The circumsporozoite protein of Plasmodium:

a mechanism of immune evasion by the malaria

parasite?
L. Schofield1

Sporozoites of malaria are covered with a repetitive surface antigen, the circumsporozoite (CS) protein. This
antigen also appears to be a major target of the host immune response. The natural immunogenicity of the
CS protein has led to attempts to develop the molecule as a vaccine candidate. It seems paradoxical,
however, that a successful parasite should present to the host an immunogenic surface molecule which
would induce protective immunity. In this paper we suggest that the CS protein is not the target of protective
immunity under natural conditions, and that naturally immunogenic repetitive antigens in malaria and other
parasites have evolved as a mechanism of immune evasion, via the induction of thymus-independent B-cell
responses.

Introduction is helpful to review this issue before turning to the

In the decade since the first identification of the
circumsporozoite (CS) protein of malaria sporozoites
(51), there have been a large number of studies de-
voted to various aspects of the biology, genetics, and
immunology of this protein. Included in this body of
work have been several attempts to vaccinate against
the parasite using CS analogues. Indeed, the hope that
the CS will serve as an anti-malarial vaccine is the
primary reason for the amount of attention given to
this molecule. Despite this considerable research ef-
fort, however, an essential question remains un-
answered, namely "what is the functional role of the
CS protein in the natural history of the parasite?" An
answer to this question is surely a basic prerequisite
for the rational, as opposed to empirical, approach to
vaccine design.
In this paper, we will review briefly some of the
salient features of the biology and immunology of the
CS protein, and describe how this molecule came to
be chosen as a vaccine candidate. We will then present
some novel data that arose from our attempts to
investigate the regulation of the immune response to
the CS protein. These data support the conclusion
that one primary role of the molecule is to act as a
mechanism of immune evasion for the parasite.
The proposition that the CS protein may func-
tion as a mechanism of immune evasion is strongly at
odds with the suggestion that this antigen is the target
of naturally acquired immunity to malaria. However,
there is little evidence to support the latter view, and it
1 Department of Medical and Molecular Parasitology, New York
University School of Medicine, New York, NY 10010, USA.
question of the function of the protein in the natural
history of the parasite.
Do sporozoites induce protective
immunity?
The solid protective immunity induced under certain
experimental conditions by irradiated sporozoites is
not a feature of all host/malaria parasite systems. It is
best described in two rather unusual host/parasite
models. In more natural host/parasite combinations,
the protection is more difficult to obtain. For ex-
ample, exposure of chickens to ultra-violet irradiated
Plasmodium gallinaceum sporozoites successfully im-
munized against subsequent viable challenge (34).
These observations were later extended to X-irradi-
ated sporozoites of P. berghei in mice (35). However,
this level of protection was considerably harder to
achieve in comparable immunization studies using
human, simian or other rodent malaria sporozoites.
Ward and Hayes (1972) immunized three rhesus mon-
keys (Macaca mulatta) against P. cynomolgi by three
rounds of bites of X-irradiated infected mosquitos
over 42 days, which failed to provide any degree of
protection against infective bite (47). Similarly, two
rhesus monkeys immunized with P. cynomolgi were
not protected upon challenge (9). In another study,
12 rhesus monkeys immunized intravenously with
P. cynomolgi were challenged when high levels of cir-
cumsporozoite precipitin antibodies were detected in
their sera. Of these animals, 2 resisted challenge and
10 were susceptible (5). After extensive immunization
of human volunteers by multiple exposure to irradi-
ated sporozoites of P. falciparum, more than half
remained susceptible to the bite of infected mosquitos,

66 Bulletin of the World Healih Organization, 66 (Suppl.): 6643 (190)


while only two of five could be protected against
P. vivax (7).
Of great importance is the recent finding that
some rodent strains cannot be successfully immunized
with irradiated sporozoites of P. yoelii (50). Both
MHC and other loci play a role in determining the
acquisition of immunity. This is in marked contrast to
the P. berghei/mouse model, where functional immun-
ity may be obtained in all mouse strains tested (25).
This difference apparently is not due to varying de-
grees of immunogenicity exhibited by the sporozoites
of each species. Furthermore, it is unlikely that such
similar sporozoites will prime very different immune
effector mechanisms. Rather, it is most likely that this
reflects instead the very different degrees of natural
infectivity of the sporozoites for the host. It is clear
that P. berghei has a very low degree of infectivity for
laboratory mice as compared with P. yoelii, which, for
example, gives rise to pre-erythrocytic liver schizonts
easily detectable by microscopy. The difference in
infectivity may be as great as two logs. Thus there are
good grounds for proposing that acquired immunity
operates best in circumstances where a high degree of
innate immunity produces conditions inimicable to
the parasite; in contrast, when natural infectivity of
sporozoites for their host is high, acquired immunity
is subject to threshold effects and the pattern of
protection will be variable, being under polygenic
control.
