Am. J. Trop. Med. Hyg., 71(1), 2004, pp.

2939
Copyright 2004 by The American Society of Tropical Medicine and Hygiene

ANALYSIS OF CIRCUMSPOROZOITE PROTEINSPECIFIC IMMUNE RESPONSES

FOLLOWING RECENT INFECTION WITH PLASMODIUM VIVAX
CHAISUREE SUPHAVILAI, SORNCHAI LOOAREESUWAN, AND MICHAEL F. GOOD
The Cooperative Research Centre for Vaccine Technology, The Queensland Institute of Medical Research, Herston, Queensland,
Australia; The Tropical Medicine Hospital, Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand

Abstract. CD8+ and CD4+ T cells are involved in immunity to the pre-erythrocytic stage of malaria. This study has
been undertaken to define T cell epitopes on the Plasmodium vivax circumsporozoite protein (CSP) and to analyze the
early induction of immune response following infection. We identified CD4+ and CD8+ T epitopes recognized by
different strains of mice as well as by humans. The CD4+ T cell response in mice was found to be similar in all strains,
but variation between strains was evident. Five H-2d-restricted CD8+ cytotoxic T lymphocyte (CTL) epitopes, but no
H-2k-or H-2b-restricted epitopes, could be defined. Non-H-2 genes were also able to regulate the response. In recently
infected Thai adults, poor immunoresponsiveness was demonstrated. CTL activity and proliferative responses of T cells
from malaria-exposed donors were very low. In contrast, exposed individuals had specific antibodies against the immu-
nodominant repeats of both common strains of the P. vivax CSP; however, titers decreased following treatment.

INTRODUCTION In Thailand, malaria is a significant problem and P. vivax

Immunity to malaria sporozoites is believed to be mediated
by both T cells and antibodies1 and evidence from both ro-
dent2,3 and human studies4 suggests that the circumsporozoite
protein (CSP) can be a target of both T cells and antibodies,
thus rendering it a leading sporozoite/liver stage vaccine can-
didate. Data from murine studies have suggested that CD8+
as well as CD4+ T cells are required for protection against
malaria,3 and humans who were immunized with irradiated
sporozoites of Plasmodium falciparum, as well as those natu-
rally exposed to P. falciparum sporozoites, developed CD8+
and CD4+ T cell responses against the P. falciparum CSP,
although such responses were often infrequently present.5,6
Although logistically very difficult, one way to define human
CD8+ T cell epitopes on the CSP of P. vivax is to expose
volunteers to the bites of hundreds of irradiated mosquitoes
heavily infected with P. vivax sporozoites. However, a num-
ber of studies have now demonstrated that the location of
epitopes on a given protein recognized by human T cells may
be very similar to, often overlapping, the region recognized by
murine T cells. For example, human and murine CD4+ T cell
epitopes map to the same region of the P. falciparum CSP7,8
and a murine CD8+ T cell epitope on the P. falciparum CSP
overlaps a human CD8+ T cell epitope.5,6,9
Analysis of antibody responses to P. vivax CSP is important
but is complicated by the fact that there are two distinct types
of repeats on the P. vivax CSP.10,11 Antibodies against re-
peats have been reported to occur in the serum of malaria-
exposed subjects,12,13 and antibody to non-repeat regions has
also been found,14 a situation different to that for P. falci-
parum, where the vast majority of antibody to sporozoites is
directed against the repeats only.15 Since it is believed that
CSP-specific CD4+ and CD8+ T cell responses, along with
CSP-specific antibodies, will be important in immunity to P.
vivax sporozoite/liver stages, and because polymorphisms
within the P. vivax CSP16,17 could render a single recombi-
nant CSP ineffective as a vaccine candidate, a major strategy
is to define T and B cell epitopes and combine these in a
synthetic vaccine. An epitope on P. vivax CSP recognized by
a protective monoclonal antibody has been defined,18 and
vaccine constructs based on this19 and CSP T cell epitopes20
have been made. Other epitopes defined from monkey21 and
murine20,22 studies may also prove useful.
accounts for nearly 30% of all infections. The incidence of P.
vivax malaria is higher during the wet season (MayAugust).
It affects mostly non-immune individuals traveling for work
to the regions of high endemicity on the borders of Myanmar,
Laos, and Cambodia. Although there have been some studies
of human immune responses to the P. vivax CSP,12,2325 these
have been limited and only two former studies have looked at
T cell responses among Thais.25,26 There have been no studies
to define Thai CD8+ T cell epitopes on the CSP of P. vivax.
The goals of this study were to facilitate vaccine design and
development by analyzing the T and B cell responses specific
to P. vivax CSP of recently infected Thais, gain insight into
the early induction of CSP-specific immunity, and define the
possible location of other important epitopes by examining
the immune responses of vaccinated mice.
MATERIALS AND METHODS
Patients. One hundred thirteen P. vivax-exposed donors
(age range 1465 years) were recruited from patients from
the Kanchanaburi area admitted to The Tropical Medicine
Hospital at Mahidol University in Bangkok, Thailand. All
exposed donors were slide positive for P. vivax at the time of
admission. Some patients had multiple (27) episodes of P.
vivax infection. The patients were bled on days 0, 7, and 28
after admission. Some follow-up patients were tested one
month to two years after their last episode. Thirty-six non-
exposed volunteers (age range 2255 years) were Bangkok
residents with no history of malaria. This research was re-
viewed and approved by the Bancroft Human Ethics Re-
search Committee. Peripheral blood mononuclear leukocytes
were prepared from heparinized blood and tested for their
proliferative response and their ability to generate cytotoxic
responses in vitro to P. vivax peptides. Serum or plasma was
also collected at each time point, as available, for estimation
of antibody responses to P. vivax peptides.
Animals. Six to eight week-old female BALB/c (H-2d),
B10.D2 (H-2d), B10.BR (H-2k), B10 (H-2b), and F1 BALB/c
B10.D2 mice were obtained from the Animal Resources Cen-
ter, Willetton, Western Australia.
Peptides and antigens. 20-mer overlapping peptides span-
ning the entire sequence of P. vivax (Belem and North Ko-
rean strains and the repeat variants of the VK strain), a CD8+

