Background
Malaria threatens public health in regions of the world where more than a third of the human population lives [
1,
2]. It has been shown that immunization with radiation-attenuated
Plasmodium sporozoites, the infective stage of the malaria parasite, confers protective immunity [
3,
4]. The role of specific antibody in conferring protection was demonstrated with passive administration of murine mAbs directed against the major repeat epitope of the circumsporozoite (CS) protein [
5] in a rodent model. The corresponding epitope of the human malaria parasite
Plasmodium falciparum is contained within the repeat tetramer peptide (Asn-Pro-Asn-Ala)
n, (NPNA)
n [
6]. In some studies of volunteers protected against malaria by immunization with radiation attenuated
P. falciparum sporozoites, protected individuals had significant elevations of anti-repeat antibodies (>19 μg/ml) [
7].
With the advent of recombinant combinatorial antibody technology [
8,
9] and phage display [
10‐
13] it is possible to attempt to dissect the human antibody response against a wide range of pathogens. In order to further investigate the role of the human antibody response in
P. falciparum sporozoite induced protection, a phage display library of antibody gene fragments isolated from the peripheral blood lymphocytes of such a protected donor (WR5) [
7] was assembled. Recombinant antibodies against the PfCSP structural repeat (NPNA)
3 epitope were selected. Recognition was restricted to a single antibody designated PfNPNA-1, encoded by V
H3 and V
κI families. This restricted humoral response has implications for rational vaccine design and the potential use of this human monoclonal antibody to prevent
P. falciparum infection.
Discussion
The recombinant antibody library construction differed from conventional antibody phage display library assembly [
10‐
13], a pre-selection step was introduced to remove antibody inserts that were either; prematurely terminated, intact but did not translate well or were intact, translated well but failed to translocate into the bacterial periplasmic space, a prerequisite for functional display. Previously an approach towards developing a vector to select for fully intact functional sequences for antibody or peptide display had shown promise with model sequences [
21], but had not been applied for large-scale random antibody library assembly. A "clean-up" vector, plasmid open reading frames expression secretion (pORFES) [
14] was developed and used to remove these non-functional sequences. Up to 50% of the clones from the initial transformed library were non-functional. Some of the non-functional antibody fragments could in part be due to errors introduced during PCR amplification resulting in frame shifts. However it may be that some sequences either did not express well or did not translocate into the periplasmic space. Irrespective of the explanation, the size of the functional library was half of the total transformation events. An initial enhancement of the initial library by removing most non-functional inserts may at first appear to be a minor improvement. However, in conventional phage display the initial expansion of the library prior to panning results in a preferential growth of phage that do not make and display encoded inserts, moreover phage that lack an insert have a greater growth advantage. This results in a phage population that is greatly biased towards non-productive elements, which impacts directly on the panning efficiency. Incorporation of the pORFES step assured that only the functional (1.3 × l0
6) sequences were subsequently transferred to the phage display vector. Panning with a functionally enhanced library resulted in very efficient enrichment and recovery. Previously it had been demonstrated that manipulating the conditions of phage production results in modulation of the density of antibody display on phage [
15]. The phage library was expanded using parameters that would result in either monovalent display (0-1 antibody/phage) or multivalent display (0–5 antibodies/phage) [
15] prior to initiating panning. It was anticipated that a range of antibodies with varying affinities would be present in the library, and modulating antibody display on phage would permit capture antibodies with a range of affinities and sequence diversity.
