796
Frequent and Persistent, Asymptomatic Plasmodium falciparum Infections
in African Infants, Characterized by Multilocus Genotyping
S. Franks,1,a K. A. Koram,3 G. E. Wagner,1 K. Tetteh,2
D. McGuinness,1 J. G. Wheeler,2 F. Nkrumah,3
L. Ranford-Cartwright,1,a and E. M. Riley1,2
1
Institute of Cell, Animal, and Population Biology, Division
of Biological Sciences, University of Edinburgh, Edinburgh,
and 2Department of Infectious and Tropical Diseases, London School
of Hygiene and Tropical Medicine, London, United Kingdom;
3
Epidemiology Unit, Noguchi Memorial Institute for Medical
Research, University of Ghana, Legon, Ghana.
To determine the duration and complexity of naturally acquired Plasmodium falciparum
infections in small children, a longitudinal cohort study of 143 newborns was conducted in
coastal Ghana. On average, children experienced 2 episodes of infection in their first 2 years
of life, the median duration of an asymptomatic infection was ! 4 weeks, and estimates of the
mean number of parasite genotypes per infection were 1.15–2.28. Nevertheless, 40% of the
children experienced infections lasting <12 weeks, and both the duration and complexity of
infections increased with age. The longest period of continual infection was 64 weeks, and
the maximum persistence of a single parasite genotype was 40 weeks. Thus, malaria infections
in infants !5 months old tend to be asymptomatic and rapidly cleared; persistent asymptomatic
parasitemia is more common in children 15 months old. The ability of very young children
to clear or control malaria infections indicates the presence of effective innate or immune
antiparasite mechanisms.
Immunity to malaria is acquired after repeated or persistent
infection for a number of years. The age at which functional
antimalarial immunity is achieved depends largely on levels of
malaria endemicity, with immunity developing more quickly in
areas of high transmission [1]. One explanation for the relatively
slow development of immunity is that it may be variant specific—i.e., infection with many different parasite variants (genotypes) may be required to develop immune responses against
most parasite variants circulating in a population. Because
parasite diversity arises from clonal antigenic variation, as well
as allelic polymorphism, the duration of an infection is also an
important parameter.
Although it is well known that malaria infections can persist
Received 20 September 2000; revised 21 November 2000; electronically
published 8 February 2001.
Presented in part: Molecular Approaches to Malaria meeting, Lorne,
Victoria, Australia, February 2000.
Ethical approval for the study was obtained from the Ghanaian Ministry
of Health (DMA-083) and the University of Edinburgh Medical Ethics
Committee, and informed consent was obtained from mothers for enrollment of themselves and their infants in the study.
Financial support: Wellcome Trust (grant 040328). S.F. and L.R.-C. were
funded by the United Kingdom Medical Research Council.
a
Present affiliations: Eilan Bann Trust, Eilan Bann, Kyleakin, Isle of
Skye, Scotland (S.F.); Division of Infection and Immunity, Institute of
Biomedical and Life Sciences, University of Glasgow, Glasgow, Scotland
(L.R.-C.).
Reprints or correspondence: Dr. Eleanor M. Riley, Dept. of Infectious
and Tropical Diseases, London School of Hygiene and Tropical Medicine,
Keppel St., London WC1E 7HT, United Kingdom ([email protected]).
The Journal of Infectious Diseases 2001; 183:796–804
q 2001 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2001/18305-0015$02.00
for many months, or even years, many of the data predate the
molecular era and come from either malaria therapy records
or from patients from nonendemic areas who contracted malaria during visits overseas. The consensus appears to be that
single, untreated malaria infections can persist for ∼18 months
in the absence of exposure to reinfection [2, 3]. Because of the
difficulty of differentiating persistent infection from reinfection,
however, few studies have attempted to determine the duration
of infections in endemic populations. Contamin et al. [4] studied
parasite diversity in African children for >3 months but concentrated on children with clinical episodes of malaria; treatment of these clinical episodes meant that asexual infections
were terminated prematurely, and thus the natural duration of
these infections is unknown. Daubersies et al. [5] have shown
persistent asymptomatic carriage of single-genotype infections
for >2 months, but, as far as we are aware, only one previous
molecular study of long-term parasite persistence and diversity
has been reported. Babiker et al. [6] sampled 75 individuals
living in an area of highly seasonal malaria transmission in
eastern Sudan for a period of 15 months and showed that
asymptomatic infections could persist for >12 months in the
absence of ongoing transmission. Persistent infections were typically seen in children and teenagers rather than adults. Typically, the persistent infection would be supplanted by a novel
genotype when transmission began again during the following
rainy season [6].
Under conditions of continuing malaria transmission, shortterm studies with frequent sampling all have reported rapid
turnover of parasite clones [5, 7, 8]. Identification of persistent
parasite clones under such conditions is extremely difficult and
JID 2001;183 (1 March)
Malaria in African Infants
would require the use of more precise genetic markers than
those employed to date. The standard genotyping procedure
involves polymerase chain reaction (PCR) amplification of
genes that encode highly polymorphic antigens of the asexual
blood stage of the parasite, coupled with comparison of the
size of the amplified product and/or hybridization to sequencespecific oligonucleotide probes. Limitations of this approach
include the facts that genes that are of the same size may not
share the same sequence and that oligoprobes will pick up sequence differences only in short regions of the gene. Levels of
diversity will thus be underestimated, and persistence of single
clones may be overestimated, unless multiple unlinked genetic
markers are used.