29
30 SUPHAVILAI AND OTHERS

epitope of P. falciparum CSP, f97 (IEKYLKTIKNSL- control for human CTL assays. The CTL response to the
STEWSPCS) recognized by BALB/c mice,27 and f57 (KP- influenza matrix peptide has previously been described to be
KDELDYENDIEKKICKMEKCS), a CD8+ epitope from P. restricted by HLA-A2.1.29 The frequency of HLA-A2 in the
falciparum CSP recognized by B10.BR mice9 were used. Thai population is approximately 10%.30 Another report has
These peptides were synthesized by the tea bag method28 at indicated that the frequency of the HLA-A2.1 subtype is
The Queensland Institute of Medical Research and their se- rare.31 Tetanus toxoid (TT) (0.3 flocculation units/mL) (Com-
quences are given (Table 1). Because of the large number of monwealth Serum Laboratories, Parkville, Victoria, Austra-
peptides to be tested and the difficult logistics of testing pe- lia) and purified protein derivative (PPD) of Mycobacterium
ripheral blood mononuclear leukocytes to each peptide from tuberculosis (Commonwealth Serum Laboratories) (60 g/
only a small volume of blood, peptides were frequently mL) were used as positive control antigens. Bovine serum
pooled, following a similar methodology in other studies.5 It is albumin (BSA)coupled peptides were used as antigens in an
possible that by pooling peptides antigenic competition may enzyme-linked immunosorbent assay (ELISA) for the deter-
result in some peptides being non-immunogenic. Major his- mination of P. vivax-specific antibodies. Peptides were
tocompatibility complex (MHC)linked and non-MHC- coupled to BSA using the glutaraldehyde coupling method as
linked genes could influence the outcome in a manner partly described.10
dependent on the conditions used in culture and assay. For Mapping of CD4+ T cell epitopes. Three to four mice were
the CD4+ T cell studies in mice, pool A consisted of v1v10, immunized with 30 g per animal of pooled peptides emul-
pool B consisted of v11v20, and pool C consisted of v21v30. sified in Freunds complete adjuvant in the foot pad unless
For the CD8+ T cell studies in mice, group I consisted of otherwise specified (50 L per footpad). Inguinal and popli-
v1v5, group II consisted of v6v10, group III consisted of teal lymph nodes were collected 89 days after immunization
v11v15, group IV consisted of v16v20, and group V con- and tested for their proliferative response to P. vivax CSP
sisted of v21v26. There were equal amounts of each peptide peptides at concentrations of 30 g/mL and 10 g/mL in a
in the pool. When used for generation of in vitro human volume of 200 L. Responses were expressed as counts per
cytotoxic T lymphocyte (CTL) cultures, they were grouped minute (mean cpm of wells with peptide mean cpm of wells
as followed: group I (v1v5), group II (v6v10), group III without peptide). In human studies, lymphocyte proliferation
(v11v15 and v27v30), group IV (v16v20), and group V assays were performed against individual peptides at a con-
(v21v26). The influenza matrix peptide (residues 5573, centration of 30 g/mL. Results are expressed as stimulation
LTKGILGFVF TLTVPSERG-amide) was used as a positive indices (SIs), where an SI > 3.0 was arbitrarily considered a
positive response. Although this is an arbitrary cut-off value,
it has been commonly used in similar studies.23,25
TABLE 1 Mapping of CD8+ T cell epitopes. Animals were pre-
Overlapping peptides representing entire sequence of the circum- primed with phosphate-buffered saline emulsified in Freunds
sporozoite protein of Plasmodium vivax (Belem and NK and re- incomplete adjuvant at the tail base three weeks prior to im-
peat variants of VK)* munizing with 50 g of peptide also emulsified in Freunds
incomplete adjuvant following an established protocol.32 In
Peptide
no. Residues Peptide sequence vitro CTLs were generated against the appropriate peptide
V1 120 MKNFILLAVSSILLVDLFPT pool at a concentration 10 g/mL with 30 units/mL of recom-
V2 1130 SILLVDLFPTHCGHNVDLSK binant interleukin-2 (rIL-2) being added on day 2. The cul-
V3 2140 HCGHNVDLSKAINLNGVNFN ture was tested for CTL activity after 79 days. Some cultures
V4 3150 AINLNGVNFNNVDASSLGAA were also restimulated with 10 g/mL of antigen, 30 units/mL
V5 4160 NVDASSLGAAHVGQSASRGR
V6 4665 SLGAAHVGZSASRGRGLDEN of rIL-2, and irradiated spleen cells every two weeks. The
V7 5675 ASRGRGLDENPDDEEGDAKK conventional chromium release assay was performed 78
V8 6685 PDDEEGDAKKKKDGKKAEPK days after restimulation. P-815 cells were used as target cells
V9 7695 KKDGKKAEPKNPRENKLKQP for BALB/c and B10.D2 CTL, while L-929 and EL-4 cells
V10 86105 NPRENKLKQPGDRADGQPAG
were used as targets for B10.BR and B10 cultured cells, re-
V11 97116 DRADGQPAGDRADGQPAGDR
V12 124143 DRAAGQPAGDRADGQPAGDR spectively.
V13 160179 DRADGQPAGDRAAGQPAGDR Human CTL cultures were set up as described5 in 24-well
V14 232251 DRAAGQPAGDRAAGQPAGDRC plates. For in vitro secondary restimulation, cultures were fed
V15 259278 DRAAGQPAGNGAGGQAAGGN with autologous mitomycin C-treated phytohemagglutinin-
V16 NKins GAGGQAAGGNAANKKAEDAG
V17 NKins AANKKAEDAGGNAGGNAGGG stimulated blasts in the presence of peptides (10 g/mL) and
V18 279298 AGGGQGQNNEGANAPNEKSV rIL-2 (30 units/ml) in 25-cm2 flasks. The influenza matrix pep-
V19 289308 GANAPNEKSVKEYLDKVRAT tide was used as a positive control (10 g/mL). The CTL
V20 299318 KEYLDKVRATVGTEWTPCSV activity was assessed seven days after restimulation. The spe-
V21 309328 VGTEWTPCSVTCGVGVRVRR
cific CTL activities were calculated and specific lysis greater
V22 319338 TCGVGVRVRRRVNAANKKPE
V23 329348 RVNAANKKPEDLTLNDLETD than 15% was arbitrarily considered positive, as described
V24 339358 DLTLNDLETDVCTMDKCAGI elsewhere.5
V25 349368 VCTMDKCAGIFNVVSNSLGL CD4+ or CD8+ T cell depletion of LN cells or CTL lines. The
V26 359378 FNVVSNSLGLVILLVLALFN LN or CTL cultured cells were incubated with excess amounts
V27 VK247 ANGAGNQPGANGAGNQPG-18mer
V28 VK247 ANGAGNQPGEDGAGNQPG-18mer of rat anti-mouse CD4 (GK1.5) or CD8 (TIB105) monoclonal
V29 VK247 EDGAGNQPGANGAGNQPG-18mer antibodies. The depleted cell population was then tested for
V30 VK247 ANGAGNQPGANGAGGQAA-18mer its T cell activities. The CD4+ T cell depletion ranged from
* Variant residues are specified in bold. 83% to 97% and CD8+ T cell depletion ranged from 91% to
 IMMUNITY TO P. VIVAX CIRCUMSPOROZOITE PROTEIN
96% as determined by fluorescence-activated cell sorting
analysis.
Detection of P. vivax peptide-specific antibody by ELISA. Hu-
man IgG antibodies against peptides representing the two
types of P. vivax CSP repeats (v14 and v30 representing the
classic repeats and the VK 247 repeats, respectively) (Table
1) were detected by an ELISA. Peptide was coupled to BSA
