36 gene isoforms with adjusted p values < 0.2 as determined
by differential gene expression analysis (DESeq2) between low- and high-episode children. We used hierarchical clustering to order children but used
k-means clustering to identify 4 subsets of gene expression patterns. Child episode category (high/low) are shown for comparison with gene profiles
activation. These clusters feature genes involved in innate downregulated genes, which demonstrates a strong upreg-
immunity, including defensins (such as DEFA1 and ulation of the “interferon” modules (M1.2, M3.4, M5.12)
DEFA1B) and T cell differentiation (SOX5) as well as in the high-episode group in line with previous work [34].
genes involved in regulating B cell (SIGLEC6) and T cell To examine the variance in modular expression within
responses (PDCD1 and LILRB3). the groups, we quantified the modular responses of indi-
vidual children [34] (Fig. 2d). While there is some het-
Immune modular analysis reveals a unique signature erogeneity within the groups, modules M1.2 and M3.4
associated with high number of clinical episodes (both annotated as “interferon-inducible”) clearly distin-
We applied a modular analysis of immune-related genes guish between the high- and low-episode groups, with
[28, 29, 34] to our RNA-seq data to see whether any pat- both modules expressed more highly in high-episode
terns would distinguish the three groups of children (see children. These modules are enriched for genes involved
the “Materials and methods” section). The overall change in inflammatory responses and both type I and type II
in the expression for each module in each epidemiological interferon signalling. Similar to the DGE analysis, the
group is shown in Fig. 2a–c as the percentage of up- or modular analysis reveals that high-episode children are
Bediako et al. BMC Medicine           (2019) 17:60 Page 7 of 14
Fig. 2 Immune modular analysis reveals a unique signature associated with a high number of clinical episodes. We performed a modular analysis
of a low-episode versus naive, b high-episode versus naive and c high-episode versus low-episode children. For each gene within each of these
previously defined modules, we performed a Mann-Whitney test between different high-, low-episode, or naive children and determined the
number of significant (p < 0.05) upregulated and downregulated genes. The overall change in expression is shown as the percentage of up- or
downregulated genes, which demonstrates a strong upregulation of the “interferon” modules (M1.2, M3.4, M5.12) in the high-episode group. d
For each child and each module, we calculated the “modular response” and then performed a Mann-Whitney test of response rates for each
module between high and low malaria episode children. Using a Benjamini-Hochberg procedure with FDR cut-off of 20%, we identified three
modules that were significantly different between high- and low-episode groups
characterised by a relative upregulation of immune-re- signature, it was reasonable to suspect associations be-
lated genes within the “interferon-inducible” modules. tween the amounts of these plasma cytokines and the
modular responses determined above. As expected,
Plasma IL-10 levels are significantly higher in children plasma levels of IL-10, TNF-α and IL-6, which we found
who have experienced high numbers of clinical episodes to be higher in the high-episode group were significantly
We next compared plasma cytokine and chemokine correlated with the “interferon” modules (M1.2, M3.4,
levels in the study participants at the time of sampling. M5.12) described above (Additional file 7: Figure S4).
Of the 65 cytokines sampled (see the “Materials and These plasma cytokine data therefore provide further
methods” section), 30 were below detectable levels. evidence of enhanced immune activation and inflamma-
Thirty-one of the remaining cytokines were not signifi- tion in the high-episode children and identify elevated
cantly different between the groups (Additional file 6: levels of the immunoregulatory cytokine IL-10 as a key
Table S3). We did however find a trend of increasing difference between children who have experienced high
levels of IL-10, IL-6, TNF-α and CCL15 (MIP-1δ) going and low numbers of episodes.
from the naïve group to the high group, with IL-10
exhibiting the largest effect (Fig. 3). Indeed, IL-10 levels Cellular subset composition reveals expansion in
were significantly higher in children with high numbers activated γδ T cells in high-episode children
of episodes than in low-episode children. Since our In addition to the whole blood transcriptomic ana-
modular analysis had revealed a cytokine-inducible gene lysis, we also characterised the cellular subset
Bediako et al. BMC Medicine           (2019) 17:60 Page 8 of 14
Fig. 3 Differences in the levels of cytokines in plasma of naive and low- and high-episode children. Cytokine levels determined by Luminex cytokine
array were parameterised as log fluorescence and tested using a three-way Kruskal-Wallis test between naive, low-episode and high-episode groups.
Post-hoc Dunn’s tests were performed on cytokines with significant differences. *p = 0.05, **p = 0.01, ***p = 0.005
composition of the blood samples isolated from each populations and excluding unidentifiable populations re-
participant by flow cytometry. Using a 14-colour sulted in 15 identifiable cellular populations (Add-
panel of antibodies against a range of human im- itional file 9: Figure S6). This analysis revealed that
mune cell surface molecules, we determined the cellular subset composition is significantly associated
relative abundance of the major cell populations in with malaria experience. The numbers of CD11c+ B cells
the peripheral blood. (populations 12, 14 and 17), γδ T cells (populations 9,
In addition to classical flow cytometric analysis using 24/10), double-negative T cells (population 25) and den-
manual gating, we maximised the objectivity and dritic cells (population 21) in naïve individuals were sig-
descriptive power of our analysis by using Cytofkit [23] nificantly different from those in either of the other two
to perform an unsupervised analysis of our FACS data. groups (Fig. 4). CD11chi B cells (population 14) in par-
Cellular subsets were identified using a clustering ticular are practically absent from the naive group while
algorithm and individual populations annotated using a found at levels over 104 cells/mL in the other groups.
heatmap displaying the median intensity values per clus- CD11c+ γδ T cells were significantly expanded in the
ter for every marker. This method identified 25 popula- high-episode group and were the only population of
tions (Additional file 8: Figure S5), which were further those we characterised to distinguish between the high-
curated manually. Merging biologically indistinguishable and low-episode groups.