The proposition that solid immunity operates
best when innate immunity is already high is obvious
and unremarkable. However, it does seem that these
conditions apply in the host/parasite model upon
which much anti-sporozoite vaccine research is based,
namely the P. berghei/laboratory mouse model.
Those host/parasite combinations in which acquired
immunity is more difficult to obtain appear to re-
semble more closely the balance as it is found under
natural conditions, i.e., they are the true co-adapted
host/parasite pair, or at least provide more accurate
models of pre-erythrocytic malaria. This raises the
question of how far the irradiated P. berghei sporo-
zoite type of protective immunity may be extended to
humans. The limited evidence so far indicates that
exposure of humans to sporozoites provides con-
siderably lower protection than that obtained in the
P. berghei/mouse model.
The question of whether sporozoites induce any
degree of protective immunity under conditions of
natural transmission is difficult to answer conclu-
sively. Several studies suggest that naturally acquired
antibodies are non-neutralizing (26, 31). Even adults
chronically exposed to natural transmission appear
to be susceptible to new infections, although they are
capable of controlling the infection at the blood-stage
level. It is generally considered that naturally acquired
WHO Bulletin OMS: Supplement Vol. 68 1990
The circumsporozolte protein of Plasmodlum
immunity does not operate against sporozoites, al-
though this must remain somewhat moot until put to
experimental test.
Does the CS protein induce
protective immunity?
Notwithstanding the limited suitability of the
P. berghei/laboratory mouse system as a model
for acquired immunity in the human malarias (and
other co-adapted host/parasite combinations), it has
nonetheless been worthwhile to dissect the functional
immunity obtained in this model as a first step in
vaccine research. Although providing considerable
proof of the natural immunogenicity of the CS pro-
tein, there has been little direct evidence from this or
other models demonstrating that the CS protein is
responsible for inducing protective immunity against
sporozoites. For example, multiple immunizations
with attenuated sporozoites lead to high levels of
antibody, directed almost exclusively to the repetitive
domain of the CS surface antigen. Nonetheless, the
high titre of anti-repeat antibody obtained in this way
is ineffectual in neutralizing parasites. Sporozoites
escape the antibody and penetrate the liver, where
protective immunity to the parasite can result from
cellular responses (41, 49). Indeed, parasite-induced
antibody is invariably ineffective in neutralizing para-
sites when passively transferred into naive recipients
(36, 41).
When considering the biological function of the
CS protein, and the involvement of this antigen in
natural or experimentally acquired immunity, it is
essential to avoid data derived from experiments that
are fundamentally irrelevant to the natural history of
the parasite. For example, pre-incubation of infectious
sporozoites for 30 minutes or longer periods with
anti-CS antibodies will inhibit their ability to penet-
rate hepatocytic target cells in vitro, and neutralize
their infectivity for susceptible hosts. These antibod-
ies, however, are routinely unable to provide pro-
tection when passively transferred into the same hosts.
Thus, although some antibody preparations may be
sufficient to neutralize sporozoite infectivity under
certain experimental conditions, this evidence cannot
be extrapolated to imply that the CS protein is a
target of protective humoral immunity under condi-
tions of natural exposure to the parasite. Similarly,
whereas monoclonal antibodies directed towards the
repetitive domain of the P. berghei CS protein can
confer protection in passive transfer (39), it should be
clear that this represents a situation not truly found
during the natural immune response to sporozoites.
Of course, these conditions may be those we wish to
achieve for a successful vaccine, and the failure of
polyclonal antibodies derived against the parasite to
67
L. Schofield
neutralize infectivity in vivo does not necessarily in-
validate the hope for an effective humoral vaccine.
However, the data clearly indicate that under natural
conditions of exposure to malaria, and even after
immunization with irradiated sporozoites, the CS
protein does not induce neutralizing antibodies.
The immunity that does result from immuniza-
tion with irradiated P. berghei sporozoites is clearly
directed towards the developing pre-erythrocytic
schizonts, and may be mediated by cytotoxic T cells
(41, 49), lymphokines (41), or other undescribed mech-
anisms (42). It is not known if the CS protein is either
the target or trigger of this effector phase. Apparently,
large numbers of CS-specific CD8+ T cells are suffi-
cient to provide protection against P. berghei upon
adoptive transfer into the somewhat resistant white
mouse host (40). However, this T-cell clone does not
protect against the more infectious P. yoelii sporo-
zoites although there can be cross-reactivity with
the corresponding P. yoelii sequence (40). As outlined
above, protection can be obtained in a similar fashion
by transfer of monoclonal antibodies directed to-
wards the repetitive domain of the CS protein.
Clearly, these observations cannot be taken as dem-
onstrating that cytotoxic T cells or antibodies direc-
ted against the CS protein provide protection against
malaria under natural conditions or following experi-
mental exposure to irradiated sporozoites.