and then coated onto the ELISA plate at a concentration of
2 g/mL in coating buffer. Samples where the net optical
density at 405 nm (OD405) at a dilution of 1:100 was greater
than the mean + 3 SD of the non-exposed samples tested
against each peptide were considered positive. In cases where
there was a high BSA-specific antibody response that sample
was omitted.
RESULTS
CD4+ T cell responses. Different strains of mice (BALB/c
[H-2d], B10.D2 [H-2d], B10.BR [H-2k], and B10 [H-2b]) were
immunized with the different pools of peptides as described
earlier in this report. Draining LN cells were then cultured
with individual peptides to define CD4+ T cell epitopes. For
some LN cell populations, cells were depleted prior to culture
using antibodies to CD4 or CD8. The CD4+ T cell epitopes
were found to be clustered in two regions of the P. vivax
CSP, one at the amino terminal region in the first 60 amino
acids and the other at a carboxyl terminal region covering
a conserved region (region II, comprising amino acids
TEWTPCSVTCGVG within peptide v21) of the protein. All
four mouse strains responded well to v21 (amino acids 309
328). BALB/c and B10.D2 mice showed the same pattern of
responses to all peptides (Figure 1a and b). B10.BR mice
responded to the classic repeat peptide (v14) (Figure 1c),
while B10 mice responded to the variant repeat peptide (v30)
(Figure 1d). The CD4+ T cell depletion prior to in vitro stimu-
lation abrogated most proliferative responses, while the effect
on the proliferative response of CD8+ T cell depletion was
much less (Figure 1).
Individual humans were also tested for their lymphoprolif-
erative responses (Figure 2). Control antigens (TT and PPD)
were included in the study. Both non-exposed and exposed
subjects responded at approximately equivalent levels to
these control antigens (96% of non-exposed subjects and 98%
of exposed subjects responded to PPD, while 63% of non-
exposed and 44% of exposed subjects responded to TT). The
P. vivax-exposed group was recently exposed and were di-
vided into those with a single exposure and those with mul-
tiple exposures. The data are expressed as the percent of
responders in a given group of subjects. As mentioned earlier,
positivity was defined as a stimulation index greater than 3.
Little responsiveness by peripheral blood mononuclear cells
of non-exposed subjects to the CSP peptides was found, with
one (4%) of 25 responding to each of v6, v7, v9, v20, and v22,
and two (8%) of 25 responding to v18. These peptides have
been reported to stimulate cells from non-exposed subjects in
other studies.23,33 However, malaria-exposed subjects also re-
sponded infrequently to the peptides. In those individuals
who had their first episode, three (16%) of 19 responded to
v2, one (5%) of 19 to v5, one (5%) of 19 to v9, v12, v22, v23,
v27, and two (11%) of 19 to v19 (Figure 2). In those individu-
als with multiple episodes of P. vivax malaria, two (8%) of 24
responded to v1 and v2, one (4%) of 24 to v19, v21, and v25,
31
and three (13%) of 24 to v22 (Figure 2). Levels of respon-
siveness to CSP peptides by both exposed and non-exposed
subjects were much less than that previously observed33 and
may reflect the fact that the individuals in this study were
recently infected, compared with the other studies.
CD8+ T cell responses. Using an established protocol (see
Materials and Methods) different strains of mice were immu-
nized with the different peptide pools for definition of puta-
tive CTL epitopes. BALB/c (H-2d) mice were found to rec-
ognize peptides v1, v2, v5, v19, v21, and v25, while B10.D2
mice (also H-2d) repeatedly recognized only v19 (Figure 3).
Peptides 21 and 25 did not always induce a CTL response in
BALB/c mice, but did so on two of four occasions. The dis-
crepancy in CTL responsiveness between the H-2-identical
strains (BALB/c and B10.D2) suggested that there may be
negative regulatory non-MHC elements in B10.D2 mice af-
fecting the responsiveness to these epitopes. To assess this
possibility, BALB/c and B10.D2 mice were crossed to pro-
duce F1 mice that were subsequently tested for immunologic
responsiveness. However, the F1 mice were found to recog-
nize the same P. vivax peptides as the parental BALB/c
strain, suggesting that the lack of responsiveness was not due
to a dominant suppressive regulatory effect. A known P. fal-
ciparum H-2d-restricted epitope, f97 (sequence given in the
Materials and Methods),27 was also tested in the three H-2d
strains and was shown to induce CTL in BALB/c, B10.D2
(Figure 3), and F1 mice. All defined epitopes were shown to
be recognized by CD8+ T cells because CTL activity was
reduced after CD8 depletion by more than 75%. No CTL
activity against P. vivax CSP peptides was demonstrated
by either B10 (H-2b) or B10.BR (H-2k) mice using this
protocol. A positive control peptide for B10.BR mice (the P.
falciparum CSP peptide f57 (sequence given in the Materials
and Methods)9 was tested and shown to be immunogenic for
CTL activity.
It was of interest that no P. vivax CSP-specific epitopes
were defined in B10 or B10.BR mice. A particular concern
was the possibility that the lack of response was the result of
a lack of suitable CD4+ T cell-mediated help for the genera-
tion of CTL. The CD4+ T cell epitopes have been defined on
the P. vivax CSP and it was noted that there was a paucity of
CD4+ T cell responsiveness among pools II and III. We thus
immunized B10 and B10.BR mice with pool II or pool III
peptides together with either v21 (well recognized by CD4+ T
cells from all strains) or the CD4+ T cell promiscuous epitope
from tetanus toxin P30,34 which was confirmed here to be
immunogenic in BALB/c, B10.D2, and B10 mice (Figure 4a).
Using peptide 21 as a putative source of help for CTL gen-
eration, we showed that there was still no generation of CTL
activity from B10.BR and B10 mice in response to pools II or
III. However, when P30 was used as a source of putative help
in B10 mice, CTL were generated to the group II pool con-
taining P30 (pool III was not tested with P30). Analysis of the
response, however, showed that the CTL were specific for
P30 itself, not the peptides in pool II (Figure 4b). P30 and v21
were similarly unsuccessful in inducing a CTL response to
peptides from either pool II or pool III in B10.BR mice. This
same approach was then used to assess whether further
epitopes could be defined in BALB/c and B10.D2 mice from
among the negative pools. No further CTL epitopes from the
P. vivax CSP were defined for these mice. However, it was
observed that P30 itself was also seen by CD8+ T cells from
32 SUPHAVILAI AND OTHERS