We conducted a prospective cohort study of malaria infection
in an area of low-to-moderate, stable Plasmodium falciparum
transmission in coastal Ghana [9, 10]; 143 newborns were followed up by use of monthly blood sampling for up to 2 years.
We found that episodes of asymptomatic malaria infection were
extremely common in infants [9], which offered us the opportunity to explore the diversity and natural duration of malaria
infections in children who were being regularly reexposed to
infection but who were not being treated with antimalarial
drugs. Some infants were more frequently parasitemic than
were others, and we were in interested knowing whether this
was due to repeated reinfection (which would imply exposure
to high levels of transmission) or to persistent infection by 1
or a few parasite genotypes (which would imply a failure to
clear infection). We therefore characterized in detail the malaria
infections in a group of 13 children, each of whom was parasitemic on >4 occasions in the first 18 months of life. To be
sure that chronic infections were due to persistence of single
parasite genotypes, we employed multilocus genotyping (comprising both antigen genes and neutral microsatellite markers)
and sequenced the highly polymorphic msp-2 locus to confirm
that PCR products of the same size represented the same allele.
Materials and Methods
Study area and cohort. The study was conducted in Prampram
(population 8000), ∼50 km east of Accra, in an area of coastal
savanna in southern Ghana. Malaria transmission in the area is
perennial (although with some seasonal variation), stable, and of
low-to-moderate intensity (5–10 infectious bites per person per
year) [11]. P. falciparum accounts for 98% of all detected infections.
After they gave informed consent, 143 mothers and their newborn
infants were recruited into the study. A heparinized heel-prick blood
sample was obtained from each child on the day of delivery, at 2,
4, and 6 weeks of age, and then every 4 weeks. The child’s axillary
temperature was measured, and a health questionnaire was completed by the mother every 2 weeks. If the child was febrile (temperature >37.57C), a blood film was made and was examined immediately so that antimalarial chemotherapy could be instituted,
if necessary. All children with fever and a parasite-positive blood
film were treated with a full course of chloroquine. Because blood
797
films from clinically well children were not read until some weeks
later, asymptomatic infections were not treated. Because free treatment was provided by study staff and was continuously available,
we believe that very few infections were treated without our knowledge. Antimalarial chemoprophylaxis was not available to pregnant
women in Prampram during the study period, and thus transfer of
antimalarials across the placenta, or via breast milk, was estimated
to be minimal. Infants did not receive malaria prophylaxis.
Parasite detection. Giemsa-stained thick blood films were examined by oil immersion microscopy. The number of parasitized
erythrocytes per 300 leukocytes was counted, and the number of
parasites per microliter of blood was calculated on the basis of an
average leukocyte count of 13,000/mL in children !1 year old [12].
Slides were classified as negative only after >1000 leukocytes had
been counted; the minimum parasite density that could reliably be
detected by microscopy was 40 parasites/mL.
To improve the sensitivity of parasite detection, DNA was extracted from the red cell pellet [13] and was amplified by using
primers to the multicopy, subtelomeric 7H8/6 gene of P. falciparum,
as described elsewhere [9]. This method had a resolution of 1 parasite/mL [9], offering 110-fold sensitivity over microscopy alone.
Parasite genotyping. All samples that were positive for P. falciparum were subjected to PCR genotyping of the msp-1 and msp2 genes. The dimorphic type of the 30 end of msp-1 was determined
by nested PCR, using outer primers derived from conserved sequences in blocks 15 and 17 of the gene, 1G: 50-AAG/CGATATT/
CTTAAATTCACG-30 and IH: 50-ATTATTTTCGTTACAAGTAGG-30. There were 40 cycles of denaturing at 957C for 60 s,
annealing at 507C for 45 s, and extension at 727C for 60 s. This
was followed by an inner PCR, using a single 30 inner primer (1C)
derived from a conserved sequence in block 17 and 2 50 primers—one specific for the MAD20 family (1A) and one for the K1
family (1B)—derived from the dimorphic block 16 [14]. MAD20like alleles give a PCR product of 475 bp, whereas K1-like alleles
give a product of 204 bp. To determine the size of the polymorphic
block 2 of msp-1, nested PCR was performed with 2 sets of primers
(outer O1, O2; inner N1, N2) that recognize conserved sequences
in block 1 and 3, as described elsewhere [15, 16].
To determine the dimorphic type of msp-2 and the size of the
repetitive polymorphic region, an outer PCR reaction amplified almost the full length of the gene, using primers in the conserved 50
(primer 13) and 30 (primer 10) sequences, as described elsewhere [17,
18]. To determine the dimorphic type, the outer PCR product was
amplified with a single 30 primer (2F) from the conserved 30 sequence
of msp-2 and 2 50 primers—one specific for the IC1 family (2E) and
one for the FC27 family (2D) [14]. FC27-like alleles give a PCR
product of 306 bp; IC1-like alleles give a product of 269 bp. To
determine the size of the msp-2 polymorphic region, the outer PCR
product was amplified by 50 (S1) and 30 (S4) primers, on the basis of
conserved sequences just internal to primers 10 and 13 [16].
Microsatellite analysis. Four microsatellite loci—chosen because they comprise trinucleotide repeats, give clear banding patterns, and are unlinked within the genome—were chosen from the
224 markers described by Su and Wellems [19]. The markers were
amplified by heminested PCR, as described by Anderson et al. [20],
using fluorescent-labeled primers (MWG Biotech) suitable for detection on an ABI automated sequencer (PE Applied Biosystems).