The function of the repetitive domain:
an hypothesis
Included within the general issue of the role of the CS
protein is the more specific question of the function
of the immunodominant repetitive domains found
within the molecule. The role of conserved, tandemly
repetitive regions within protein antigens poses an
unsolved question that applies to a multitude of the
higher eukaryotic parasites such as Plasmodium,
Toxoplasma, Leishmania, Trypanosoma, Schistosoma,
etc., which share in common this unusual feature. As
observed in a recent review, of 30 malarial antigens
sequenced to date, 29 contained tandem repeats (48).
These repetitive domains are strong B-cell epi-
topes, being often the only antigenic sites on these
molecules recognized by the humoral immune system
of the host. The repetitive (NANP). sequence of the
falciparum CS protein is able to absorb the entire
serological reactivity of human immune sera against
the sporozoite stage of the parasite, demonstrating
that this region alone is the single immunodominant
epitope (53). Indeed, the natural immunogenicity of
these antigens appears to be the major reason for their
selection as vaccine candidates. Notwithstanding the
repeated failure of polyclonal antisera to transfer
immunity from donor to recipient (36, 41), these same
s8
antibodies or their monoclonal derivatives have been
used to identify antigens expressed in recombinant
DNA libraries. These naturally immunogenic mole-
cules were subsequently developed as vaccine candid-
ates, under the assumption that antibodies would
confer protective immunity. Thus there are currently
attempts to immunize against malaria by utilizing
synthetic peptide or recombinant protein analogues
of the repetitive domains of the CS protein and the
ring-infected erythrocyte surface antigen (3, 8, 24).
The suggestion that repetitive antigens may in-
duce neutralizing antibodies raises a central paradox:
why would the parasite readily display on its surface
highly immunogenic antigens which might induce
parasiticidal antibody? This question is closely related
to one more fundamental: what is the functional role,
or selective (Darwinian) rationale of repetitive anti-
gens in Plasmodium and other parasites? One hypo-
thesis postulates that each unit of a repetitive domain
acts as a ligand for a host protein, the multimeric
nature of the repeats serving to increase greatly the
affinity of the interaction (37). However, (a) struc-
turally distinct repeats occur among proteins that are
phylogenetic homologues, despite the fact that they
must perform similar functions; (b) the hypothesis
does not explain the immunogenicity of the repeats;
and (c) the central requirement of this hypothesis
remains unfulfilled, because despite extensive
investigation no evidence exists that repeat regions
are ligands for host cells such as hepatocytes or
erythrocytes.
Clearly, questions concerning (a) the functional
role of these repetitive domains, (b) the reasons under-
lying their immunodominant nature, and (c) whether
their interaction with the host immune system leads
to protective immunity, assume central importance,
given the predominance of such elements and their
candidacy as malaria vaccines. We suggest that recent
experimental findings, outlined below, might provide
some answers to these questions, and provide a plaus-
ible selective (Darwinian) rationale for the presence of
repetitive domains in protein antigens of malaria and
other eukaryotic parasites. These findings, which sug-
gest that these vaccine candidates may actually func-
tion as a novel mechanism of immune evasion, arose
from the elucidation of the pathway of antibody
production to the repeats, as follows:
The anamnestic antibody response to synthetic
peptides and most proteins is under Ir gene control.
Following the studies of Chestnut & Grey, which
demonstrate the ability of B cells to serve as antigen-
presenting cells, the requirements for hapten-carrier
linkage and MHC-restricted physical interaction
among lymphocytes during antibody production is
thought to result from the ability of antigen-specific B
cells to take up antigen by surface Ig, and present
WHO Bulletin OMS: Supplement Vol. 68 1990

the relevant immunogenic peptides to antigen-specific
helper T cells in the context of self MHC (4). This
framework is assumed to apply in the case of antibody
responses to the repetitive domains of malarial anti-
gens. The major impetus for accepting this view was
the finding that a synthetic peptide corresponding to
the immunodominant (NANP). repeat sequence of
the P. falciparum CS protein was a poor inducer of
antibody formation in human volunteers for vaccina-
tion. It was also shown that all mice except those
bearing the H2b haplotype were unable to respond to
the (NANP). peptide (12, 15). This non-responsive-
ness resulted from the failure of the synthetic peptide
B-cell epitope to associate with MHC class II mole-
cules and induce a helper T-cell response. In contrast,
conjugation of (NANP). to a carrier such as tetanus
toxoid allowed a response to the repeat to develop
during immunization (24). Moreover, mice of all
MHC haplotypes develop antibody to (NANP). upon
exposure to parasites. It was inferred therefore that in
responses to intact parasites, one or more helper
T-cell sites on the CS protein provide the capacity
for antibody production via cognate T/B-cell
cooperation.
In accordance with these premises, considerable
effort has been devoted to mapping CS-specific T-cell
sites, by the use of synthetic peptides which induce a
proliferative response from peptide-primed murine
lymphocytes, or act as carriers in Ir gene-controlled
responses to synthetic constructs containing CS re-
petitive B-cells epitopes (15, 16, 18-20, 22, 44, 45). This
is deemed to be essential to the development of anti-
malarial vaccines, as it is hoped that these putative
T-cell sites will enable T-cell activation and Ir
gene-controlled boosting of antibody in response to
parasites.