FIGURE 1. Proliferative response of undepleted, CD4+ T cell-depleted, and CD8+ T cell-depleted cells in BALB/c (a), B10.D2 (b), B10.BR
(c), and B10 (d) mice immunized with Plasmodium vivax peptide pools (see Materials and Methods for details). Data are expressed as delta counts
per minute (CPM). The background cpm for the results shown here ranged from 91 to 868.

BALB/c mice (Figure 4b). Curiously, B10.D2 mice failed to
recognize this peptide. Presumably, non-MHC genes are con-
tributing to the regulation of this response as has been shown
in other systems.35
A polymorphism occurs naturally in peptide v19, the
epitope recognized by CD8+ CTL from BALB/c and B10.D2
mice. A single amino acid mutation occurs in the Chinese
strain36 at position 300, resulting in a change of amino acid
from glutamic acid to valine. A peptide representing the Chi-
nese strain polymorphism (v19var) was thus constructed to
determine whether this change affected CTL recognition. For
BALB/c and B10.D2 mice, v19-specific CTL (induced follow-
 IMMUNITY TO P. VIVAX CIRCUMSPOROZOITE PROTEIN
FIGURE 2. Number of responders by lymphoproliferative response to Plasmodium vivax circumsporozoite protein peptides and antigens in
non-exposed individuals (n 25), in the first malaria episode group (n 19), and in the multiple episodes group (n 24). See Materials and
Methods for definition of responder. TT tetanus toxoid; PPD purified protein derivative of Mycobacterium tuberculosis.
ing immunization with group IV) completely failed to recog-
nize v19var.
CTL activity against P. vivax peptides in recently exposed
and non-exposed Thais. Peptide-specific responses to all P.
vivax CSP peptides tested as pools were low. For the majority
of individuals studied, the net peptide-specific lysis was be-
tween 10% and +10%, as previously shown in a study of P.
falciparum-specific CTL.23 For the group I peptides (Figure
5), CTL from four of 53 individuals tested were greater than
15% (peptide-specific lysis). One individual responded to
group V peptides. There were no responses to any other
groups. While these may represent genuine peptide-specific
lytic activity, it was not possible to re-test these individuals,
who had left hospital and returned to their homes (all outside
Bangkok) by the time the results of the study were at hand.
While the positive results could not be confirmed, it is clear
that the frequency of CTL to P. vivax CSP at a population
level is very low. This low frequency is similar to the very low
frequency of P. falciparum CSP-specific CTL observed in
Caucasians, northwestern Thais, Kenyans, Gambian,37 and
Papua New Guineans.26
The CTL responses to the influenza matrix peptide were
generated in eight of 74 individuals. Although it was not pos-
sible to retest individuals, the data do suggest that at the
population level there was a low but finite frequency of in-
fluenza peptide-specific CTL. Although the influenza pep-
tide-specific response is low, it is similar to the low frequency
of HLA.A2.1 individuals in the Thai population,30 and may
also represent a lack of recent exposure to influenza. The
ability to generate CTL using this methodology both here and
previously5,26 give us certainty that our methodology is work-
ing.
Antibody response. Sera of P. vivax-exposed and non-
exposed subjects were tested for the presence of antibodies to
the classic and variant repeats of the P. vivax CSP (v14
and v30). The average OD405 + 3 SD of 34 non-exposed sera
was used as a cut-off value. None of the non-exposed subjects
had antibodies to either v14 or v30. However, unlike the pau-
33
city of T cell responsiveness to P. vivax CSP peptides, IgG-
specific antibody responses among P. vivax-exposed subjects
to the CSP were significant. The classic P. vivax repeat pep-
tide (v14) was recognized by 63% (12 of 19) of the sera from
individuals having their first episode when tested on admis-
sion, 68% (17 of 25) when tested on day 7, and 33% (8 of 24)
when tested on day 28 (Figure 6a). In multiply exposed indi-
viduals 50% (14 of 28) were positive on admission, 50% (15 of
30) on day 7, and 18% (6 of 34) on day 28.
The IgG-specific antibody response to the variant P. vivax
repeat peptide (v30) was 42% (8 of 19) in first exposed indi-
viduals on admission, 52% (13 of 25) at day 7, and 29% (7 of
24) at day 28 (Figure 6b). In multiply exposed individuals
60% (17 of 28) were positive at day 0, 53% (16 of 30) at day
7, and 50% (17 of 34) at day 28.
DISCUSSION
This study defines CSP-specific immune responses in indi-
viduals recently infected with P. vivax. Using an overlapping
set of synthetic 20-mer peptides, CD4+ and CD8+ T cell
epitopes were initially defined for different strains of mice.
LN cells from all strains of mice gave a very similar response
to peptides spanning two areas of the P. vivax CSP. The first
area is amino terminal to the conserved region (region I [RI]).
This region, particularly within the first 60 amino acids, is
polymorphic, with the polymorphism restricted to amino acid
positions 11, 13, 38, 49, and 52.17 The second area is carboxyl
terminal to the repeat region, in the vicinity of another con-
served region (region II [RII]). Again, this area is also poly-
morphic.16,17 One peptide (v21) was identified that was rec-
ognized by CD4+ T cells from all strains of mice studied. The
data concurred with a previous report identifying a peptide
with an overlapping sequence as being immunogenic in dif-
ferent mouse strains.22 From available sequence data, this
peptide (v21) has only one amino acid variation, V/A at po-
sition 322.17 This peptide also includes the conserved region
34 SUPHAVILAI AND OTHERS

FIGURE 3. a, Cytotoxic T lymphocyte (CTL) activity in BALB/c mice against pooled Plasmodium vivax peptides and P. falciparum peptide
(f97). b, CTL activity in B10.D2 mice against pooled P. vivax peptides and P. falciparum peptide (f97). All assays was performed at a 50:1
effector:target ratio.