Fig. 4 Differences in cellular subset composition of whole blood from naïve and low- and high-episode children. Cellular composition
was determined via flow cytometry and analysed as described in the “Materials and methods” section. We used a three-way Kruskal-Wallis
test to determine if cell concentrations changed between child categories. We then performed a post-hoc Dunn’s test between individual
groups to determine where significant differences occurred. Dendritic cells (population 21) and CD11c+ B cells (populations 12, 14 and
17) were clearly able to distinguish between naive and malaria-experienced children. However, we observed only subtle differences
between low- and high-episode children with only γδ T cells (merged population 24/10) differing between the two groups. *p = 0.05,
**p = 0.01, ***p = 0.005
Bediako et al. BMC Medicine           (2019) 17:60 Page 9 of 14
Children with high numbers of episodes show differences in gene expression between children who have
transcriptionally altered CD8+ T cells, B cells and experienced a high number of episodes compared with
neutrophils others who have experienced fewer episodes. High-epi-
A major caveat of whole blood transcriptomic analysis is sode children were associated with increased expression
that observed differences in transcript levels might repre- of genes involved in immune activation and regulation,
sent changes in the abundance of certain cellular popula- with modular analysis revealing the enrichment in genes
tions but not necessarily changes in gene expression involved in responses to type I and II interferons. The
within individual cell populations [44]. We therefore per- transcriptomic signature of enhanced immune activation
formed cellular deconvolutions where cell-specific gene in high-episode children is supported by our findings that
expression profiles were inferred based on FACS measure- levels of IL-10 and numbers of a subset of γδ T cells are
ments of cell proportions and RNA-seq transcript levels. significantly higher in these children compared to
To validate the method, we first demonstrated that RNA low-episode children. Through cellular deconvolution of
expression of canonical lineage-associated markers associ- the transcriptomic data, we found that high-episode chil-
ated well with the inferred cell profiles and found good dren may have transcriptionally altered CD8+ T cells, B
correlations between each inferred subset and its respect- cells and neutrophils.
ive subset associated marker (Fig. 5a). Notably, we observed a modular transcriptional signa-
As cellular deconvolution can be confounded by the ture that differs between high- and low-episode children.
sensitivity of smaller populations to changes in larger High-episode children were characterised by higher ex-
subsets, we limited the analysis to the populations that pression of three modules containing interferon-inducible
accounted for 10% or more of the total cell population. genes. These three modules (M1.2, M3.4 and M5.12) are
Three populations (CD8+ T cells, B cells and neutro- part of the transcriptional signature associated with pro-
phils) demonstrated a transcriptionally altered pheno- tection of malaria-naïve adults following the administra-
type with ten or more altered genes in high-episode tion of the RTS,S malaria vaccine [34]. They have also
children relative to low-episode children (Fig. 5b). The been shown to become sequentially activated in systemic
majority of immunologically relevant genes upregulated lupus erythematosus (SLE) patients [30] and form part of
in high-episode children were associated with CD8+ T the transcriptional signature associated with the trivalent
cells. Gene set enrichment analysis (GSEA) on the sig- influenza vaccine [29]. While module M1.2 is enriched for
nificantly altered genes for these cells showed enrich- genes induced by IFN-α, modules M3.4 and M5.12 are
ment for genes involved in positive activation of capable of also being driven by IFN-β and IFN-γ [30]. This
lymphocytes (including LAT, LCK and CD40), which appears to suggest a role for both type I and type II inter-
could suggest more active CD8+ T cells in high-episode ferons in shaping the immune system within high-episode
children (Additional file 10: Table S4). The B cell decon- individuals. Cellular immunity to malaria is typically
volution profile also features mostly upregulated gene thought to involve IFN-γ produced by Th1 CD4+ T cells;
sets in this group of children. Many of these genes have however, both type I and II interferons have been impli-
been implicated in B cell receptor signalling and regulat- cated in the immune response to malaria. Type I inter-
ing antibody responses (and include TNFRSF13B, ZBTB32 ferons, produced by a number of cell types following
and MSC). B cells from high-episode children also malaria infection [31, 32, 45–47], have been implicated in
expressed higher levels of IGHE (gene encoding IgE). regulating CD4+ T cell responses and promoting the dif-
Neutrophils in high-episode children were associated with ferentiation of IL-10-producing Tr1 cells [48], which are
the upregulation of genes involved in the defence and in- known to be significantly expanded in highly exposed chil-
flammatory responses including OSM and TNFAIP6. Our dren [49]. This immunoregulation is thought to reflect an
deconvolution approach thus suggests that high-episode attempt by the immune system to limit inflammation-in-
children are distinguished by transcriptionally altered duced immunopathology but comes at the cost of limiting
CD8+ T cells, B cells and neutrophils, characterised by the anti-parasite immunity and may interfere with the induc-
upregulation of key immune-related genes. tion of robust vaccine-induced immunity.