However, this inferential logic rests on untested
assumptions. This is because in all studies to date, the
mapping of T-cell sites has been solely by homologous
immunization and challenge with synthetic peptide or
recombinant constructs. These studies therefore dem-
onstrate that certain regions are non-self and can
associate with Class II MHC, but do not address the
question of their involvement in responses to intact
parasites. Thus, the studies to date do not undertake
the necessary tests required to show that parasite-
immunized mice acquire CS-specific helper activity,
or that any of these putative helper sites actually
mediate Ir gene-controlled antibody formation in
response to intact parasites. Indeed, Ir gene control of
the antibody response to the native CS protein of
sporozoites is itself a premise that remains to be
validated by experimental test. This is critically im-
portant because cognate interaction among T and B
cells is not the only mechanism by which antibody can
be generated to protein antigens. In contrast to syn-
WHO Bulletin OMS: Supplement Vol. 68 1990
The circumsporozolte protein of Plasmodlum
thetic peptides or monomeric protein antigens, there
exist cases in which antibody formation to complex or
repetitive protein antigens, and some lipoproteins, is
not under Ir gene control (1, 33).
Accordingly, we sought to test the assumption
that cognate T-cell help is required for the production
of antibodies to the repetitive domains of the
P.falciparum and P. berghei CS proteins, by utilizing a
variety of classical protocols originally designed to
elucidate the phenomenon of Ir gene-controlled
MHC-restricted cognate interactions among antigen-
specific T and B cells. First among those we chose was
the protocol of Katz et al., which allows Ig production
among MHC matched, but not mismatched, T and B
lymphocytes in irradiated, semiallogeneic F, re-
cipients (27). Using intact irradiated sporozoites as the
immunizing antigen, the experiments demonstrate
that the anamnestic IgG response to the repetitive
region of native CS proteins does not require cognate
interaction with sensitized T cells, and is not MHC
restricted (43).
In simple adoptive transfer experiments using
primed T and B cells from syngeneic donors, we were
able to show that the secondary response to the native
CS protein of sporozoites was thymus-independent,
i.e., proceeded without requirement for antigen-speci-
fic helper T cells (Table 1). In addition, functional
helper T cells from donors immunized with homolog-
ous sporozoites, conserved CS T-cell epitope peptides,
or prototype CS vaccines, were unable to influence the
magnitude of the thymus-independent response to the
native antigen of parasites (Tables 1 and 2).
However, as euthymic animals produce IgG in
response to sporozoites, it is clear that some T-cell
information is required for the class switch. We have
obtained evidence that this T-cell contribution is non-
cognate (MHC-unrestricted help), by the use of irra-
diated, allogeneic bone-marrow chimeras. These
animals, constitutively unable to mount cognate co-
operation among their MHC mismatched T and B
cell populations, were nonetheless able to respond to
priming and boost with sporozoites with the pro-
duction of anti-CS IgG, as were syngeneic controls
(unpublished data).
One may question the use of the term T-inde-
pendent when an MHC-unrestricted contribution
by T cells may be evident. However, an in vivo T- 1
response may be defined as one in which antigen
drives antibody formation directly, without require-
ment for the MHC-restricted cognate interaction of
antigen-specific T and B lymphocytes in secondary
responses to antigen. That a factor-driven contribu-
tion may be sufficient to influence Ig isotype does not
alter the T-independent status of an antigen.
Finally, we determined that exposure of mice to
sporozoites, which leads to high levels of anti-CS Ig,
69
L. Schofield

Table 1: Antigen-specffic T cells are not required for scondary response to the CS protein of sporozolesa
Titre:
T cells B cells Challenge CS TT
T (naive) B (naive) SPZ 64 ND
TT NO 128
T (naive) B (SPZ+TT) SPZ 4096 32
TT ND 512
T (SPZ+TT) B (SPZ+TT) SPZ 8192 64
TT 32 65536
8A/J mice were primed with 10 pg TT in FCA i.p., or to5 irradiated P. berghei sporozoites i.v. After four weeks, some were sacrificed and
splenic lymphocytes from tetanus toxoid (U) and sporozoite-primed donors were mixed. Adherent cells were removed by plating at 37°C,
and B-cells removed by serial panning on goat anti-mouse Ig (H + L specific) coated plates; 1.5 x 107 T-cells were inoculated i.v. into naive
AIJ recipients. After 24 hours, these animals were irradiated (600 rad) and inoculated i.v. with 1.75 x 107 purified B-cells, prepared by
mixing equal numbers of sporozoite and TT-primed spleen cells followed by two cycles of anti-Thy-1 + C treatment and centrifugation on
Lympholyte M (Accurate Chem. Co.). T-cell lysis (> 98%) was monitored by surface ftuorescence assay. Twenty-four hours after transfer of
B-cells, separate groups of mice (n = 4) were boosted with 10 pg TT in IFA i.p., or t05 irradiated P. berghei sporozoites i.v. After 8 days,
sera were drawn and pooled, titrated serially 2-fold in 1% BSAIPBS, and the IgG (gamma-chain specific) titre to TT or repeat region of