Allele sizes were scored by using GENESCAN and GENOTYPER
798
Franks et al.
JID 2001;183 (1 March)
Figure 1. Frequency and duration of Plasmodium falciparum malaria infections in Ghanaian infants and children 0–2 years old. A, Distribution
of infected blood samples within the cohort. Samples were collected every 4 weeks from 143 children for >2 years. B, Distribution of episodes
of infection within the cohort of 143 children. C, Duration of episodes of infection. D, Relationship between age of child and duration of malaria
episode, expressed as percentage of episodes of infection lasting 1 4 weeks vs. age of child at onset of infection. l, Percentage of all episodes; m,
percentage of children in which the longest episode they experienced was ! 4, 4–11, 12–20 weeks, and so on.
software (PE Applied Biosystems). Multiple alleles per locus were
scored if the minor peaks gave a fluorescence intensity of >25%
of the major allele. Peaks of !200 fluorescent units were discounted.
DNA sequencing.
Msp-2 polymorphic sequences were reamplified from the 10, 13 outer PCR product by use of primers
designed to include either an EcoR1 (50 primer 2C) or a BamH1
restriction site (30 primers 2A, 2B) [14], to allow cloning and expression of the product (for subsequent studies). PCR products
were run out on a 1.5% agarose/0.5 Tris-borate EDTA buffer gel;
if 11 band was seen (indicating a mixed clone infection), individual
products were excised from the gel and were purified with the PrepA-Gene kit (BioRad). Single-band PCR products were purified with
the High Pure PCR product purification kit. DNA concentration
was estimated by agarose gel electrophoresis, and ∼75 ng DNA,
in a volume of 2–5 mL, was added to a reaction mixture containing
4 mL Dye Terminator Ready mix (PE Applied Biosystems), 1.6
pmol of primers, and distilled, autoclaved H2O, to a total reaction
volume of 10 mL. The sequencing reaction involved 25 cycles of
denaturation at 957C for 30 s, annealing at 507C for 30 s, and
extension at 657C for 4 min. The product was precipitated in 50
mL 100% ethanol and 2 mL 4M CH2COONa (pH 4.8; Sigma, Poole)
at 2207C for 10 min, was washed in 75% ethanol, and was airdried. The pellet was resuspended in formamide/EDTA/dextran
blue loading buffer, was denatured at 957C for 2 min, and was
loaded onto an ABI 373 automated DNA sequencer (PE Applied
Biosystems). Because of a failure to pick up base G in some reactions, DNA sequences were reanalyzed by using Sequence Analysis 2.1.1, were edited using Seqed version 1.0.3 (both PE Applied
Biosystems), and were compiled in Seqlab GCG9 (University of
Wisconsin Genetics Computer Group).
For brevity, complete sequence data are not shown in this article,
but sequences have been deposited in GenBank (accession numbers
AF217004–AF217043). For each child, each unique sequence is designated with the child’s identification number, the dimorphic family
type (A p IC1 or B p FC27 type), and a sequential numeral. Thus,
15A1 is the first IC1-like sequence isolated from child 15, and 62B1
is the first FC27-like sequence isolated from child 62.
Sickle cell typing. Carriage of sickle cell mutations was determined by hemoglobin electrophoresis, as described elsewhere [9].
Results
P. falciparum infection rates in children. The number of
parasite-positive samples per child showed an overdispersed
distribution (figure 1A). The median (25th and 75th percentiles)
number of parasite-positive samples per child was 3 (1 and 8).
Parasite densities ranged from ! 40/mL (i.e., detectable by PCR
JID 2001;183 (1 March)
Malaria in African Infants
but not by microscopy) to 12 million parasites/mL (median [25th
and 75th percentiles], 480 [43 and 3900] parasites/mL). There
was no significant difference in parasite prevalence between
children who carried the sickle cell trait (AS, AC, SS, or CC)
and children without the sickle cell trait (AA). Twenty-five percent (484 of 1953) of samples from AA children were parasite
positive by either microscopy or PCR, whereas, in sickle cell
trait carriers, 25.7% (223 of 867) samples were infected.
Fifty-six percent (121 of 215) of febrile episodes were associated with some degree of parasitemia (detected by PCR or
microscopy). By use of definitions of clinical malaria specific
for this population [10], 89 cases of clinical malaria were detected for 52 different children; 31 clinical episodes were in
children <12 months old, and 58 were in children 112 months
old. Twenty-one percent (19 of 89) of clinical infections were
in sickle cell trait carriers (i.e., AS, AC, or SC genotype).
Of the 34 children with 1 4 parasite-positive blood samples
in the first 12 months of life, 13 were randomly selected for
more detailed investigation. In this subgroup, parasite densities
ranged from detectable by PCR only to 1170,000 parasites/mL
(median [25th and 75th percentiles], 303 [43 and 2800] parasites/
mL. Only 6 infections were accompanied by febrile symptoms
at the time of sampling; all were of very low parasite density
(!100 parasites/mL), were not detected at the time the child was
examined, and thus were not treated.
