Evidence implicating MHC genes in the immunological nonresponsiveness to the Plasmodium falciparum CS protein.

Evidence implicating MHC genes in the immunological nonresponsiveness to the Plasmodium falciparum CS protein M.F. Good,' 2 S. Kumar,' A.S. De Groot,' W.R. Weiss,13 I.A. Quakyi,1 F. Dontfraid,1 G.E. Smith,4 M. Cochran,4 J.A. Berzofsky,5 & L.H. Miller1 The circumsporozoite (CS) protein is a major candiate vaccine antigen for the sporozoite stage of malaria. Both cytotoxic Tcells (CTL) and antibody specific for the CS protein are thought to be important in protection. By examining the immune response in mice and humans we have shown that genes mapping to the major histocompatibility complex (MHC) are important for immune responsiveness. Fl mice between high antibody responders and low antibody responders are high antibody responders, suggesting that in this model immune suppressor genes do not control the immune response. Using synthetic peptides to map epitopes for CTL and helper T cells (which are important for the antibody response) we have shown that the T-cell epitopes are located in the polymorphic region of the protein, and we hypothesize that T cells have indeed selected the variation observed in the CS protein. The success of subunit vaccines will depend on the pattern of variation in different geographical locations, the ability to construct multivalent vaccines containing different variant epitopes from this protein, and on the existence of other sporozoite and liver- stage proteins involved in protection. Introduction Humans and laboratory animals can be immunized to malaria sporozoites (1-3). In these experiments, im- munization was by the bite of large numbers of irradiated infectious mosquitos or by intravenous injection of irradiated sporozoites. Natural sporozoite immunity, however, may not readily develop in people from endemic regions (4). Since the circum- sporozoite (CS) protein appears to be a target for protective immunity (see ref. 5 for a recent elegant demonstration), we analysed the cell types responsible for immunity to sporozoites and the immunogenicity of the.CS protein in humans and laboratory animals. Chen and colleagues demonstrated that p-sup- pressed mice (which lacked antibodies or circulating B cells) could be protectively immunized with irradiated sporozoites (6). Since athymic mice could not be immunized, these data demonstrated that T-cells, in the absence of antibody could protect from viable sporozoite challenge. To ask which subset of T-cells 1 The Laboratory of Parasitic Diseases, N.I.A.I.D., The National Institutes of Health, Bethesda, MD, USA. 2 The Queensland Institute of Medical Research, Bramston Terrace, Herston 4006, Brisbane, OLD, Australia. Correspondence should be sent to this address (Dr M.F. Good). 3 Naval Medical Research Institute, Bethesda, MD, USA 4 MicroGeneSys Inc., West Haven, CT, USA. 5 Metabolism Branch, N.C.I., The National Institutes of Health, Bethesda, MD, USA. was responsible for antibody-independent immunity, we depleted Plasmodium yoelii-immune mice of differ- ent T-cell subsets by the parenteral administration of either anti-CD8 or anti-CD4 monoclonal antibodies. Mice depleted of CD8, but not CD4 immune T cells were not protected, suggesting that cytotoxic T lymphocytes (CTL) were responsible for immunity (7). Similar findings were reported using the P. berghei system (8). Methods and discussion Although the CS protein is known to be a target for protective antibodies (9), it was not known whether it was a target for protective CTL. As a start in ad- dressing this question, we transfected the gene for the CS protein of P. falciparum into murine L cells (10). We then showed that spleen cells from mice (B1O. BR (H-2k)) immunized with either irradiated P.falciparum sporozoites or CS-recombinant vaccinia (kindly pro- vided by Dr B. Moss) contained CS-specific CTL. Using overlapping synthetic peptides, we then map- ped the epitope recognized by the CTL to residues 368-390 of the CS protein. We are currently studying the human CTL response to this protein, as well as the murine CTL response to the P. yoelii CS protein. However, since human and murine T-cells often rec- ognize similar regions of proteins (11), it is likely that the human response to the protein will include this carboxyterminal region recognized by mice. Unfortu- nately, this region is known to be variant. BIO.BR mice appeared to recognize only this single segment of so Bulletin of the World Health Organization, U (Suppl.): 80-84 (1990) immunogenicity of the circumsporozolte (CS) protein the protein. Another strain (BIO.D2 (H-2d)) was examined, but these mice failed to recognize the pro- tein (12). Thus, from two different strains examined, expressing between them 5 different MHC Class I genes, only a single CTL epitope was recognized. A similar paucity of CTL epitopes has been reported in other systems (13). We have hypothesized that the variation ob- served