P. berghei CS determined by RIA.

Table 2: Lack of boosting by recCS-specific T cells"

Titreb following challenge with:
T cells B cells SPZ recCS TT
T (naive) B (naive) 64 64 256
T (naive) B (SPZ + TT) 2048 128 2048
T (SPZ + TT) B (SPZ + TT) 4096 256 262144
T (SPZ + TT) B (recCS + UT) 4096 256 262144
T (naive) B (recCS + TT) 4096 256 2048
T (recCS + TT) B (recCS + TT) 2048 16384 262144
T (recCS + TT) B (SPZ + TT) 4096 65536 262144
8 A/J mice were primed with 1Og recCS or TT in FCA i.p., or 105 P. falciparum sporozoites i.v. After four weeks, some were sacrificed and
2 x 107 T-cells from each donor group, purified by panning on goat anti-mouse IgG (H + L specific) coated plates, were inoculated i.v. into
unprimed syngeneic recipients. After 24 hours, recipients were irradiated, and inoculated with 2 x 107 sporozoite, TT, or recCS-primed
B-cells, prepared by treatment with anti-Thy I + complement and centrifugation over Lyrnpholyte M (TT-primed lymphocytes mixed
with recCS or sporozoite-primed lymphocytes prior to lysis). Twenty-four hours later, animals received to5 P. faiciparum sporozoites
i.v., or 10 pg recCS or TT i.p. After 7 days, anti-U and ant-(NANP)3 titres were determined by RIA.
b Titres to (NANP)40 determined for recipients of sporozoites or recCS only; titres to TT determined for TT recipients only.

does not appear to result in the sensitization of anti-
CS T-ceIls, i.e., such mice show little carrier-specific
helper activity in response to the T-dependent
recombinant CS analogue (43).
Thymus-independent antigens are typically poly-
mers with simple, repetitive, antigenic structures. Al-
though often thought of as being comprised only of
carbohydrates, this class does include proteins with
a repetitive B-cell epitope, such as collagen and poly-
meric flagellin. Repetitiveness is understood to pro-
vide an activation signal to antigen-specific B cells by
cross-linking of hapten-specific surface Ig. Both the
degree and duration of cross-linking influence the
B-cell response. By careful manipulation of the vari-
ous side-chains of acrylamide, Dintzis et al. demon-
strated that cross-linking of 12-16 immunoglobulin
70
receptors by regularly spaced haptenic groups was
sufficient to induce T-independent activation of
hapten-specific B cells (13). Interestingly, the repet-
itive domain of the P. fakciparum CS protein contains
15 highly reiterated haptenic groups (NANP)3,
and may therefore behave as proteinaceous bacterial
flagellae (14) by inducing thymus-independent
antibody responses.
In addition to the "cross-linking" models of
thymus-independent B-cell activation (13, 14, 33),
other models (10) propose that some thymus-inde-
pendent antigens are constitutive mitogens. In ac-
cordance with this view, certain lipoproteins (30) are
able to induce thymus-independent B-cell responses;
in these cases, the lipid moiety activates B cells upon
internalization following recognition of the protein
WHO Bulletin OMS: Supplement Vol. 68 1990

antigen by surface Ig. Recent evidence suggests that
many repetitive malarial antigens are acylated
(23, 46). Sequence analysis reveals that the CS protein
shares the amino acid motif, which is thought to
represent the acylation site, exhibited by these acyl-
ated malarial antigens and the Trypanosoma brucei
VSA (46).
We conclude from these findings that there exist
two pathways for the production of IgG to the CS
protein of malaria sporozoites: a T-dependent re-
sponse to synthetic peptide and recombinant protein
analogues of the CS proteins (i.e., current vaccines);
and a thymus-independent response induced by the
parasite itself. This appears to develop in the absence
of functional helper T cell sensitization to the CS
protein; a non-cognate contribution by T cells, how-
ever, may be responsible for the formation of IgG.
These features allow us to propose a plausible select-
ive (Darwinian) rationale for immunodominant re-
petitive antigens in eukaryotic parasites (i.e., as a
mechanism of immune evasion), and suggest that
immunodominant repetitive domains within malarial
proteins are involved in the induction of thymus-
independent responses, either by cross-linking of sur-
face Ig, or by acting as potent B-ell epitopes to
concentrate a PBA moiety on B cells of the appropri-
ate specificity. This is proposed to constitute a mech-
anism of immune evasion, and confer a selective
advantage to the parasite, as follows:
(1) In T-dependent responses, somatic mutation
of VH genes and competition among B-cell clones to
present antigen to T cells leads, through clonal selec-
tion of B cells, to an affinity maturation of Ig. This
process is also responsible for T and B memory cell
formation and regulation of Ig isotype. In contrast,
thymus-independent responses result in few memory
cells, and antibodies that are homogeneous and of
low affinity. Qualitative differences of this type may
explain why anti-CS antibodies generated through the
T-dependent route by a synthetic vaccine may protect
against sporozoite challenge (17), whereas antibodies
raised against sporozoites are invariably ineffectual
(36, 41, 49).