Parasite diversity. Because we were typically working at
the limit of parasite detection, samples that—using the multicopy 7H8/6 locus—were PCR positive for P. falciparum were
frequently negative when typed at the single copy msp-1 or msp2 loci. Eight hundred sixty-eight samples, from 126 children,
tested positive for P. falciparum by 7H8/6 PCR; of them, 690
gave a positive result in >1 of the 4 genotyping assays (dimorphic and polymorphic markers of msp-1 and msp-2). An
MSP-1 dimorphic type was determined for 433 samples, of
which 426 (98%) were of MAD20 type, 4 (0.9%) were of KI
type, and 5 (1.1%) contained both types. K1 alleles were isolated
from only 6 children and established a persistent infection in
only 1 child. Twenty-five different msp-1 block 2 alleles (defined
as different-sized PCR products) were detected in 153 samples,
with band sizes of 300–1200 bp. More than 1 band (indicating
a mixed infection) was observed in only 5 (3.3%) samples.
An MSP-2 dimorphic type was determined for 351 samples,
of which 203 (57.8%) were IC1 type, 96 (27.4%) were FC27
type, and 52 (14.8%) were mixed infections with both types
present. Polymorphic msp-2 alleles were detected in 229 samples, with band sizes of 450–1000 bp. More than 1 band (indicating a mixed infection) was observed in 41 (17.9 %) samples.
In the msp-1 and msp-2 markers taken together, evidence for
mixed-genotype infections was found in 101 (14.6%) samples,
but, in light of the relatively low positivity rate in the genotyping
assays, this is likely to be an underestimate.
In the group of 13 frequently infected children, parasites were
also typed for 4 microsatellite markers. The range of allele sizes
799
and the estimated frequency of each allele for the microsatellite
loci are shown in table 1. Microsatellite allele frequencies were
calculated by using only the major allele detected in each sample
and by scoring each allele only once for each child. Thus, although allele TA42-188 was present in several samples from
child 5, it was scored only once.
The most diverse locus tested was msp-2. In 13 children, 36
Table 1. Estimated microsatellite allele frequencies in a
population of Plasmodium falciparum parasites isolated
from Ghanaian infants.
Microsatellite locus
TA42
TA1
PfPK2
TA87
Allele
Frequency
179
182
188
191
199
203
218
233
245
136
145
160
163
166
169
172
175
178
181
184
187
190
193
196
154
163
166
169
172
175
178
181
184
187
190
193
196
199
91
94
97
100
103
106
109
112
115
—
0.095
a
0.476
0.095
0.048
0.095
0.048
0.048
0.095
—
0.027
0.135
0.081
0.108
0.108
—
0.135
0.108
0.108
a
0.162
—
0.027
—
—
—
0.091
0.182
0.121
a
0.212
0.091
—
0.061
—
0.121
—
0.091
0.030
0.060
—
—
0.048
a
0.238
0.143
0.190
0.143
0.143
0.095
NOTE. Frequency is based on use of major allele in each sample
only. Each allele was scored only once for each child (N p 13 ). “Allele” represents size (in bp) of the amplified polymerase chain reaction
fragment. Because alleles differ in the number of trinucleotide repeats, alleles differ in size by multiples of 3. Alleles with no frequency
given (represented by dashes [—]) were seen as minor alleles only.
a
Most common allele.
800
Franks et al.
Table 2.
Plasmodium falciparum infection patterns in individual
Ghanaian infants and children 0–2 years old.
Parasitepositive
samples
No. of
a
Study samples PCR1 BF1
5
12
15
17
29
45
47
52
62
65
71
81
120
9
14
16
12
15
13
18
17
19
20
20
18
20
4
4
8
8
8
4
6
4
11
8
6
5
5
Maximum Duration
Maximum
genotypes of single
No. of
parasite
per
genotype,
b
c
d
e
genotypes
density
sample
weeks
0
3
5
2
2
2
f
5
5
g
10
7
4
5
5
ND
1720
1300
516
258
46,440
4212
172,000
1462
10,621
8514
12,298
93,998
4
3
6
4
6
7
8
5
11
8
6
4
5
2
3
4
2
4
4
4
3
4
3
2
3
4
12
8
8
30
40
4
8
32
24
8
4
8
16
NOTE. BF1, blood film–positive; ND, no data; PCR1, polymerase chain
reaction–positive.
a
Number of blood samples collected (of a possible 22).
b
Maximum parasite density (per microliter) detected in any sample.
c
Total number of different parasite genotypes detected over the duration of
the study (i.e., number of different alleles detected at any 1 locus).
d
Maximum number of different alleles at any one locus detected in any 1
blood sample.
e
Duration of the same allele at >3 different loci. Detected in 1 sample only,
<4 weeks; detected in 2 sequential samples only, 4 weeks; detected in samples 8
weeks apart, 8 weeks; and so on.
f
One of these infections was not detected by PCR. Thus, the total number of
infected samples is 7.
g
One of these infections was not detected by PCR. Thus, the total number
of infected samples is 12.
different msp-2 alleles were identified by size polymorphism in
PCR, and, from the 73 complete msp-2 sequences obtained, 41
different alleles were identified; 29 of them were IC1-like, 11
were FC27-like, and 1 was a hybrid form with an IC1-like Nterminal sequence and an FC27-like C-terminal sequence. Of
the 41 msp-2 alleles, 37 were isolated from 1 child only, 3 alleles
were each found in 2 different children, 1 was found in 4 different children, and 1 was found in 5 different children.
By use of msp-1 and msp-2 data, the number of parasite
genotypes detected in each of the 13 children over the study
period was determined to be 3–8 (median, 4; mean, 4.62), and
Table 3.