in the CS protein may be selected by CTL (14). Since T cells only recognize antigenic epitopes in association with MHC molecules, such a selection would have to occur during a cellular stage in the life- cycle. We believe that CTL recognize and destroy hepatocytes infected with sporozoites expressing a CS protein epitope to which the CTL have previously been exposed. Hepatocytes containing sporozoites bearing variant CS proteins would escape destruction and thus be selected. Such variation, selected by CTL, would also affect the antibody response to the protein, as discussed below. In murine systems, monoclonal antibodies direc- ted against the CS protein repetitive epitope can passively protect an animal from live sporozoite chal- lenge (9). This suggested that human vaccine develop- ment programmes should consider creating an immunogen capable of inducing a high-titre, high- affinity antibody response (15-17). Such an immuno- *gen would consist of a B-cell epitope (the repetitive NANP epitope) conjugated to one or more T-cell epitopes as a source of T-cell help. T-cell epitopes could be of either parasite or non-parasite origin. However, for continual antibody-mediated protec- tion, titres must remain high at all times. This is because there is no time following sporozoite inocu- lation and prior to penetration of hepatocytes for the antibody response to be boosted and to protect from the immediate sporozoite challenge. For a vaccine to be effective, titres would have to be high following vaccination, and would require natural boosting by the parasite itself. Antibodies would thus block sporo- zoites and the sporozoites would in turn boost anti- body levels in readiness for the next sporozoite challenge. B cells derive help from T cells following uptake and processing of the antigen by the B cell. If the antigen (CS protein) contains a T-cell epitope against which the person has previously been immun- ized, there will be a secondary or boosted antibody response. If the immunogen contained T-cell epitopes from a protein other than the CS protein, such boost- ing would obviously not occur following sporozoite challenge. We initially examined the immunogenicity of the CS protein, in terms of the antibody response by different individuals, and then determined the helper T-cell epitopes on the protein. At first we determined whether the antibody response to the protein was controlled by genes map- ping within the MHC region. Different strains of H-2 congenic mice were immunized with either a P. falci- parum CS-recombinant vaccinia virus (18) or with purified CS protein prepared from CS-recombinant baculovirus-infected cells (11). From nine different H-2 congenic strains examined, only two were high responders to the protein. A number of strains were either low responders or non-responders. It was pos- sible to map the MHC genes responsible for a high response to I-Ab and I-A'. To demonstrate that re- sponsiveness was controlled by a 'responder' gene, rather than negatively controlled by a 'suppressor gene', the immunogenicity of the CS protein was determined for 'high responder x low responder' (F1) animals. Thus, BlO.BR (H-2"), BlO.D2 (H-2d), and BlO.BR x BlO.D2 (Fl) were immunized and boosted with purified CS protein produced by CS-recombin- ant baculovirus-infected cells (Table 1). We observed that the BlO.BR and the Ft animals were high responders, but that the BlO.D2 animals were low responders. Thus, at least in these strain combina- tions, immune suppressor genes do not control the response to the CS protein. Table 1: Immune response genes within MHC control the humoral response to purifld recombinant CS protein" Anti-(NANP)n titre Strain Primary Secondary B10.BR (H-2k) 1 256 4 096 2 512 >16384 3 512 >16384 4 64 256 5 512 >16384 B1O.D2 (H-2d) 1 <16 <16 2 <16 <16 3 <16 <16 B1O.BR x B1O.D2 (Fl) 1 1024 >16384 2 64 4 096 3 64 4 096 4 1024 4096 5 256 >16384 8Mice were immunized with 10 pg purified circumsporozoite pro- tein (P. falciparum, 7G8 strain) (produced by recombinant bac- ulovirus-infected cells, as described (11)) emulsified in CFA. Three weeks later, the animals were boosted with 10 pg aqueous CS protein. 'Primary' sera were taken just prior to the boost and 'secondary' sera were taken 10 days after the boost. Antibodies binding R32tet32 (15) were determined by ELISA as described (18). Titre is defined as the highest dilution of serum which produced an absorbance reading of at least 3 SD above the mean absorbance of a panel of normal mouse sera at the same dilution. WHO Bulletin OMS: Supplement Vol. 68 1990 S1 M.F. Good et al. We (19) and others (20) had already shown that I-Ab-bearing animals recognized the repetitive epi- tope (NANP) as a T-cell epitope. However, no other strains responded to this epitope. Thus, there must have been another T-cell epitope present on the CS protein recognized by BIO.BR mice. Using an al- gorithm to predict T-cell epitopes on the basis of helical