RII, which is thought to be important in hepatocyte inva- ticular merit in a P. vivax CSP vaccine by increasing the
sion.3840 While not recognized by acutely/very recently in- chance that immunity will be boosted by natural infection.
fected Thais, v21 is known to be recognized by up to 40% of Further investigation of the v21 variant is required.
Caucasians previously exposed to P. vivax.23 It may have par- For murine CD8+ T cells, the response was restricted to one
 IMMUNITY TO P. VIVAX CIRCUMSPOROZOITE PROTEIN 35

FIGURE 4. Proliferative responses to P30 (a) and cytotoxic T lymphocyte responses (b). In the lymphoproliferative studies, CD4+ and CD8+
T cells were depleted post-immunization but prior to assay.

H-2 haplotype (H-2d). Since CTL of BALB/c mice responded
to more epitopes than B10.D2 mice and both strains are of the
H-2d haplotype, it is likely that non-H-2 genes are contribut-
ing to immunologic responsiveness. Similar observations have
been previously reported.35 The CTL responses in BALB/c
B10.D2 F1 mice were found to be similar to those in BALB/c
mice. This suggests that dominant negative regulatory ele-
ments on the B10 background are not responsible for the lack
of responsiveness of B10.D2 mice. It is also likely that non-
HLA genes may be contributing to the human low CTL im-
munologic responsiveness to the CSP. The CD8+ T cell
epitopes can often be predicted from motifs. Although this
approach was not used here to identify epitopes, an analysis
of the identified epitopes with respect to motifs has been
undertaken. A class I anchor for H-2KdDd cannot be defined
on v19 peptide, compared with MHC ligands and peptide
motifs listed.41 However, one possible motif on the v19 pep-
tide was identified for H-2Ld (APNEKSVKVYL, where an-
chor residues are P at position 2 and L at position 11).
Possible MHC class I motifs on v1 and v2 were identified in
this study. The preferred H-2Kd anchor residue at position 2
is F and that at position 9 is I.41 Peptide v1 has F at position
2 and I at position 10. Possible anchor residues on v2 are F at
position 2 and L at position 12 (LFPTHCGHNVDL). Peptide
5 in group I pool elicited some limited CTL activity and a
possible motif could not be identified. A possible class I motif
on v21 could not be identified. For v25, possible H-2Kd motif
anchor residues are F at position 2 and L at position 9.
Peptide v19 has three known amino acid polymorphisms,
P/T (proline to threonine), E/V (glutamic acid to valine), and
K/I (lysine to isoleucine),17 and one of these polymorphisms
(E/V) was found to completely abrogate CTL recognition.
Cross-recognition between the variants requires further
study; however, it is tempting to consider that polymorphisms

FIGURE 5. Cytotoxic T lymphocyte activity against group I Plasmodium vivax circumsporozoite protein peptides in non-exposed- and P.
vivax-exposed individuals at different time points. There were 28 subjects in the non-exposed (NE) group, 53 subjects studied at day seven after
admission (D7) group, 58 subjects studied at day 28 after admission (D28) group, and 19 subjects studied at follow-up (F) group.
36 SUPHAVILAI AND OTHERS

FIGURE 6. Antibody response to a, Plasmodium vivax classic repeat peptide (v14) and b, P. vivax variant repeat peptide (v30) in non-exposed
(NE), single and multiply exposed individuals at a serum dilution of 1:100 at different time points. Data are expressed as net optical density at
405 nm (OD405) (see Materials and Methods). Some symbols represent multiple points (see Results for percent of positive responses for each
group). For definitions of other abbreviations, see Figure 5.