JID 2001;183 (1 March)
the number of genotypes detected in any 1 sample was 1–3
(median, 2; mean, 2.28). When microsatellite data were included
(table 2), the number of genotypes per child was 3–11 (median,
6; mean, 5.92), and the number of genotypes per sample was
1–4 (median, 3; mean, 3.23).
Persistence of P. falciparum infections. Parasitemia in consecutive blood samples was classified as a single episode of
infection; parasitemia in 2 samples separated by an intervening
negative sample was classified as 2 separate episodes of infection. By use of this approach, the median number of discreet
episodes of infection per child (in the cohort of 143 children)
was 2 (range, 0–7; figure 1B). The duration of a single episode
was calculated as the number of weeks between the first and
last consecutive positive samples (e.g., if samples at weeks 10,
14, 18, and 22 all were parasite positive, it was defined as a
single episode lasting 12 weeks). If only a single sample was
positive (with negative samples collected 4 weeks earlier and 4
weeks later), the episode was deemed to have lasted !4 weeks.
By this definition, most malaria episodes lasted ! 4 weeks
(range, ! 4 to 164 weeks), but ∼40% of the children experienced
>1 malaria episode that persisted for >12 weeks (figure 1C).
Infections acquired in early infancy tended to be of shorter
duration than those acquired later in life. For example, only 6
(7.5%) of 80 infections acquired by children !5 months old
lasted for 1 4 weeks, whereas, on average, 38% of infections
acquired by children 15 months old lasted for 1 4 weeks (figure
1D). The pattern thus seems to be one of frequent, transient
infections in children !5 months old and slightly less frequent
but more persistent infections in children 15 months old.
Mixed-genotype infections were more commonly detected in
older children; 50 (18.1%) of 276 infections in children 15
months old were mixed infections, compared with only 7 (8.8%)
of 80 mixed infections in children !5 months old. This difference
is statistically significant (x 2 p 4.05 , 1 df; P ! .05) and may
reflect a truly higher proportion of mixed infections in older
children or may be an artifact due to higher parasite load and
increased likelihood of detection of minor clones.
It is not clear from this analysis, however, whether these
apparently persistent infections are due to persistence of a single
Plasmodium falciparum infection data from Ghanaian infant PP17.
Age, weeks
Parameter
Parasite density/mL
msp-1 Dimorphic
msp-1 Polymorphic
msp-2 Dimorphic
msp-2 Polymorphic
msp-2 Sequence
TA42
TA1
Pfpk2
TA87
2
18
26
!40
MAD20
570
FC27
490
NS
NP
169,184
172
97,109
!40
K1
NP
IC1
550
A2
NP
169,184
NP
97,109
!40
MAD20
520
IC1
550
A2
188
160
163
109
31
!40
MAD20
520
IC1
550
A2
188
160
163,166
109
34
344
MAD20
520
IC1
550
A2
188
160
163,166
109
38
!40
MAD20
520
IC1
550
A2
188
160
163,166
109
42
516
MAD20
520
IC1
550
A2
188
160
163,166
109
46
!40
MAD20
520
IC1
550
A2
188
160,166
163
109
NOTE. Data are polymerase chain reaction (PCR) band sizes or dimorphic genotype, except where noted. NP,
no product in PCR; NS, no sequence obtained.
JID 2001;183 (1 March)
Table 4.
Malaria in African Infants
801
Plasmodium falciparum infection data from Ghanaian infant PP29.
Age, weeks
Parameter
Parasite density/mL
msp-1 Dimorphic
msp-1 Polymorphic
msp-2 Dimorphic
msp-2 Polymorphic
msp-2 Sequence
TA42
TA1
Pfpk2
TA87
18
22
30
34
!40
MAD20
520
IC1
560
A1
188
178,184
181
ND
!40
MAD20
520
IC1
560
A2
188
175,178,184
181
ND
!40
MAD20
520
IC1
560
A2
188
175,178,184
166,181
103
86
MAD20
610
IC1
610
A2
NP
163,178,184
166,181
100
38
46
50
58
!40
!40
MAD20
640
IC1
640
A1
188
184,190
187
ND
MAD20
520
IC1
610
A2
188
178,181,184
166,181
103,106
260
MAD20
520
IC1
610
A2
188
178,184
166,172,181187
103
!40
MAD20
520
IC1
610
A1
NP
ND
ND
ND
NOTE. Data are polymerase chain reaction (PCR) band sizes or dimorphic genotype, except where noted. ND, no data; NP, no
product in PCR.
parasite genotype or to overlapping infections with parasites
of different genotypes. Persistence of individual parasite genotypes could be reliably determined only in the subgroup of 13
children whose infections were typed at 6 unlinked loci. A clear
example of long-term persistence for child PP17 is shown in
table 3. A parasite with the genotype msp1 MAD20/520, msp2 IC1/550-A2, TA42-188, TA1-160, Pfpk2-163, and TA87-109
was detected in 6 consecutive monthly samples (26–46 weeks
old), and the same clone was probably also present at week 18
as part of a mixed infection.
An example of a persistent, mixed-clone infection is shown
in table 4 (child PP29). Two clones (characterized by msp-2
sequences A1 and A2) were detected in multiple blood samples
from ages 18–58 weeks. Because not all alleles could be typed
in each sample, it is not possible to determine the genotype of
each clone, but, for several loci, 2 alleles are seen to persist
(msp-2 IC1-560 and IC1-610, TA1 178 and 184, and Pfpk2 166
and 181); the 2 clones appear to share the same TA42 allele.