Inflammatory innate and adaptive immune responses
Discussion are crucial for parasite clearance; however, these effector
In this multi-dimensional assessment of the association functions can result in significant immunopathology
between repeated malaria infections and immune pheno- without appropriate regulation [50, 51]. IL-10 plays a
type, we combined data from whole blood transcriptomic crucial role in modulating the inflammatory response
analysis, multi-parameter flow cytometry, multiplex during malaria [51], and it is notable that even in
plasma cytokine analysis and active malaria surveillance to non-parasitaemic children, of the 65 cytokines measured
identify the immunological features associated with clin- in plasma, IL-10 was the only cytokine observed at sig-
ical malaria experience. We observed subtle but detectable nificantly different levels between high- and low-episode
Bediako et al. BMC Medicine           (2019) 17:60 Page 10 of 14
Fig. 5 Cell-specific gene expression profiles inferred from cellular deconvolution of transcriptome reveal transcriptionally altered CD8+ T cells,
neutrophils and B cells in high-episode children. a RNA expression of canonical cellular lineage-associated markers strongly correlated with
inferred cell-specific profiles, suggesting that cell-specific gene expression patterns can be successfully deconvolved from cell sub-type proportions
and RNA levels. b Using scrambled data to infer false discovery rate, cell-specific gene contributions were selected. Focusing on cell populations with
at least ten genes altered, high-episode children show transcriptionally altered CD8+ T cells, neutrophils and B cells. Heatmap illustrates the expression
level of cell-specific genes in high-episode children relative to low-episode children
individuals. While a number of different cell types pro- inhibit malaria-specific pro-inflammatory cytokine pro-
duce IL-10, a major source in malaria infections is CD4+ duction [52]. Though we do not address the cellular
T cells that co-produce IFN-γ and IL-10 (Tr1 cells) [49]. source of IL-10 in this study, we found that increased
These cells are prevalent in children living in endemic plasma levels of IL-10 were significantly associated with
areas, and the IL-10 they produce has been shown to increased expression of the “interferon-inducible”
Bediako et al. BMC Medicine           (2019) 17:60 Page 11 of 14
signature in high-episode children, in keeping with the More unexpected was our finding that increased mal-
known role of interferons in inducing the development aria experience results in more activated CD8+ T cells.
of Tr1 cells [48]. CD8+ T cells have clear roles in the immune response to
At the cellular level, a subset of γδ T cells (population pre-erythrocytic stages of infection [64, 65] and have
24/10) was significantly expanded in high- relative to been implicated in mediating pathology in a murine
low-episode children. γδ T cells are activated during model of cerebral malaria [66, 67]. There is evidence
malaria, but their function in anti-malarial immunity re- that CD8+ T cells specific to blood-stage antigens are ac-
mains unclear. These cells have been shown to expand tivated via cross-presentation by dendritic cells [68] and
during acute malaria infection in previously naïve indi- may indirectly promote immunity through secretion of
viduals [53, 54] and can produce inflammatory cytokines IFN-γ [11, 69]. It is interesting to note that a recent
including TNF-α and IFN-γ [55] in addition to being study in Western Kenya has described the expansion of
able to directly kill merozoites in vitro [56, 57]. More re- an unconventional innate-like CD8+ T cell population in
cently, a subset of γδ T cells has been shown to associate children living in an area of high parasite burden [70].
with the protection in irradiated sporozoite vaccination While future studies will be needed to confirm the pres-
[58]. In this study, we found that a specific subset of γδ ence of this cellular subset among our cohort, this study
T cells, expressing CD11c, accumulates in high-episode provides further evidence of transcriptional alteration of
children. While not previously reported in the context of CD8+ T cells in the context of malaria exposure.
malaria, CD11c+ γδ T cells have been described as a In our study, participants were selected on the basis of
highly activated subset with enhanced effector function numbers of preceding episodes of malaria within the
and high migratory potential [59]. past 8 years; however, it is important to note that unsur-
Deconvolution analysis integrating cellular proportions prisingly from an epidemiological standpoint, these chil-
(as determined by flow cytometry) with transcriptomic dren also differ in two other important aspects. Despite
data allowed us to infer altered gene expression profiles the fact that none of the participants in either group had
in neutrophils, CD8+ T cells and B cells in high-episode experienced an episode of malaria for more than 110
children. Very little is known about the role of neutro- days, there was a significant difference in the calculated
phils in malaria although neutrophils isolated from exposure indices and time to the last episode between
Pf-infected children in the Gambia were shown to tem- the low- and high-episode groups. While the modular
porarily exhibit reduced effector function until about 8 transcriptional signature we observed does not appear to
weeks after infection [60]. Our results suggest that re- correlate with time to the last episode (Spearman corre-
peated episodes of malaria result in the development of lations: M1.1 = 0.027, p value = 0.87; M1.2 = − 0.05, p
an activated neutrophil phenotype that persists even in value = 0.76; M3.4 = − 0.08, p value = 0.63; M5.12 = −
the absence of detectable infection. 0.13, p value = 0.43), we cannot discount the possibility
Our finding of high levels of B cell expression of genes that the effects that we observe are due in part at least
including TNF receptor superfamily member 13B to the more recent immunological stimulation in the
(TNFRSF13B), a receptor found on the surface of B cells, high-episode group rather than the number of previous
responsible for regulating humoral responses and sur- episodes per se. Carefully designed longitudinal studies
vival of plasma cells [61], is in line with the studies dem- would be required to disentangle the contributions of
onstrating that repeated exposure to malaria is necessary these and other parameters to the development of a mal-
for the development of appropriate humoral responses aria immune response. These studies could prospectively
[3–5]. Furthermore, our finding that B cells from relate individual pre-existing immunological status to
high-episode children also express high levels of IgE subsequent risk of clinical infection and thus determine
supports previous studies showing increased plasma IgE which immune responses are directly related to clinical
levels in individuals living in high transmission settings protection.