have been selected by P. vivax CSP-specific CTL, as has been
postulated to occur with P. falciparum CSP.8 Clearly, how-
ever, further consideration of this hypothesis requires confir-
mation that human CTL recognize v19. Human proliferative
T cells (type undefined) have been shown to commonly rec-
ognize v19,23 although the ability of the polymorphisms to
ablate recognition of human T cells has not been determined.
It was of interest that P30 (from TT, 947-967, FNNFTVS-
FWLRVPKVSASHLE), as well as functioning as a promis-
cuous helper epitope (as previously described34), was also
seen as a cytotoxic T epitope in both H-2d and H-2b mice. P30
contains possible anchor residues for defined motifs in both
strains. There are two possible sets of H-2K d anchors
(NFTVSFWL, and SFWLRVPKV) and one H-2Kb anchor
(FTVSFWLRV).41
The reasons for the low CD8+ responsiveness are unknown.
However, we have previously shown that the frequency of
CTL responses to epitopes of the P. falciparum CSP was very
low among visitors to and residents of endemic areas, and that
recent exposure to sporozoites (within two years) was a pre-
requisite for detection of specific CTL.5 A study by Malik and
others6 demonstrated that P. falciparum CSP-specific CTL
were detected in four of five recently and heavily exposed
volunteers. Given that the subjects in the study presented
here were not heavily exposed, it appears that in general,
recent and heavy exposure of humans may be required if
CSP-specific CTL are to be present in sufficiently high fre-
quency to be detected in the in vitro assays used in this and
the previous studies. Other factors referred to in different
studies42,43 have also been postulated to affect CSP-specific T
cell responsiveness, including paucity of epitopes within the
CSP and polymorphism within epitopes resulting in limited
exposure to any particular epitope.
In a recent Colombian study,44 three HLA 0201-restricted
CD8+ epitopes were defined on the P. vivax CSP. These were
located in the carboxyl terminal region of the protein corre-
 IMMUNITY TO P. VIVAX CIRCUMSPOROZOITE PROTEIN
sponding to residues 301-309, 365-374, and 341-349, closely
overlapping some of the regions recognized by the murine
CD8+ T cells in our study. Although our human volunteers
did not recognize these epitopes, this may relate to the fact
that the patients in our study were recently infected and para-
sitemic at the time of our study, whereas the Columbian vol-
unteers were not, but had 20 years of permanent exposure.
The use of tetramers could enable a more detailed analysis of
the basis for poor CD8 T cell responsiveness.
The low level of human lymphoproliferative responsiveness
was unexpected. We previously showed23 that both malaria-
exposed and non-exposed Caucasians responded to multiple
epitopes within the P. vivax CSP, but that the exposed donors
responded uniquely to a subset of epitopes. We also found
that responses were more frequently found in those individu-
als more recently exposed. In a study in Karen Thais (an
ethnic population different from the Thais in this study33), we
also demonstrated significantly more responsiveness using
similar methodology to that used here. An important differ-
ence, however, between the previous studies and the study
reported here was that the two previous studies did not ex-
amine responsiveness in acutely and very recently infected
subjects. Given that the same peptide sequences were studied
and that the methodologies of all three studies were very
similar, it appears that peptide-specific CD4+ T cells are not
present in the peripheral blood of individuals acutely infected
in numbers sufficient to detect in the assay used here. Fur-
thermore, since both previous studies demonstrated that re-
sponses were found in non-exposed donors (postulated to be
the result of immunologic cross-reactivity between CSP and
environmental organisms),45 as well as exposed donors, the
lack of responsiveness observed here is likely not to be due to
insufficient time since exposure to generate a response. This
hypothesis is supported by the observation that P. vivax CSP-
specific antibodies were readily detected in a number of in-
dividuals tested. A likely explanation for the data observed
here is that CSP-specific T cells are present in a location other
than the peripheral blood. An obvious site would be the liver.
In the murine model, CSP-specific protective CTL have been
shown to home to the liver.46 It is worth noting that in a
recent Brazilian study proliferative responses to recombinant
CSP were observed in 50% of the adults not continuously
exposed to malaria, but in only 10% of subjects from a trans-
mission area.47 As in our study, antibody responses were simi-
lar in both groups. These studies are very complementary and
point to the likely effect of recent infection on immunomodu-
lation/lymphocyte trafficking, and also highlighted the im-
portance of timing of blood collection when analyzing T cell
responses. Many of the factors postulated to be relevant to
human T cell non responsiveness (such as polymorphism, sup-
pression, and sequestering of T cells in the liver) would not be
operative in mice following peptide vaccination.
The reasons for the decrease in levels of CSP-specific an-
tibodies following recrudescence/ relapse are unknown. How-
ever, it has been extensively reported that infection is associ-
ated with immunosuppression,4850 and it has been demon-
strated in rodents5153 that malaria infection can cause
apoptosis of specific T cells, and in humans of mononuclear
cells.54
This study of CSP-specific immune responses in individuals
recently infected with P. vivax points to an early engagement
of the immune system following infection. A long-term fol-
37
low-up of individuals post-infection may give important in-
sights into the kinetics of CSP-specific immune responses rel-
evant to protection, which will assist our progress in vaccine
development.
In summary, the results demonstrate the presence of a sig-
nificant number of murine T cell epitopes on the P. vivax CSP
and a paucity of human T cell responses, possibly as a result
of recent infection. In other systems, including the P. falci-
parum CSP5,9,24,25,27 and the influenza nucleoprotein,29,54,55
there has been a close overlap of murine and human T cell
epitopes, both for CD4+ and CD8+ T cell responses. This
study may then point to the location of human CD8+ T cell
epitopes. It is of concern, however, but not totally unex-
pected, that of the limited amount of described polymor-
phisms, variation within P. vivax CSP,16,17 a CTL epitope