(2) The (NANP)3 sequence is capable of absorb-
ing almost the entire serological reactivity of human
immune sera against falciparum sporozoites: thus the
repetitive surface domain is the major immunodomin-
ant epitope (21) As B cells of different specificities
compete for antigen they are preferentially clonally
expanded (clonal selection). Immunodominant repeti-
tive domains, whilst the target of thymus-independent
antibody responses, may suppress the development of
antibody to adjacent epitopes by a process of B-cell
clonal dominance (epitopic suppression), a strategy of
immunodominance may succeed if the immunodom-
WHO Bulletin OMS: Supplement Vol. 68 1990
The circumsporozolte protein of Pamdum
inant epitope itself does not induce neutralizing anti-
body, which is apparently the case under conditions of
natural exposure to malaria.
(3) Antigen-specific B cells have been shown to
be required for antigen presentation and helper T-cell
priming in vivo (28). B cells reactive with CS protein
repeats, induced into a thymus-independent differen-
tiation pathway, may fail to present antigen to T cells,
thereby suppressing T/B-cell cooperation. This would
account for the failure to detect CS specific helper
T-cell sensitization in association with T-independent
antibody responses, and the failure to boost these
responses with functional helper cells.
Immune evasion via allelic
polymorphism?
No discussion of immune evasion by the CS protein
can be complete without reference to the observation
that a large number of amino acid substitutions exists
in certain regions of the molecule outside the repet-
itive domain (11). It has also been shown that T cells
were specific for one of two variants at position 339
(20). It was therefore suggested that the observed
amino acid substitutions represent polymorphisms
maintained by natural selection, resulting from
variants escaping destruction by immune T cells.
A recent study has shown that the amino acid
substitutions are a widespread feature of natural
populations of the parasite (29). However, it is not
clear that this polymorphism is maintained by a
selective pressure exerted by MHC-restricted immune
T cells (2, 42). This hypothesis requires accepting a
number of stringent conditions, namely:
- these regions must be the target of a protective
herd immunity involving simultaneous recognition
and destruction of parasites;
- the selective advantage offered to variants must be
approximately equivalent to that accruing to drug
resistance mutations, to account for the strong
preferential fixation of non-synonymous substitu-
tions;
the amino acid substitutions must lead to a loss of
recognition by the polyclonal repertoire of T-cell
receptors available in the immune hosts.
It should be stressed that the observed variants
cannot be maintained by MHC polymorphism in the
host population. This is because MHC polymorphism
does not provide a mechanism for frequency-depend-
ent selection. To the contrary, the distribution of host
MHC haplotypes remains constant over the time
scales under consideration here, and thus would pro-
vide a force promoting conformity of CS alleles (max-
imum fitness accruing to those able to associate the
least with any MHC haplotypes).
71
L. Schofield
Despite the uncertainty surrounding the mech-
anism whereby allelic polymorphism is maintained in
the parasite population, it is clear if these regions are
the only areas accessible to cytotoxic T cells, a vaccine
based upon CD8+ T-cell recognition of the CS pro-
tein is unlikely to be effective.
Conclusions
The proposition that the repetitive domain of the CS
protein has evolved to be a good B-cell epitope and to
be the target of a thymus-independent antibody re-
sponse is, in our opinion, the most plausible selective
rationale for this unusual structure yet forwarded.
In addition to providing an explanation for a large
number of disparate observations, it makes the
following predictions:
(1) Closely related taxa may contain divergent
repeat sequences within a conserved CS framework,
as there are a few constraints on the repetitive domain
other than it be repetitive and able to function as a
good B-cell epitope.
(2) The immunogenicity and immunodominance
of the repeats are a predictable consequence of this
strategy of immune evasion.
(3) The antibodies produced by this mech-
anism under natural conditions are likely to be non-
neutralizing.
(4) Low levels of helper T-cell activation will
result from T-1 responses.
The question of whether antibodies against the
CS protein (and other antigens of malaria) can protect
against infection should only be addressed with clear
reference to the source of antibodies and the expe-
rimental conditions. Antibodies produced under nat-
ural circumstance, or by immunization with the
parasite itself, are unlikely to provide protection, but
antibodies generated through other pathways may yet
prove effective. Thus the use of repetitive domains in
anti-malarial vaccines is not necessarily excluded as a
result of these findings, as several effective vaccines
currently use T-1 antigens (e.g., pneumococcal vac-
cines). However, the vaccination strategy to be
pursued should be rationalized accordingly.
References
1. Abbas, A. A reassessment of the mechanisms of
antigen-specific T-cell dependent B-cell activation.
Immunology today, 9: 89-94 (1988).
2. Arnot, D. Malaria and the major histocompatibility
complex. Parasit. today, 5: 138-142 (1989).