Other minor clones appear transiently. Similarly, a persistent
mixed infection was detected in PP12 (table 5); 2 msp-2 alleles
and 2 TA1 alleles were detected in 3 consecutive samples. The
most complicated infection pattern seen was for child 62 (table
6), in whom infection with persistent clones (e.g., msp-2 A1,
A6, and B1) was supplemented by transient superinfections.
By detection of the same allele at >3 loci to define persistence,
the estimated duration of individual infections for each child was
4–40 weeks (table 2). Some clones appeared to persist longer but
could be typed at only 1 or 2 loci. For the microsatellite loci,
where the number of alleles in the population was low, persistence
at a single locus is an unreliable marker of persistence, but for
other markers, such as msp-2 sequence, persistence can be more
reliably inferred. Longer-term persistence at levels below the limit
of detection by PCR is also possible.
Discussion
Numerous studies in the past decade have examined parasite
genetic diversity within individual human hosts [4–6, 8, 21–23],
but only one of these has looked at long-term parasite per-
sistence in an endemic population. Babiker et al. [6] documented asymptomatic persistence of single parasite clones for
>12 months in young Sudanese adults and estimated that a
chronic infection was established in ∼40% of infected individuals (of all ages). The study reported here differs from previous
genotyping studies in several ways. First, because most of the
infections in this study were asymptomatic, clearance of parasite populations by drug treatment was infrequent, thus allowing the natural duration of infections to be estimated. Second,
the age group being studied was restricted to children !2 years
old, the age group in which rapid acquisition of antimalarial
immunity is occurring. Third, parasites were typed at 6 unlinked
and highly polymorphic loci.
The greatest allelic diversity was observed in the msp-2 locus,
with 41 different alleles detected by sequencing in just 1200
blood samples from 13 children. Simple PCR amplification of
this region revealed almost identical levels of diversity, however,
which indicates that differences in PCR band size are almost
as sensitive as sequencing in picking up novel msp-2 alleles. The
microsatellite loci were less diverse than were msp-2, with 9–15
alleles at each locus. The TA42-188 allele was found as a major
allele in almost half the children; persistence of this allele was
thus not a particularly good marker (on its own) of persistent
infection by a single parasite clone. The greatest diversity
Table 5.
PP12.
Plasmodium falciparum infection data from Ghanaian infant
Age, weeks
Parameter
Parasite density/mL
msp-1 Dimorphic
msp-1 Polymorphic
msp-2 Dimorphic
msp-2 Polymorphic
msp-2 Sequence
TA42
TA1
Pfpk2
TA87
42
340
MAD20
520
IC1
650
ND
ND
ND
ND
ND
46
645
MAD20
520
IC1
610,515
ND
188
175,172
175
100
50
1730
MAD20
520
IC1
610,515
ND
188
175,172,178
175,172
100
54
!40
MAD20
520
IC1
610,515
ND
188
175,172
175
100
NOTE. Data are polymerase chain reaction band sizes or dimorphic genotype, except where noted. ND, no data.
NOTE.
10
MAD20
555,570
IC1,FC27
510,580
A2,B1
NP
NP
NP
NP
!40
14
860
MAD20
555,570
IC1,FC27
510,580
A3,B1
188
163,175
175
106
18
645
MAD20
555,570
IC1,FC27
510,580
A4,B1
188,233
136,163,169
166,172,178
100,106
22
86
MAD20
590
IC1
510,580
A1
233
160,169,175,187
166,172
100,106
26
Age, weeks
42
1462
MAD20
490
IC1
510,540,680
A4,A5
182,188,191,203
163,178
169,172
103,106
46
86
NP
520
IC1
670
A6
184,199
175
154,166172
100,106
50
129
MAD20
520
IC1
670
A6
NP
181
166
100,106
Data are polymerase chain reaction (PCR) band sizes or dimorphic genotype, except where noted. NP, no product in PCR; NS, no sequence obtained.
86
MAD20
MAD20
590
555,570
IC1
FC27
530
510
A1
NS
NP
182
145,163169,184 160,166
NP
169
NP
100,115
!40
2
Plasmodium falciparum infection data from Ghanaian infant PP62.
Parasite density/mL
msp-1 Dimorphic
msp-1 Polymorphic
msp-2 Dimorphic
msp-2 Polymorphic
msp-2 Sequence
TA42
TA1
Pfpk2
TA87
Parameter
Table 6.
62
192
MAD20,K1
520
IC1
670
A3,A6,A7
182
178
181
106
70
120
NP
NP
NP
NP
NS
191
181,184
199
111
JID 2001;183 (1 March)
Malaria in African Infants
among the microsatellite markers was seen at the TA1 locus,
for which 15 different alleles were identified. Accordingly, TA1
revealed higher levels of multiplicity of infection than the other,
less-polymorphic, markers.
Genotyping of infections in this study was complicated by
the fact that most infections were of very low parasite density
(at the limit of detection by PCR); it seems likely that, in many
cases, not all the alleles present in the individual have been
amplified. Because mixed infections may arise by multiple inoculations of individual clones or by a single inoculation of a
mixed infection that may have undergone crossing (and recombination) in the mosquito vector [16], we could not assume the
presence of discrete genotypes. Thus, the number of genotypes
in a single sample was estimated conservatively as the highest
number of alleles detected at a single locus. Estimates of the
total number of genotypes with which an individual was infected over the period of the study were also made conservatively, on the assumption that the detection of a particular allele
in 11 sample was due to the persistence of a parasite genotype
rather than to reinfection with another parasite carrying the
same allele. In light of the fact that the number of distinct alleles
detected at each locus was 9–42, and all 6 loci are unlinked in
the parasite genome, we estimate that the chances of 2 different
parasite clones sharing the same alleles at 3 loci, of which msp2 sequence is one, range from 1 in 11,000 to 1 in 225,000.