[62, 63]. We also revealed a clear expansion of CD11c+
B cells in malaria-experienced children. Atypical B cells Conclusion
(a population that includes CD11c+ B cells) have previ- In summary, in this exploratory study, we show that our
ously been identified in high frequencies among individ- approaches of transcriptomic analysis together with
uals living in malaria-endemic regions [9]. While these cytokine and immune cell profiling of peripheral blood
cells were completely missing in naïve children, we did can robustly detect immune differences between high-
not observe significant differences in CD11c+ B cell and low-episode children. Multiple (and possibly recent)
numbers between the high- and low-episode groups. episodes of malaria are associated with the modification
This suggests that although malaria most certainly leads of the immune system in children. Individuals who have
to the initial expansion of these cells, they may not accu- experienced repeated episodes demonstrate enhanced
mulate with subsequent episodes. activation of neutrophils, B cells and CD8+ T cells;
Bediako et al. BMC Medicine           (2019) 17:60 Page 12 of 14
upregulation of interferon-inducible genes; and a clear MB and AR are supported by the Wellcome Trust (Grant #: WT 206194)
increase in circulating levels of the immunoregulatory
Availability of data and materials
cytokine IL-10. Such elevated IL-10 levels suggest a de- Transcriptomic data has been deposited in the European Genome-phenome
gree of immune modulation that may be important for Archive (EGA) https://www.ebi.ac.uk/ega/studies/EGAS00001003167. For
avoiding immunopathology but could interfere with metadata please submit a request to the KEMRI Wellcome Trust Data
Governance Committee via Data Requests dgc@kemri-wellcome.org using
parasite clearance. This skewing may also affect the in- the form available on their website: http://kemri-wellcome.org/about-us/
duction of protective immune responses by vaccines and #ChildVerticalTab_15.
hence have significant implications for the efficacy of
Authors’ contributions
such vaccines in endemic populations. JL, KM, PB, MB, CN, MR and FN conceptualised the project. YB, RA, AR, JJV,
MEL, DK, JM and JWGA carried out the experiments and/or analysed the
Additional files data. JMN, JW, EPV, MS, OK, JS, J-WL, SM and GN provided the essential data.
YB, RA, JL, AR and JJV wrote the manuscript. All authors read and approved
the final manuscript.
Additional file 1: Table S1. Flow cytometry antibody panel. (XLSX 9 kb)
Additional file 2: Figure S1. Gating strategy used to define cellular Ethics approval and consent to participate
subsets used in deconvolution analysis. (PDF 425 kb) Informed, written consent was obtained from the parents/guardians of the
research participants prior to enrolment in the study. The study was reviewed
Additional file 3: Figure S2. Distribution of temperature and log- and approved by the Kenyan Medical Research Institute National Ethics
parasitaemia for each clinical episode in the high- and low-episode Committee. (Reference number: SSC 2887)
groups over period of follow-up. High-episode group: red dots, n = 21;
low-episode groups: blue dots, n = 21 (PDF 160 kb) Consent for publication
Additional file 4: Table S2. k-means clustering of differentially Not applicable.
expressed genes between low- and high-episode children. (XLSX 10 kb)
Additional file 5: Figure S3. Principal component analysis (PCA) plot of Competing interests
the transcriptome profiles of study participants. Naïve (green), low episodes The authors declare that they have no competing interest.
(blue) and high episodes (red). (PDF 39 kb)
Additional file 6: Table S3. Mean levels of detectable plasma cytokines. Publisher’s Note
(XLSX 11 kb) Springer Nature remains neutral with regard to jurisdictional claims in published
Additional file 7: Figure S4. Association between immune modular maps and institutional affiliations.
expression and plasma cytokine levels. (a) Spearman correlations
between significant cytokines and modules are shown along with (b) Author details
1 2
their respective p values. (PDF 278 kb) Francis Crick Institute, London, UK. Wellcome Genome Campus, Wellcome
Sanger Institute, Hinxton, Cambridgeshire, UK. 3University of Exeter, Exeter,
Additional file 8: Figure S5. Unsupervised cellular subset identification. UK. 4KEMRI/Wellcome Trust Research Programme, Kilifi, Kenya. 5Weatherall
Flow cytometry data was analysed using the integrated analysis pipeline Institute of Molecular Medicine, University of Oxford, Oxford, UK. 6Nuffield
Cytofkit. (a) Collective t-SNE dimensionality reduced CD45+ live cell data Department of Medicine, University of Oxford, Oxford, UK. 7Present Address:
derived from 69 participants. Every dot represents a single cell, and the Transla TUM, Zentralinstitut für translationale Krebsforschung der
colour of the cells indicates the expression values for a given marker Technischen Universität München, Munich, Germany. 8Present Address: West
analysed. (b) Cellular subsets were identified using Cluster X. (c) Heatmap African Centre for Cell Biology of Infectious Pathogens, University of Ghana,
displaying hierarchical clustering of median surface marker expression Accra, Ghana. 9Present Address: Division of Pediatric Infectious Diseases, State
levels of indicated populations. Bracketed clusters were condensed into Key Laboratory of Biotherapy, Sichuan University and Collaboration
one population. (Populations 13, 7, 18, 19 and 15 determined to be Innovation Centre, Chengdu, China.