Induction of protective immunity against sporozoite challenge by exposure to radiation attenuated malaria sporozoite has been demonstrated in humans [
4,
7,
22]. Protection is thought by most investigators to be primarily cellular in nature [
23], but there is no question that antibodies with significant sporozoite neutralizing activity are elicited [
22] and may play a role in protection. The antibody response is primarily directed against the repeat region of the PfCSP. Studies of subunit vaccines which induce antibodies only against the repeat region demonstrate that protective immunity can be induced in some individuals [
24,
25]. At the onset of this study it was proposed that the dissection of the anti-
P.
falciparum sporozoite antibody response by combinatorial antibody library phage display would permit individual selected antibodies to be evaluated for protective potential and the information generated could be used in vaccine design. In particular, attention was focused on antibodies against the structural motif (NPNA)
n. Despite using two different strategies for the elution of repeat region peptide specific antibodies (acid and peptide specific) it would appear that the anti-structural repeat response by this protected individual is restricted to a single V
H/V
L combination observed in the panel of selected phage (n = 25). Sequencing of randomly picked phage prior to panning revealed that a diverse range of V
H and V
L families were represented in the library as shown in Table
2. Moreover the PfNPNA-1 V
H/V
L was not represented in the sampling and was only detected after enrichment.
Comparison of the monovalent PfNPNA-1 molecule with the conventional bivalent murine mAb, such as the in vitro inhibitory 2A10 against
P. falciparum sporozoites indicates that they recognize the repeat epitope(s) with equivalent affinities [
18]. The sequence revealed extensive somatic hyper mutations in both the V
H and V
L genes suggesting antigen driven affinity maturation. Based on these observations, PfNPNA-1 may be a good candidate to develop and evaluate as a protective antibody.
Analysis of field samples in rural Gambia [
26], Thailand [
27] Indonesia [
28] and Kenya [
29], suggest that anti-sporozoite antibody is poorly developed under natural conditions of exposure and does not protect against clinical malaria. In contrast to exposure to
P. falciparum sporozoites under natural conditions in the field, immunization with irradiated
P. falciparum sporozoites induces in general higher levels of antibodies against the PfCSP repeats, and does induce sterile protective immunity [
4,
7,
30‐
38]. In the study by Egan et al., 3 of the 4 volunteers were protected against challenge with
P. falciparum sporozoites. The generally accepted explanation for the lack of protection in the one volunteer is that the volunteer did not receive an adequate immunizing dose of irradiated sporozoites (less than 1000 infective bites [
4,
7]). However, it is of interest that this non-protected volunteer (WR1, [
22])had significantly lower levels of antibodies against the PfCSP repeat than did the protected volunteer who donated cells for this study (WR5, [
22]) (2.4 μg/ml vs 50 μg/ml of specific antibody). This raises the question as to whether the antibodies are markers for adequate immunization or are actually major mediators of protection. Regardless, this anti-repeat response in this protected individual appeared to be restricted to a single antibody. This does not preclude that antibodies directed against non-repeat epitopes on PfCSP and other sporozoite proteins [
39] play a role in protection. It is not possible to conclude that the response against the structural repeat epitope is restricted to a single antibody of moderate affinity, since only a single protected donor has been used in this study. One may speculate that in concordance with the argument put forward by Saul [
40] that the inability to recover high affinity antibody, may reflect that high affinity antibodies may not be required for protection. Due to the repetitive nature of the antigen one can further speculate that only limited affinity maturation is required to obtain physiologically relevant efficacy. The restricted recovery of antibodies is unlikely to be a technical limitation on the phage technology since others have generated panels of very high affinity human antibodies against a range of antigens [
13]. Very few examples of different approaches of generating human antibodies from immune donors are described in the literature, in particular when attempting to make antibodies against the same antigen. Currently it is not possible to fully understand the limitations of a technology. Using an alternative technology of engrafting immune human PBL's directly into SCID mice from donors vaccinated against anthrax vaccine adsorbed, boosting with protective antigen (PA), recovering immortalizing antibody-producing cells via conventional hybridoma technology [
41] resulted in a panel of very high affinity potent neutralizing antibodies against anthrax toxin. Independently, an antibody phage display library from a similar (not identical) immune donor PBL's was constructed and panned against PA [
42] also resulted in a panel of high affinity anti-anthrax PA antibodies. This would suggest that the methodology is not limiting. However in this example, unlike CSP, the PA antigen does not contain repeating epitopes.