3. Ballou, W.R. et al. Safety and efficacy of a recombin-
ant DNA Plasmodium falciparum sporozoite vaccine.
Lancet, 1: 1277-1279 (1987).
72
4. Chestnut, R.W. & Grey, H.M. Studies on the capacity
of B cells to serve as antigen-presenting cells.
J. immunol., 126: 1075 (1981).
5. Chen, D. Aspects of host-parasite interaction in the
rhesus-P. cynomolgi-A. stephensi system. Ph.D.
thesis, New York University School of Medicine, 1974.
6. Clyde, D. et al. Immunization of man against sporo-
zoite-induced falciparum malaria. Am. j. med. sci.,
266: 169-177 (1973).
7. Cochrane, A. et al. Immunization against sporozoites.
In: Kreier, J., ed. Malaria. New York, Academic Press,
1980, pp. 163-202.
8. Collins, W.E. et al. Immunization of Aotus monkeys
with recombinant proteins of an erythrocyte surface
antigen of Plasmodium falciparum. Nature, 323:
259-262 (1986).
9. Collins, W.E. & Contacos, P.G. Immunization of mon-
keys against Plasmodium cynomolgi by X-irradiated
sporozoites. Nature new biol., 236: 176-177 (1972).
10. Coutinho, A. & G. Moller. Thymus-independent B-cell
induction and paralysis. Adv. immunol., 21:114 (1975).
11. de la Cruz, V.F. et al. Sequence variation In putative
functional domains of the circumsporozoite protein of
Plasmodium falciparum. Implications for vaccine de-
velopment. J. biol. chem., 262: 11935-11939 (1987).
12. Del Gludlce, G. et al. The antibody response in mice to
carrier-free synthetic polymers of Plasmodium falci-
parum circumsporozoite repetitive epitope is l-Ab re-
stricted: possible implications for malaria vaccines.
J. immunol., 137: 2952-2955 (1986).
13. Dlntzis, I.M. et al. Molecular determinants of immuno-
genicity: the immunon model of immune response.
Proc. Natl Acad. Sci., USA, 73: 3671-3675 (1976).
14. Feldmann, M. Induction of immunity and tolerance
in vitro by hapten protein conjugates. I. The relation-
ship between the degree of hapten conjugation and
the immunogenicity of dinitrophenylated polymerized
flagellin. J. exp. med., 135: 735-753 (1972).
15. Good, M.F. et al. Genetic control of the immune re-
sponse in mice to a Plasmodium falciparum sporo-
zoite vaccine. Widespread non-responsiveness to
single malaria T epitope in highly repetitive vaccine.
J. exp. med., 164: 655-660 (1986).
16. Good, M.F. et al. Human T-cell recognition of the
circumsporozoite protein of Plasmodium falciparum:
immunodominant T-cell domains map to the poly-
morphic regions of the molecule. Proc. Natl Acad.
Sci., USA, 85: 1199-1203 (1988).
17. Zavala, F. et al. Synthetic peptide vaccine confers
protection against murine malaria. J. exp. med., 166:
1591-1595 (1987).
18. Good, M.F. et al. Construction of synthetic immuno-
gen: use of new T-helper epitope on malaria circum-
sporozoite protein. Science, 235: 1059-1062 (1987).
19. Good, M.F. et al. The real difficulties for malaria
sporozoite vaccine development: non-responsiveness
and antigenic variation. Immunology today, 9: 351-355
(1988).
20. Good, M.F. et al. Parasite polymorphism present with-
in minimal T-cell epitopes of Plasmodium falciparum
circumsporozoite protein. J. immunol., 140: 1645-1650
(1988).
WHO Bulletin OMS: Supplement Vol. 68 1990

21. Zavala, F. et al. Rationale for development of a syn-
thetic vaccine against Plasmodium falciparum mal-
aria. Science, 228: 1436-1440 (1985).
22. Guttinger, M. et al. Human T-cells recognize poly-
morphic and non-polymorphic regions of the Plasmo-
dium falciparum circumsporozoite protein. Europ.
Mol. Biol. Org. j., 7: 2555-2557 (1988).
23. Haldar, K. et al. Acylation of a Plasmodium falci-
parum merozoite surface antigen via sn-1,2-diacyl
glycerol. J. biol. chem., 260: 4969-4974 (1983).
24. Herrlngton, D.A. et al. Safety and immunogenicity in
man of a synthetic peptide malaria vaccine against
Plasmodium falciparum sporozoites. Nature, 328:
257-259 (1987).
25. Hoffman, S. et al. Immune response gene regulation
of immunity to Plasmodium berghei sporozoites and
circumsporozoite protein vaccines: overcoming gene-
tic restriction with whole organisms and subunit vac-
cines. J. immunol., 142: 3581-3586 (1989).
26. Hoffman, S.L. et al. Naturally acquired antibodies to
sporozoites do not prevent malaria: vaccine develop-
ment implications. Science, 237: 639-642 (1987).