unidentifiable). (PDF 1238 kb)
Additional file 9: Figure S6. Cellular composition of whole blood from Received: 19 November 2018 Accepted: 18 February 2019
naïve and low- and high-episode children. The initial clusters in Add-
itional file 8: Figure S5 were manually curated, merging biologically indis-
tinguishable clusters resulting in 15 identifiable cellular populations. We References
used a 3-way Kruskal-Wallis test to determine if cell concentrations chan- 1. World Health Organization. World malaria report. Geneva: World Health
ged between child categories. We then performed a post-hoc Dunn’s test Organization; 2017.
between individual groups to determine where significant differences oc- 2. Snow RW, Sartorius B, Kyalo D, Maina J, Amratia P, Mundia CW, Bejon P,
curred. *p = 0.05, **p = 0.01, ***p = 0.005. (PDF 781 kb) Noor AM. The prevalence of Plasmodium falciparum in sub-Saharan Africa
Additional file 10: Table S4. Results of GSEA analysis of CD8+ T cell since 1900. Nature. 2017;550(7677):515–8.
deconvolution signature associated with high-episode children. (XLSX 13 kb) 3. Langhorne J, Ndungu FM, Sponaas AM, Marsh K. Immunity to malaria: more
questions than answers. Nat Immunol. 2008;9(7):725–32.
4. Marsh K, Kinyanjui S. Immune effector mechanisms in malaria. Parasite
Acknowledgements Immunol. 2006;28(1–2):51–60.
The authors acknowledge the participants and their parents who graciously 5. Tran TM, Li S, Doumbo S, Doumtabe D, Huang CY, Dia S, Bathily A, Sangala
consented for their samples to be used in this study. J, Kone Y, Traore A, et al. An intensive longitudinal cohort study of Malian
The authors thank the WTSI DNA pipelines for generating the RNA libraries children and adults reveals no evidence of acquired immunity to
and sequencing. Plasmodium falciparum infection. Clin Infect Dis. 2013;57(1):40–7.
6. Walther M, Tongren JE, Andrews L, Korbel D, King E, Fletcher H, Andersen
Funding RF, Bejon P, Thompson F, Dunachie SJ, et al. Upregulation of TGF-beta,
The study received funding from the UK Medical Research Council, (MRC FOXP3, and CD4+CD25+ regulatory T cells correlates with more rapid
Programme grant #: MR/M003906/1) parasite growth in human malaria infection. Immunity. 2005;23(3):287–96.
YB, RA, JL, JWL, SM and JS are supported by the Francis Crick Institute, which 7. Kurup SP, Obeng-Adjei N, Anthony SM, Traore B, Doumbo OK, Butler NS,
receives its funding from the UK Medical Research Council, Cancer Research Crompton PD, Harty JT. Regulatory T cells impede acute and long-term
UK, and the Wellcome Trust, UK. The Wellcome Trust provides core support immunity to blood-stage malaria through CTLA-4. Nat Med. 2017;23(10):
to the Kenya Programme (203077_Z_16_Z) 1220–5.
Bediako et al. BMC Medicine           (2019) 17:60 Page 13 of 14
8. Pinzon-Charry A, Woodberry T, Kienzle V, McPhun V, Minigo G, Lampah DA, exploration reveals quantitative and qualitative differences in response to
Kenangalem E, Engwerda C, Lopez JA, Anstey NM, et al. Apoptosis and influenza and pneumococcal vaccines. Immunity. 2013;38(4):831–44.
dysfunction of blood dendritic cells in patients with falciparum and vivax 30. Chiche L, Jourde-Chiche N, Whalen E, Presnell S, Gersuk V, Dang K,
malaria. J Exp Med. 2013;210(8):1635–46. Anguiano E, Quinn C, Burtey S, Berland Y, et al. Modular transcriptional
9. Weiss GE, Crompton PD, Li S, Walsh LA, Moir S, Traore B, Kayentao K, repertoire analyses of adults with systemic lupus erythematosus reveal
Ongoiba A, Doumbo OK, Pierce SK. Atypical memory B cells are greatly distinct type I and type II interferon signatures. Arthritis Rheumatol. 2014;
expanded in individuals living in a malaria-endemic area. J Immunol. 2009; 66(6):1583–95.