Further it is speculated that antibodies directed against the structural (NPNA)n repeat play a role in conferring protection against P. falciparum sporozoites in some of the protected volunteers and this protection may be associated with circulating levels of this specific antibody against the structural repeat.
Efforts are being directed towards producing a fully human IgG based on the PfNPNA-1 V
H and V
L domains for further in vitro and in vivo evaluation. The use of a human monoclonal antibody as a preventive measure against
P. falciparum malaria, would be independent of factors which hinder active vaccination, such as adjuvant effects, the requirement to be effectively presented in a diverse range of human leukocyte class I and II molecules, and immunlogical antagonism [
43,
44]. In practice, the utility of monoclonal antibodies as anti-infectious agents is often negated by the presence and or the inevitable emergence of variants with altered surface epitopes (in particular with viral targets). Fortunately, there has never been a
P. falciparum isolate that does not contain the (NPNA)
n repeats on the PfCSP [
45], and the number of tandem array of repeats on the PfCSP reduces the likelihood of variants arising which evade antibody recognition. This would suggest that an effective antibody directed against the repeats would be effective against all
P. falciparum. If this restricted antibody response to the repeat epitope plays a role in preventing
P. falciparum infection, PfNPNA-1 may be a useful prophylactic agent. Moreover, if PfNPNA-1 is shown to be protective in passive immunization in humans or monkeys as previously demonstrated for anti-
P. vivax CSP murine mAb, NVS3 [
46], it would provide a template that could be used in defining the precise conformation of the structural repeat required for the induction of desired antibodies that can neutralize parasites.
Conclusions
Over the past 25 years the antibody response against the PfCSP repeat epitope has been pursued as a target for active vaccination, with encouraging results [
47]. Our attempt to dissect the protective antibody response against the structural PfCSP repeat revealed that the response was restricted to a single V
H/V
L pairing, designated PfNPNA-1 encoded by V
H3 and V
κ I families (with evidence of somatic mutations). The affinity for the ligand was in the μM range, which in the context of a whole antibody may be more than sufficient for retention on a polyvalent surface such as the
P. falciparum CSP. It is speculated that the induction and the maintenance of high circulating levels of antibodies against the structural PfCSP repeat may be more important than intrinsic high affinity for the ligand for protection against
P. falciparum infection. The absence of high affinity anti-repeat antibodies is in concordance with the expected response against a multivalent antigen (i.e. sporozoite surface). Under physiological conditions a whole IgG antibody and a multimeric ligand result in bivalent binding. Such complexes can have avidities estimated to be approaching the product of two independent monomeric interactions. In this case, the 1 × 10
-6M monovalent affinity of PfNPNA-1 may approach a theoretical higher avidity (1 × 10
-12 M) in the context of a whole antibody. This implies that further affinity maturation either in vivo or in vitro may not necessarily increase physiological effectiveness of the whole IgG antibody. Public health officials have acknowledged the urgency for development of an effective anti-
P.
falciparum malaria vaccine. One of the key criteria of such a putative vaccine may be the induction and maintenance of high levels of anti-(NPNA)n antibodies. The fully human PfNPNA-1 IgG could be used as a positive control in evaluating sera from immunized donors, or possibly be developed as a prophylactic agent that could be used alone or in combination with various vaccination strategies. One immediate hurdle for the development of such an antibody as a prophylactic would be the anticipated high cost of commercial manufacture in mammalian cells. However, advances in alternative antibody production technology may one day provide some more cost effective solutions [
48,
49].
With the availability of an antibody phage display library constructed from a protected individual immunized via bites of irradiated
P. falciparum infected
Anopheles mosquitoes, it should be possible to further dissect the antibody response against "other" sporozoite antigens [
39].
Authors' contributions
JAC was the postdoctoral researcher on this project. WOR and SLH co-investigators. ASK was the PI and recipient of the Department of Army award. All authors read and approved the final manuscript