Persistence of alleles at >3 loci is thus a very robust definition
of long-term infection by the same parasite clone.
Most infections in very young children (!5 months old) were
transient, lasting ! 4 weeks, and were spontaneously cleared
without reaching high densities or causing clinical disease.
Rapid clearance of infections may explain why these infections
tended to be genetically less complex than infections in older
infants. This combination of short-lived and genetically simple
infections implies that children !5 months old would experience
a narrower spectrum of antigenic diversity than would older
infants. In contrast, infections in children 15 months old tended
to last longer and to be more genetically complex, which indicates that the repertoire of responses to variant antigens may
develop rapidly after ∼5 months of age.
The finding that malaria infections in infants tend to be rapidly cleared without causing clinical symptoms agrees with data
from large epidemiological studies, in which seroconversion and
splenic enlargement were detected in children in the absence of
obvious clinical malaria [24]. Most children in our study experienced multiple infections in the first 2 years of life, with >7
discrete episodes of infection detected per child. Our monthly
screening program, however, will have missed many short-lived
infections, and real exposure levels will be higher than indicated
by our data.
Although many infections were rapidly cleared, infections in
some children persisted for many months. Two children were
found to be infected on every occasion that they were tested
for 50 and 64 weeks, respectively, and more than half the chil-
803
dren had >1 episode of infection that lasted for 13 months.
The longest persistence of a single parasite genotype (documented by persistence of the same allele at >3 loci) was 10
months. It is interesting that children with very long-term infections were likely to be infected with msp-2 sequences that
were also found in other children (data not shown), which suggests that they are a reservoir for infection in the community.
Our estimate of infection complexity in the cohort as a whole
(∼15% mixed infections, giving an MOI of 1.15) is likely to be
a substantial underestimate, in light of the fact that we were
able to pick up many very low-level infections by using the
multicopy 7H8/6 sequence, but not all these infections could
be genotyped by using single-copy gene targets. In the 13 children whose infections were typed exhaustively (with negative
PCRs being repeated >10 times until a positive result was obtained) at 6 loci, the MOI was 2.28, which is likely to be quite
an accurate estimate and accords with previous estimates of
MOI in areas with similar levels of transmission [4, 25].
It is also interesting that many of the children were clearly
infected with novel parasite genotypes without developing clinical symptoms, which indicates that, although novel infections
may lead to clinical disease [4], they frequently do not. These
data throw doubt on the notion that acquisition of clinical
immunity to malaria is simply a matter of building up a repertoire of responses to variant antigens. Even very young children can cope quite well with most parasite infections they
acquire; even in this age group, clinical disease appears to be
a relatively uncommon consequence of malaria infection. This
raises a very important question: what causes some infections
to become symptomatic when so many infections appear to be
rather innocuous? Is it a function of immunity or of the child’s
age or some other variable? Is it primarily determined by the
parasite or the host?
Children !5 months old appear to be genuinely resistant to
malaria, in that infections occur but do not become established.
There are a number of potential immunological and physiological explanations for this, including the presence of fetal
hemoglobin, immaturity of infant red blood cells, and the lack
of essential parasite growth factors in the blood [26]. By contrast, children 15 months old tend to develop persistent infections. These children, presumably, possess effective immune
mechanisms for limiting (but not eliminating) parasite growth.
In conclusion, this study demonstrated that malaria infections in infants !5 months old are frequent but short lived. In
children 15 months old, however, persistent, genetically complex, subclinical malaria infections occur frequently in otherwise healthy infants and young children. Thus, even in areas
of moderate-to-low levels of malaria transmission, most children will have been exposed to a wide variety of polymorphic
or variant malaria antigens over a relatively short period of
time. These findings have implications for the way in which we
view acquired immunity to malaria.
804
Franks et al.
Acknowledgments
We thank the mothers and children of Prampram, Ghana, for taking
part in this study; members of staff of the Epidemiology Unit and
Immunology Unit of the Noguchi Memorial Institute for Medical Research (Legon, Ghana), for their assistance in carrying out this study;
Michael Addae (Noguchi Memorial Institute for Medical Research)
for sickle cell electrophoresis; Martin Waterfall (Institute of Cell, Animal, and Population Biology, University of Edinburgh, Edinburgh,
United Kingdom) for performing some of the genotyping; David Conway and Cally Roper for assistance with microsatellite analysis; and
Steve Bennett and Jim Todd (Department of Infectious and Tropical
Diseases, London School of Hygiene and Tropical Medicine, London,
United Kingdom) for statistical advice.
References
1. McGregor IA. The development and maintenance of immunity to malaria
in highly endemic areas. Clin Trop Med Commun Dis 1986; 1:1–29.
2. Bruce-Chwatt L. Essential malariology. 2d ed. London: Heinemann, 1985.
3. Krajden S, Panisko D, Tobe B, Yang J, Keystone J. Prolonged infection with
Plasmodium falciparum in a semi-immune patient. Trans R Soc Trop Med
Hyg 1991; 85:731–2.