183(3):2176–82. 31. Haque A, Best SE, Ammerdorffer A, Desbarrieres L, de Oca MM, Amante FH,
10. Illingworth J, Butler NS, Roetynck S, Mwacharo J, Pierce SK, Bejon P, de Labastida RF, Hertzog P, Boyle GM, Hill GR, et al. Type I interferons
Crompton PD, Marsh K, Ndungu FM. Chronic exposure to Plasmodium suppress CD4(+) T-cell-dependent parasite control during blood-stage
falciparum is associated with phenotypic evidence of B and T cell Plasmodium infection. Eur J Immunol. 2011;41(9):2688–98.
exhaustion. J Immunol. 2013;190(3):1038–47. 32. Kim CC, Nelson CS, Wilson EB, Hou B, DeFranco AL, DeRisi JL. Splenic red
11. Horne-Debets JM, Faleiro R, Karunarathne DS, Liu XQ, Lineburg KE, Poh pulp macrophages produce type I interferons as early sentinels of malaria
CM, Grotenbreg GM, Hill GR, MacDonald KP, Good MF, et al. PD-1 infection but are dispensable for control. PLoS One. 2012;7(10):e48126.
dependent exhaustion of CD8+ T cells drives chronic malaria. Cell Rep. 33. Chaussabel D, Quinn C, Shen J, Patel P, Glaser C, Baldwin N, Stichweh D,
2013;5(5):1204–13. Blankenship D, Li L, Munagala I, et al. A modular analysis framework for
12. Bediako Y, Ngoi JM, Nyangweso G, Wambua J, Opiyo M, Nduati EW, Bejon blood genomics studies: application to systemic lupus erythematosus.
P, Marsh K, Ndungu FM. The effect of declining exposure on T cell- Immunity. 2008;29(1):150–64.
mediated immunity to Plasmodium falciparum - an epidemiological “natural 34. Rinchai D, Presnell S, Vidal M, Dutta S, Chauhan V, Cavanagh D, Moncunill G,
experiment”. BMC Med. 2016;14(1):143. Dobano C, Chaussabel D. Blood interferon signatures putatively link lack of
13. Jagannathan P, Kim CC, Greenhouse B, Nankya F, Bowen K, Eccles-James I, protection conferred by the RTS,S recombinant malaria vaccine to an
Muhindo MK, Arinaitwe E, Tappero JW, Kamya MR, et al. Loss and antigen-specific IgE response. F1000Res. 2015;4:919.
dysfunction of Vdelta2(+) γδ T cells are associated with clinical tolerance to 35. Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD,
malaria. Sci Transl Med. 2014;6(251):251ra117. Diehn M, Alizadeh AA. Robust enumeration of cell subsets from tissue
14. Saito F, Hirayasu K, Satoh T, Wang CW, Lusingu J, Arimori T, Shida K, expression profiles. Nat Methods. 2015;12(5):453–7.
Palacpac NMQ, Itagaki S, Iwanaga S, et al. Immune evasion of Plasmodium 36. Jin H, Wan YW, Liu Z. Comprehensive evaluation of RNA-seq quantification
falciparum by RIFIN via inhibitory receptors. Nature. 2017;552(7683):101–5. methods for linearity. BMC Bioinformatics. 2017;18(Suppl 4):117.
15. Bejon P, Mwacharo J, Kai O, Todryk S, Keating S, Lowe B, Lang T, Mwangi 37. Fabian P, Gael V, Alexandre G, Vincent M, Bertrand T, Olivier G, Mathieu B,
TW, Gilbert SC, Peshu N, et al. The induction and persistence of T cell IFN- Peter P, Ron W, Vincent D, et al. Scikit-learn: machine learning in Python. J
gamma responses after vaccination or natural exposure is suppressed by Mach Learn Res. 2011;12:2825–30.
Plasmodium falciparum. J Immunol. 2007;179(6):4193–201. 38. Forman-Mackey DH, David, Lang D, Goodman J. emcee: the MCMC
16. Rts SCTP, Agnandji ST, Lell B, Fernandes JF, Abossolo BP, Methogo BG, Hammer. Publ Astron Soc Pac. 2013;125(925):306.
Kabwende AL, Adegnika AA, Mordmuller B, Issifou S, et al. A phase 3 trial of 39. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA,
RTS,S/AS01 malaria vaccine in African infants. N Engl J Med. 2012;367(24): Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al. Gene set enrichment
2284–95. analysis: a knowledge-based approach for interpreting genome-wide
17. Mbogo CM, Mwangangi JM, Nzovu J, Gu W, Yan G, Gunter JT, Swalm C, expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50.
Keating J, Regens JL, Shililu JI, et al. Spatial and temporal heterogeneity of 40. Liberzon A, Subramanian A, Pinchback R, Thorvaldsdottir H, Tamayo P,
Anopheles mosquitoes and Plasmodium falciparum transmission along the Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics.
Kenyan coast. Am J Trop Med Hyg. 2003;68(6):734–42. 2011;27(12):1739–40.
18. O’Meara WP, Mwangi TW, Williams TN, McKenzie FE, Snow RW, Marsh K. 41. Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw
Relationship between exposure, clinical malaria, and age in an area of R, Jassal B, Korninger F, May B, et al. The Reactome pathway
changing transmission intensity. Am J Trop Med Hyg. 2008;79(2):185–91. Knowledgebase. Nucleic Acids Res. 2018;46(D1):D649–55.
19. Mwangi TW, Mohammed M, Dayo H, Snow RW, Marsh K. Clinical algorithms 42. Fabregat A, Sidiropoulos K, Garapati P, Gillespie M, Hausmann K, Haw R,
for malaria diagnosis lack utility among people of different age groups. Jassal B, Jupe S, Korninger F, McKay S, et al. The Reactome pathway
Tropical Med Int Health. 2005;10(6):530–6. Knowledgebase. Nucleic Acids Res. 2016;44(D1):D481–7.