27. Katz, D. et al. Cell interactions between histoincom-
patible T and B lymphocytes. II. Failure of physiolo-
gical cooperative interactions between T and B
lymphocytes from allogeneic donor strains in humoral
responses to hapten-protein conjugates. J. exp. med.,
137: 1405-1427 (1973).
28. Kurt-Jones, E. et al. The role of antigen-presenting
B cells in T-cell priming in vivo: studies of B-cell
deficient mice. J. immunol., 140: 3773-3778 (1988).
29. Lockyer, M. et al. Wild isolates of Plasmodium falcipa-
rum show extensive polymorphism in T-cell epitopes
of the circumsporozoite protein. Mol. biochem. para-
sitol., 37: 275-280 (1989).
30. Malchers, F. et al. The lipoprotein of the outer mem-
brane of Escherichia coli: a B-lymphocyte mitogen.
J. exp. med., 142: 473 (1975).
31. Mellouk, S. et al. In vivo and in vitro results sug-
gesting that anti-sporozoite antibodies do not totally
block Plasmodium falciparum sporozoite infectivity.
New Eng. j. med., 315: 648 (1986).
32. Mitchison, N.A. The carrier effect in the secondary
response to hapten-protein conjugates. I. Measure-
ment of the effect with transferred cells and objections
to the local environment hypothesis. Eur. j. immunol.,
1: 10 (1971).
33. Mosler, D. & Subbarao, B. Thymus-independent anti-
gens: complexity of B lymphocyte activation revealed.
Immunology today, 3: 217-222 (1982).
34. Mulligan, R. et al. Active immunization of fowls
against Plasmodium gallinaceum by injections of kil-
led homologous sporozoites. J. Mal. Instit. of India, 4:
25-34 (1941).
35. Nussenzweig, R. et al. Protective immunity produced
WHO Bulletin OMS: Supplement Vol. 68 1990
The circumsporozolte protein of Plasmodlum
by the injection of X-irradiated sporozoites of Plasmo-
dium berghei. Nature, 216: 160-162 (1967).
36. Nussenzweig, R.S. et al. Plasmodium berghei: accel-
erated clearance of sporozoites from blood as part of
immune mechanism in mice. Exp. parasitol., 31: 88-97
(1972).
37. Nuseenzweig, V. & Nuss.nzwelg, R. Rationale for the
development of an engineered sporozoite malaria
vaccine. Advances in immunology, 45: 283-334 (1989).
38. Pike, B. & Nossal, G. A reappraisal of "T-independ-
ent" antigens. I. Effect of lymphokines on the re-
sponse of single adult hapten-specific B lymphocytes.
J. immunol., 132: 1687-1695 (1984).
39. Potocnjak, P. et al. Monovalent fragments (Fab) of
monoclonal antibodies to a sporozoite surface anti-
gen (Pb44) protect mice against malaria infection.
J. exp. med., 151: 1504 (1980).
40. Romero, P. et al. Cloned cytotoxic T cells recognize an
epitope in the circumsporozoite protein and protect
mice against malaria. Nature, 341: 323-326 (1989).
41. Schofield, L. et al. Gamma-interferon, CD8+ T cells,
and antibodies required for immunity to malaria
sporozoites. Nature, 330: 664-666 (1987).
42. Schofield, L. T-cell immunity to malaria sporozoites.
Exp. parasitol., 68: 357-364 (1q89).
43. Schofield, L. & Uadla, P. Lack of Irgene control in the
immune response to malaria. I. A thymus-independ-
ent antibody response to the repetitive surface protein
of sporozoites. J. immunol., 144: 2781-2788 (1990).
44. SInigaglia, F. et al. Epitopes recognized by human T
lymphocytes on malaria circumsporozoite protein.
Eur. j. immunol., 18: 633-636 (1988).
45. SInigaglia, F. et al. A malaria T-cell epitope recog-
nized in association with most mouse and human
MHC Class II molecules. Nature, 336: 778-780 (1988).
46. Smythe, J.A. et al. Identification of two integral mem-
brane proteins of Plasmodium falciparum. Proc. Natl
Acad. Sci., USA, 85: 5195 (1988).
47. Ward, R. & Hayes, D. Attempted immunization of
rhesus monkeys against cynomolgi malaria with irra-
diated sporozoites. Proc. Helminthol. Soc., Wash., 39:
525-529 (1972).
48. Weber, J.L. Molecular biology of malaria parasites.
Exp. parasitol., 66: 143-170 (1988).
49. Weiss, W.R. et al. CD8+ T cells (cytotoxic/suppres-
sors) are required for protection in mice immunized
with malaria sporozoites. Proc. Natl Acad. Sci., USA,
85: 573-575 (1988).
50. Weiss, W. et al. Genetic control of immunity to
Plasmodium yoelii sporozoites. J. immunol., 143:
4263-4266 (1989).
51. Yoshida, N. et al. Hybridoma produces protective
antibodies directed against the sporozoite stage of
malaria parasite. Science, 207: 71-73 (1980).
73