4. Contamin H, Fandeur T, Rogier C, et al. Different genetic characteristics of
Plasmodium falciparum isolates collected during successive clinical malaria
episodes in Senegalese children. Am J Trop Med Hyg 1996; 54:632–43.
5. Daubersies P, Sallenave-Sales S, Magne S, et al. Rapid turnover of Plasmodium falciparum populations in asymptomatic individuals living in a
high transmission area. Am J Trop Med Hyg 1996; 54:18–26.
6. Babiker H, Abdel-Muhsin A, Ranford-Cartwright L, Satti G, Walliker D. Characteristics of Plasmodium falciparum parasites that survive the lengthy dry
season in eastern Sudan where malaria transmission is markedly seasonal.
Am J Trop Med Hyg 1998; 59:582–90.
7. Farnert A, Snounou G, Rooth I, Bjorkman A. Daily dynamics of Plasmodium
falciparum subpopulations in asymptomatic children in a holoendemic
area. Am J Trop Med Hyg 1997; 56:538–47.
8. Smith T, Felger I, Kitua A, Tanner M, Beck HP. Dynamics of multiple
Plasmodium falciparum infections in infants in a highly endemic area of
Tanzania. Trans R Soc Trop Med Hyg 1999; 93:(Suppl 1)35–9.
9. Wagner G, McGuinness D, Koram K, Bennett S, Nkrumah FK, Riley E.
High incidence of asymptomatic malaria infections in a birth cohort of
children under 1 year of age in Ghana, detected by multicopy gene polymerase chain reaction. Am J Trop Med Hyg 1998; 59:115–23.
10. McGuinness D, Koram K, Bennett S, Wagner G, Nkrumah FK, Riley E.
Clinical case definitions for malaria: clinical malaria associated with very
low parasite densities in African infants. Trans R Soc Trop Med Hyg
1998; 92:527–31.
JID 2001;183 (1 March)
11. Afari EA, Koram K, Dunyo S, Nkrumah FK. The epidemiology of malaria
with special emphasis on transmission, morbidity, mortality and disease
control in Ghana. Accra: University of Ghana, 1996.
12. Trape J-F, Peelman P, Morault-Peelman B. Criteria for diagnosing malaria
among a semi-immune population exposed to intense and perennial transmission. Trans R Soc Trop Med Hyg 1985; 79:435–42.
13. Foley M, Ranford-Cartwright LC, Babiker HA. Rapid and simple method
for isolating malaria DNA from fingerprick samples of blood. Mol
Biochem Parasitol 1992; 53:241–4.
14. Reeder J, Marshall V. A simple method for typing Plasmodium falciparum
merozoite surface antigens 1 and 2 (MSA-1 and MSA-2) using a dimorphicform specific polymerase chain reaction. Mol Biochem Parasitol 1994; 68:
329–32.
15. Ranford-Cartwright LC, Balfe P, Carter R, Walliker D. Genetic hybrids of
Plasmodium falciparum identified by amplification of genomic DNA from
single oocysts. Mol Biochem Parasitol 1991; 49:239–44.
16. Ranford-Cartwright L, Balfe P, Carter R, Walliker D. Frequency of crossfertilisation in the human malaria parasite Plasmodium falciparum. Parasitology 1993; 107:11–18.
17. Fenton B, Clark JT, Khan CMA, et al. Structural and antigenic polymorphism
of the 35- to 48-kilodalton merozoite surface antigen (MSA-2) of the malaria
parasite Plasmodium falciparum. Mol Cell Biol 1991; 11:963–71.
18. Ranford-Cartwright LC, Taylor RR, Smith DB, et al. Differential antibody
recognition of FC27-like Plasmodium falciparum merozoite surface protein
MSP2 antigens which lack 12 amino acid repeats. Parasite Immunol
1996; 18:411–20.
19. Su X-Z, Wellems TE. Toward a high-resolution Plasmodium falciparum linkage map: polymorphic markers from hundreds of simple sequence repeats.
Genomics 1996; 33:430–44.
20. Anderson T, Su X-Z, Bockarie M, Day K. Twelve microsatellite markers for
characterisation of Plasmodium falciparum from finger prick blood samples. Parasitology 1999; 119:113–25.
21. Mercereau-Puijalon O. Revisiting host/parasite interactions: molecular analysis of parasites collected during longitudinal and cross-sectional surveys
in humans. Parasite Immunol 1996; 18:173–80.
22. Ntoumi F, Contamin H, Rogier C, Bonnefoy S, Trape J-F, Mercereau-Puijalon O. Age-dependent carriage of multiple Plasmodium falciparum merozoite surface antigen–2 alleles in asymptomatic malaria infections. Am J
Trop Med Hyg 1995; 52:81–8.
23. Roper C, Richardson W, Elhassan I, et al. Seasonal changes in the Plasmodium falciparum population in individuals and their relationship to
clinical malaria: a longitudinal study in a Sudanese village. Parasitology
1998; 116:501–10.
24. Brabin B. An analysis of malaria parasite rates in infants: 40 years after
Macdonald. Trop Dis Bull 1990; 87:R1–R21.
25. Conway D, Greenwood B, McBride J. The epidemiology of multiple-clone
Plasmodium falciparum infections in Gambian patients. Parasitology 1991;
103:1–6.
26. Riley E, Wagner G, Akanmori B, Koram K. Do maternally acquired antibodies protect infants from malaria infection? Parasite Immunol (in press).