(v19) was shown to be a site of major polymorphisms (P/T,
E/V and K/I; 11; see also Table 1) and that one of the v19
polymorphisms tested was able to abrogate CTL recognition.
A detailed knowledge of the specificity and development of
CSP-specific immune responses and the likely location of
epitopes not readily defined by studying humans naturally
exposed to malaria may help in the ultimate design of vac-
cines incorporating part or all of the CSP.
Received April 2, 2003. Accepted for publication October 29, 2003.
Acknowledgments: We thank the Department of Immunology,
Armed Force Research Institute of Medical Sciences (Bangkok,
Thailand) for providing tissue culture facilities for the work done in
Bangkok. We also thank Dr. Henry Stephens for very helpful infor-
mation and discussions on HLA class I in Thai population.
Authors addresses: Chaisuree Suphavilai, Research Institute for
Health Sciences, PO Box 80CMU, Chiang Mai University, Chiang
Mai 50202, Thailand. Sornchai Looareesuwan, The Tropical Medi-
cine Hospital, Faculty of Tropical Medicine, Mahidol University,
Bangkok 10400, Thailand. Michael F. Good, The Cooperative Re-
search Centre for Vaccine Technology, The Queensland Institute of
Medical Research, 300 Herston Road, Herston QLD 4006, Queens-
land, Australia, Telephone: 61-7-3362-0203, Fax: 61-7-3362-0110,
Email: michaelG@qimr.edu.au.
REFERENCES
1. Good MF, Doolan DL, 1999. Immune effector mechanisms in
malaria. Curr Opin Immunol 11: 412419.
2. Tsuji M, Romero P, Nussenzweig RS, Zavala F, 1990. CD4+ cy-
tolytic T cell clone confers protection against murine malaria.
J Exp Med 172: 13531357.
3. Weiss WR, Sedegah M, Beaudoin RL, Miller LH, Good MF,
1988. CD8+ T cells (cytotoxic/suppressors) are required for
protection in mice immunized with malaria sporozoites. Proc
Natl Acad Sci USA 85: 573576.
4. Hoffman SL, Oster CN, Mason C, Beier JC, Sherwood JA, Bal-
lou WP, Mugambi M, Chulay JD, 1989. Human lymphocyte
proliferative response to a sporozoite T cell epitope correlates
with resistance to falciparum malaria. J Immunol 142: 1299
1303.
5. Doolan DL, Houghten RA, Good MF, 1991. Location of human
cytotoxic T cell epitopes within a polymorphic domain of the
Plasmodium falciparum circumsporozoite protein. Int Immu-
nol 3: 511516.
6. Malik A, Egan JE, Houghten RA, Sadoff JC, Hoffman SL, 1991.
Human cytotoxic T lymphocytes against the Plasmodium fal-
ciparum circumsporozoite protein. Proc Natl Acad Sci USA 88:
33003304.
7. Dontfraid F, Cochran MA, Pombo D, Knell JD, Quakyi IA, Ku-
mar S, Houghten RA, Berzofsky JA, Miller LH, Good MF,
1988. Human and murine CD4 T cell epitopes map to the same
region of the malaria circumsporozoite protein: limited immu-
38 SUPHAVILAI AND OTHERS
nogenicity of sporozoites and circumsporozoite protein. Mol
Biol Med 5: 185196.
8. Good MF, Kumar S, Weiss WR, Miller LH, 1989. T-cell antigenic
sites of the malaria circumsporozoite protein. Blood 74: 895
900.
9. Kumar S, Miller LH, Quakyi IA, Keister DB, Houghten RA,
Maloy WL, Moss B, Berzofsky JA, Good MF, 1988. Cytotoxic
T cells specific for the circumsporozoite protein of Plasmo-
dium falciparum. Nature 334: 258260.
10. Avrameas S, Ternynck T, 1969. The cross-linking of proteins with
glutaraldehyde and its use for the preparation of immunosor-
bents. Immunochemistry 6: 5366.
11. Rosenberg R, Wirtz RA, Lanar DE, Sattabongkot J, Hall T,
Waters AP, Prasittisuk C, 1989. Circumsporozoite protein het-
erogeneity in the human malaria parasite Plasmodium vivax.
Science 245: 973976.
12. Franke ED, Lucas CM, Chauca G, Wirtz RA, Hinostroza S, 1992.
Antibody response to the circumsporozoite protein of Plasmo-
dium vivax in naturally infected humans. Am J Trop Med Hyg
46: 320326.
13. Park CG, Chwae YJ, Kim JI, Lee JH, Hur GM, Jeon BH, Koh JS,
Han JH, Lee SJ, Park JW, Kaslow DC, Strickman D, Roh CS,
2000. Serologic responses of Korean soldiers serving in ma-
laria-endemic areas during a recent outbreak of Plasmodium
vivax. Am J Trop Med Hyg 62: 720725.
14. Franke ED, Lucas CM, Roman San E, 1991. Antibody response
in humans to the circumsporozoite protein of Plasmodium
vivax. Infect Immun 59: 28362838.
15. Zavala F, Cochrane AH, Naedin EH, Nussenzweig RS, Nussen-
zweig V, 1983. Circumsporozoite proteins of malaria parasites
contain a single immunodominant region with two or more
identical epitopes. J Exp Med 157: 19471957.
16. Qari SH, Goldman IF, Povoa MM, Oliveira S, Alpers MP, Lal
AA, 1991. Wide distribution of the variant form of the human
malaria parasite Plasmodium vivax. J Biol Chem 266: 16297
16300.
17. Qari SH, Goldman IF, Povoa MM, di Santi S, Alpers MP, Lal
AA, 1992. Polymorphism in the circumsporozoite protein of
the human malaria parasite Plasmodium vivax. Mol Biochem
Parasitol 55: 105113.
18. Charoenvit Y, Collins WE, Jones TR, Millet P, Yuan L, Camp-
bell GH, Beaudoin RL, Broderson JR, Hoffman SL, 1991.
Inability of malaria vaccine to induce antibodies to a protective
epitope within its sequence. Science 251: 668671.
19. Collins WE, Sullivan JS, Morris CL, Galland GG, Jue DL, Fang
S, Wohlhueter R, Reed RC, Yang C, Hunter RL, Lal AA,
1997. Protective immunity induced in squirrel monkeys with a
multiple antigen construct against the circumsporozoite pro-
tein of Plasmodium vivax. Am J Trop Med Hyg 56: 200210.
20. Thomas BE, Sridevi K, Chopra N, Haq W, Rao DN, 2001. In-
ducing a cell-mediated immune response against peptides of
the Plasmodium vivax circumsporozoite protein. Ann Trop
Med Parasitol 95: 573586.
21. Nardin E, Clavijo P, Mons B, van Belkum A, Ponnudurai T,
Nussenzweig RS, 1991. T cell epitopes of the circumsporozoite
protein of Plasmodium vivax. Recognition by lymphocytes of a
sporozoite-immunized chimpanzee. J Immunol 146: 1674
1678.
22. Rodrigues MM, Paiva AC, Dutra AP, Yoshida N, Nakaie C,
1991. Identification of epitopes within the circumsporozoite
protein of Plasmodium vivax recognized by murine T lympho-
cytes. Exp Parasitol 72: 271277.
23. Bilsborough J, Carlisle M, Good MF, 1993. Identification of Cau-
casian CD4 T cell epitopes on the circumsporozoite protein of
Plasmodium vivax. J Immunol 151: 890899.
24. Herrera S, Escobar P, de Plata C, Avila GI, Corradin G, Herrera
MA, 1992. Human recognition of T cell epitopes on the Plas-
modium vivax circumsporozoite protein. J Immunol 148: 3986
3990.
25. Zevering Y, Houghten RA, Frazer IH, Good MF, 1990. Major
population differences in T cell response to a malaria sporo-
zoite vaccine candidate. Int Immunol 2: 945955.
26. Doolan DL, Khamboonruang C, Beck H-P, Houghten RA, Good
MF, 1993. Cytotoxic T lymphocyte (CTL) low-responsiveness
to the Plasmodium falciparum circumsporozoite protein in