20. Mwangi TW, Ross A, Snow RW, Marsh K. Case definitions of clinical malaria 43. Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K. KEGG: new
under different transmission conditions in Kilifi District, Kenya. J Infect Dis. perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res.
2005;191(11):1932–9. 2017;45(D1):D353–61.
21. Aidoo M, Terlouw DJ, Kolczak MS, McElroy PD, ter Kuile FO, Kariuki S, Nahlen 44. Shen-Orr SS, Gaujoux R. Computational deconvolution: extracting cell type-
BL, Lal AA, Udhayakumar V. Protective effects of the sickle cell gene against specific information from heterogeneous samples. Curr Opin Immunol.
malaria morbidity and mortality. Lancet. 2002;359(9314):1311–2. 2013;25(5):571–8.
22. Hermsen CC, Telgt DS, Linders EH, van de Locht LA, Eling WM, Mensink EJ, 45. Luty AJ, Perkins DJ, Lell B, Schmidt-Ott R, Lehman LG, Luckner D, Greve B,
Sauerwein RW. Detection of Plasmodium falciparum malaria parasites in Matousek P, Herbich K, Schmid D, et al. Low interleukin-12 activity in severe
vivo by real-time quantitative PCR. Mol Biochem Parasitol. 2001;118(2): Plasmodium falciparum malaria. Infect Immun. 2000;68(7):3909–15.
247–51. 46. Sharma S, DeOliveira RB, Kalantari P, Parroche P, Goutagny N, Jiang Z, Chan
23. Chen H, Lau MC, Wong MT, Newell EW, Poidinger M, Chen J. Cytofkit: a J, Bartholomeu DC, Lauw F, Hall JP, et al. Innate immune recognition of an
bioconductor package for an integrated mass cytometry data analysis AT-rich stem-loop DNA motif in the Plasmodium falciparum genome.
pipeline. PLoS Comput Biol. 2016;12(9):e1005112. Immunity. 2011;35(2):194–207.
24. Maaten vd: Accelerating t-SNE using tree-based algorithms. J Mach Learn 47. Wu J, Tian L, Yu X, Pattaradilokrat S, Li J, Wang M, Yu W, Qi Y, Zeituni AE,
Res 2014, 15:3221–3245. Nair SC, et al. Strain-specific innate immune signaling pathways determine
25. Rodriguez A, Laio A. Machine learning. Clustering by fast search and find of malaria parasitemia dynamics and host mortality. Proc Natl Acad Sci U S A.
density peaks. Science. 2014;344(6191):1492–6. 2014;111(4):E511–20.
26. Bray NL, Pimentel H, Melsted P, Pachter L. Near-optimal probabilistic RNA- 48. Zander RA, Guthmiller JJ, Graham AC, Pope RL, Burke BE, Carr DJ, Butler NS.
seq quantification. Nat Biotechnol. 2016;34(5):525–7. Type I interferons induce T regulatory 1 responses and restrict humoral
27. Love MI, Huber W, Anders S. Moderated estimation of fold change and immunity during experimental malaria. PLoS Pathog. 2016;12(10):e1005945.
dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. 49. Jagannathan P, Eccles-James I, Bowen K, Nankya F, Auma A, Wamala S,
28. Chaussabel D, Baldwin N. Democratizing systems immunology with Ebusu C, Muhindo MK, Arinaitwe E, Briggs J, et al. IFNgamma/IL-10 co-
modular transcriptional repertoire analyses. Nat Rev Immunol. 2014;14(4): producing cells dominate the CD4 response to malaria in highly exposed
271–80. children. PLoS Pathog. 2014;10(1):e1003864.
29. Obermoser G, Presnell S, Domico K, Xu H, Wang Y, Anguiano E, Thompson- 50. Lyke KE, Burges R, Cissoko Y, Sangare L, Dao M, Diarra I, Kone A, Harley R,
Snipes L, Ranganathan R, Zeitner B, Bjork A, et al. Systems scale interactive Plowe CV, Doumbo OK, et al. Serum levels of the proinflammatory cytokines
Bediako et al. BMC Medicine           (2019) 17:60 Page 14 of 14
interleukin-1 beta (IL-1beta), IL-6, IL-8, IL-10, tumor necrosis factor alpha, and 69. Horne-Debets JM, Karunarathne DS, Faleiro RJ, Poh CM, Renia L, Wykes MN.
IL-12(p70) in Malian children with severe Plasmodium falciparum malaria Mice lacking programmed cell death-1 show a role for CD8(+) T cells in
and matched uncomplicated malaria or healthy controls. Infect Immun. long-term immunity against blood-stage malaria. Sci Rep. 2016;6:26210.
2004;72(10):5630–7. 70. Falanga YT, Frascoli M, Kaymaz Y, Forconi C, Ong’echa JM, Bailey JA, Berg LJ,
51. Freitas do Rosario AP, Langhorne J. T cell-derived IL-10 and its impact on Moormann AM. High pathogen burden in childhood promotes the
the regulation of host responses during malaria. Int J Parasitol. 2012;42(6): development of unconventional innate-like CD8+ T cells. JCI Insight. 2017;
549–55. 2(15):e93814.