naturally-exposed endemic populations: analysis of human
CTL response to most known variants. Int Immunol 5: 3746.
27. Blum-Tivouvanziam U, BeghdadiRais C, Roggero MA, Valmori
D, Bertholet S, Bron C, Fasel N, Corradin G, 1994. Elicitation
of specific cytotoxic T cells by immunization with malaria
soluble synthetic polypeptides. J Immunol 153: 41344141.
28. Houghten RA, 1985. General method for the rapid solid-phase
synthesis of large numbers of peptides: Specificity of antigen-
antibody interaction at the level of individual amino acids.
Proc Natl Acad Sci USA 82: 51315135.
29. Gotch F, Rothbard J, Howland K, Townsend A, McMichael A,
1987. Cytotoxic T lymphocytes recognize a fragment of influ-
enza virus matrix protein in association with HLA-A2. Nature
326: 881882.
30. Vejbaesya S, Chantangpol R, Longta P, Chandanayingyong D,
1997. HLA class I typing by one-dimensional isoelectric focus-
ing and identification of the new variants in Thai population.
Asian Pac J Allergy Immunol 15: 2127.
31. Chiewsilp P, Mongkolsuk T, Sujirachato K, 1997. A*02 in south-
ern Thai Muslims and central Thais. J Med Assoc Thai 80:
S25S29.
32. Valmori D, Romero JF, Men Y, Maryanski JL, Romero P, Cor-
radin G, 1994. Induction of a cytotoxic T cell response by
co-injection of a T helper peptide and a cytotoxic T lympho-
cyte peptide in incomplete Freunds adjuvant (IFA): further
enhancement by pre-injection of IFA alone. Eur J Immunol
24: 14581462.
33. Zevering Y, Khamboonruang C, Rungruengthanakit K, Tungvi-
boonchai L, Ruengpipattanapan J, Bathurst I, Barr P, Good
MF, 1994. Life-span of human T-cell responses to determi-
nants from the circumsporozoite proteins of Plasmodium fal-
ciparum and Plasmodium vivax. Proc Natl Acad Sci USA 91:
61186122.
34. Panina-Bordignon P, Tan A, Termijtelen A, Demotz S, Corradin
G, Lanzavecchia A, 1989. Universally immunogenic T cell
epitopes: promiscuous binding to human MHC class II and
promiscuous recognition by T cells. Eur J Immunol 19: 2237
2242.
35. Weiss WR, Good MF, Hollingdale MR, Miller LH, Berzofsky
JA, 1989. Genetic control of immunity to Plasmodium yoelii
sporozoites. J Immunol 143: 42634266.
36. Huang T, Cheng Q, Allan S, Huang Y, 1994. DNA sequencing of
circumsporozoite protein genes of Plasmodium vivax from
four different countries in west Pacific region: comparative
study on the flank sequences. Zhongguo Ji Sheng Chong Xue
Yu Ji Sheng Chong Bing Za Zhi 12: 8592.
37. Lalvani A, Hurt N, Aidoo M, Kibatala P, Tanner M, Hill AV,
1996. Cytotoxic T lymphocytes to Plasmodium falciparum
epitopes in an area of intense and perennial transmission in
Tanzania. Eur J Immunol 26: 773.
38. Aley SB, Bates MD, Tam JP, Hollingdale MR, 1986. Synthetic
peptides from the circumsporozoite protein of Plasmodium
falciparum and Plasmodium knowlesi recognize the human
hepatoma cell line HepG2-A16 in vitro. J Exp Med 164: 1915
1922.
39. Pancake SJ, Holt GD, Mellouk S, Hoffman SL, 1992. Malaria
sporozoites and circumsporozoite proteins bind specifically to
sulfated glycoconjugates. J Cell Biol 117: 13511357.
40. Rich KA, George FW IV, Law JL, Martin WJ, 1990. Cell-
adhesive motif in region II of malarial circumsporozoite pro-
tein. Science 249: 15741577.
41. Rammensee HG, Friede T, Stevanovic S, 1995. MHC ligands and
peptide motifs: first listing. Immunogenetics 41: 178228.
42. Quakyi IA, Currier J, Fell A, Taylor DW, Roberts T, Houghten
RA, England RD, Berzofsky JA, Miller LH, Good MF, 1994.
Analysis of human T cell clones specific for conserved peptide
sequences within malaria proteins. Paucity of clones respon-
sive to intact parasites. J Immunol 153: 20822092.
43. Zevering Y, Khamboonruang C, Good MF, 1994. Effect of poly-
morphism of sporozoite antigens on T-cell activation. Res Im-
munol 145: 469476.
44. Arevalo-Herrera M, Valencia AZ, Vergara J, Bonelo A, Fleis-
chhauer K, Gonzalez JM, Restrepo JC, Lopez JA, Valmori D,
Corradin G, Herrera S, 2002. Identification of HLA-A2 re-
stricted CD8(+) T-lymphocyte responses to Plasmodium vivax
 IMMUNITY TO P. VIVAX CIRCUMSPOROZOITE PROTEIN
circumsporozoite protein in individuals naturally exposed to
malaria. Parasite Immunol 24: 161169.
45. Zevering Y, Amante F, Smillie A, Currier J, Smith G, Houghten
RA, Good MF, 1992. High frequency of malaria-specific T
cells in non-exposed humans. Eur J Immunol 22: 689696.
46. Rodrigues M, Nussenzweig RS, Romero P, Zavala F, 1992. The in
vivo cytotoxic activity of CD8+ T cell clones correlates with
their levels of expression of adhesion molecules. J Exp Med
175: 895905.
47. Braga EM, Carvalho LH, Fontes CJ, Krettli AU, 2002. Low cel-
lular response in vitro among subjects with long-term exposure
to malaria transmission in Brazilian endemic areas. Am J Trop
Med Hyg 66: 299303.
48. Goonewardene R, Carter R, Gamage CP, Del Giudice G, David
PH, Howie S, Kendis KN, 1990. Human T cell proliferative
responses to Plasmodium vivax antigens: evidence of immu-
nosuppression following prolonged exposure to endemic ma-
laria. Eur J Immunol 20: 13871391.
49. Ho M, Webster HK, Looareesuwan S, Supanaranond W, Phillips
RE, Chanthavanich P, Warrell DA, 1986. Antigen-specific im-
munosuppression in human malaria due to Plasmodium falci-
parum. J Infect Dis 153: 763771.
50. Troye-Blomberg M, Romero P, Patarroyo ME, Bjorkman A,
Perlmann P, 1984. Regulation of the immune response in Plas-
39
modium falciparum malaria. III. Proliferative response to an-
tigen in vitro and subset composition of T cells from patients
with acute infection or from immune donors. Clin Exp Immu-
nol 58: 380387.
51. Hirunpetcharat C, Good MF, 1998. Deletion of Plasmodium
berghei-specific CD4+ T cells adoptively transferred into re-
cipient mice following challenge with homologous parasite.
Proc Natl Acad Sci USA 95: 17151720.
52. Wipasa J, Xu H, Stowers A, Good MF, 2001. Apoptotic deletion
of helper T cells specific for the 19 kDa carboxyl terminal
fragment of merozoite surface protein 1 (MSP119) during ma-
laria infection. J Immunol 167: 39033909.
53. Xu H, Wipasa J, Yan H, Zeng M, Makobongo MO, Finkelman
FD, Kelso A, Good MF, 2002. The mechanism and significance
of deletion of parasite-specific CD4+ T cells in malaria infec-
tion. J Exp Med 195: 881892.
54. Toure-Balde A, Sarthou JL, Aribot G, Michel P, Trape JR, Ro-
gier C, Roussilhon C, 1996. Plasmodium falciparum induces
apoptosis in human mononuclear cells. Infect Immun 64: 744
750.
55. Gould K, Cossins J, Bastin J, Brownlee GG, Townsend A, 1989.
A 15 amino acid fragment of influenza nucleoprotein synthe-
sized in the cytoplasm is presented to class I-restricted cyto-
toxic T lymphocytes. J Exp Med 170: 10511056.