52. Portugal S, Moebius J, Skinner J, Doumbo S, Doumtabe D, Kone Y, Dia S,
Kanakabandi K, Sturdevant DE, Virtaneva K, et al. Exposure-dependent
control of malaria-induced inflammation in children. PLoS Pathog. 2014;
10(4):e1004079.
53. Roussilhon C, Agrapart M, Ballet JJ, Bensussan A. T lymphocytes bearing the
gamma delta T cell receptor in patients with acute Plasmodium falciparum
malaria. J Infect Dis. 1990;162(1):283–5.
54. Teirlinck AC, McCall MB, Roestenberg M, Scholzen A, Woestenenk R, de Mast
Q, van der Ven AJ, Hermsen CC, Luty AJ, Sauerwein RW. Longevity and
composition of cellular immune responses following experimental
Plasmodium falciparum malaria infection in humans. PLoS Pathog. 2011;
7(12):e1002389.
55. Troye-Blomberg M, Worku S, Tangteerawatana P, Jamshaid R, Soderstrom K,
Elghazali G, Moretta L, Hammarstrom M, Mincheva-Nilsson L. Human
gamma delta T cells that inhibit the in vitro growth of the asexual blood
stages of the Plasmodium falciparum parasite express cytolytic and
proinflammatory molecules. Scand J Immunol. 1999;50(6):642–50.
56. Elloso MM, van der Heyde HC, Vande Waa JA, Manning DD, Weidanz WP.
Inhibition of Plasmodium falciparum in vitro by human gamma delta T cells.
J Immunol. 1994;153(3):1187–94.
57. Hviid L, Kurtzhals JA, Adabayeri V, Loizon S, Kemp K, Goka BQ, Lim A,
Mercereau-Puijalon O, Akanmori BD, Behr C. Perturbation and
proinflammatory type activation of V delta 1(+) gamma delta T cells in
African children with Plasmodium falciparum malaria. Infect Immun. 2001;
69(5):3190–6.
58. Zaidi I, Diallo H, Conteh S, Robbins Y, Kolasny J, Orr-Gonzalez S, Carter D,
Butler B, Lambert L, Brickley E, et al. gammadelta T cells are required for the
induction of sterile immunity during irradiated sporozoite vaccinations. J
Immunol. 2017;199(11):3781–8.
59. Qualai J, Li LX, Cantero J, Tarrats A, Fernandez MA, Sumoy L, Rodolosse A,
McSorley SJ, Genesca M. Expression of CD11c is associated with
unconventional activated T cell subsets with high migratory potential. PLoS
One. 2016;11(4):e0154253.
60. Cunnington AJ, Njie M, Correa S, Takem EN, Riley EM, Walther M.
Prolonged neutrophil dysfunction after Plasmodium falciparum malaria
is related to hemolysis and heme oxygenase-1 induction. J Immunol.
2012;189(11):5336–46.
61. Ou X, Xu S, Lam KP. Deficiency in TNFRSF13B (TACI) expands T-follicular
helper and germinal center B cells via increased ICOS-ligand expression but
impairs plasma cell survival. Proc Natl Acad Sci U S A. 2012;109(38):15401–6.
62. Seka-Seka J, Brouh Y, Yapo-Crezoit AC, Atseye NH. The role of serum
immunoglobulin E in the pathogenesis of Plasmodium falciparum malaria
in Ivorian children. Scand J Immunol. 2004;59(2):228–30.
63. Elghazali G, Perlmann H, Rutta AS, Perlmann P, Troye-Blomberg M. Elevated
plasma levels of IgE in Plasmodium falciparum-primed individuals reflect an
increased ratio of IL-4 to interferon-gamma (IFN-gamma)-producing cells.
Clin Exp Immunol. 1997;109(1):84–9.
64. Doolan DL, Hedstrom RC, Wang R, Sedegah M, Scheller LF, Hobart P,
Norman JA, Hoffman SL. DNA vaccines for malaria: the past, the present, &
the future. Indian J Med Res. 1997;106:109–19.
65. Ewer KJ, O’Hara GA, Duncan CJ, Collins KA, Sheehy SH, Reyes-Sandoval A,
Goodman AL, Edwards NJ, Elias SC, Halstead FD, et al. Protective CD8+ T-
cell immunity to human malaria induced by chimpanzee adenovirus-MVA
immunisation. Nat Commun. 2013;4:2836.
66. Howland SW, Claser C, Poh CM, Gun SY, Renia L. Pathogenic CD8+ T cells in
experimental cerebral malaria. Semin Immunopathol. 2015;37(3):221–31.
67. Howland SW, Poh CM, Gun SY, Claser C, Malleret B, Shastri N, Ginhoux F,
Grotenbreg GM, Renia L. Brain microvessel cross-presentation is a hallmark
of experimental cerebral malaria. EMBO Mol Med. 2013;5(7):984–99.
68. Lundie RJ, de Koning-Ward TF, Davey GM, Nie CQ, Hansen DS, Lau LS,
Mintern JD, Belz GT, Schofield L, Carbone FR, et al. Blood-stage Plasmodium
infection induces CD8+ T lymphocytes to parasite-expressed antigens,
largely regulated by CD8alpha+ dendritic cells. Proc Natl Acad Sci U S A.
2008;105(38):14509–14.