amphipathicity, we located a major T-cell epitope in the carboxyterminal region of the protein (18). This epitope stimulated proliferating and helper T cells in BlO.BR mice. It is also an immunodominant CD4 T-cell epitope for humans (21). To locate other possible CD4 T-cell epitopes, we used the overlapping set of peptides referred to above, which span the entire sequence. Both human and murine studies were per- formed (11, 21). Both studies gave similar findings and demonstrated that for both species, most of the CD4 epitopes were located in the carboxyterminal, variant region of the protein. The human study, performed with lymphocytes from adults in the Gambia, also found that many adults (about 40%) were unable to respond to any of the overlapping set of peptides. This finding was analogous to the murine studies in two ways: firstly, many different H-2 congenic mouse strains were unable to respond to the CS protein, or were 'low responders' suggesting that there are few CD4 helper T-cell epitopes on the protein; secondly, similar regions of the protein were recognized by 'responders' from both species. This evidence supports the concept that physico-chemical properties of pep- tides are important in determining immunogenicity for T cells (22, 23). According to the hypotheses, the structure of the processed peptide is important in determining whether it initially binds to the MHC molecule as a first step in stimulating T cells. Since MHC molecules from different species are similar in basic structure and share sequence homologies, a peptide that is immunogenic in one species may be expected to be immunogenic in another species. This would not imply that a minimal T-cell site in a mouse would be a minimal T-cell site in a human, merely that T-cell epitopes would come from similar or over- lapping regions of the protein. The demonstration by this unbiased test that murine and human CD4 T-cell epitopes map to the same regions of the CS protein make the mouse a suitable candidate for mapping T-cell epitopes of vaccine interest for human patho- gens. Although we have not performed family studies with humans from endemic areas, we believe that MHC genes also control the human response: firstly, humans and mice both recognize similar regions of the protein (11), suggesting that in humans, as in mice, MHC genes play an important role; secondly, we have preliminary evidence of association between certain DR types and response to the immunodominant pep- tides (24). However, by examining the proliferative response of non-immune individuals to peptides from the CS protein, Sinigaglia and colleagues have found that many different MHC DR antigens can present a slightly modified peptide from the constant region of the protein to a CS-specific T-cell clone (25). The discrepancy between these results and the restricted immune response observed in the Gambia need to be resolved. Thus, we know that T cells play an important role in sporozoite immunity, both as effector cells (CTL) and as helpers for an antibody response; T-cell epitopes that stimulate CTL and helper T cells have been defined for the CS protein; but the epitopes occur in variant regions of the protein-an observa- tion made more significant by the demonstration that the variant T-cell epitopes usually do not cross-react (26). How do we now use this information to help us design a vaccine? Firstly, we must realize that many other sporozoite-stage or liver-stage proteins may play important roles in a 'sporozoite' vaccine, and the immunogenicity and relevance of these proteins has scarcely been investigated. Secondly, the pattern of variation of the CS protein must be further examined. At present, the genes for six different CS proteins have been sequenced. The different strains represent labor- atory isolates from different parts of the world. Whether such strains are at present occurring in endemic regions is not known. Neither is the degree of variation within given geographical regions known. The knowledge of such a pattern of variation will have major ramifications for vaccine design and may de- cide the feasibility of a CS-based vaccine. If variation within a given geographical region is limited, then the CS protein or proteins representative of that region may form a useful vaccine. From the perspective of an antibody-based vaccine, constant re-exposure to only a few different strains should boost the antibody response to the protein following vaccination. If, how- ever, there was extensive variation within a locale, an individual may only rarely be exposed to vaccine strain sequences. Thus, even if he responded to the vaccine, his immune response would rarely be boos- ted. Although the pattern of variation is not known, we have recently observed that T lymphocytes from most individuals in an endemic region are capable of responding to at least one of six different peptides representing the immunodominant domains of three different CS proteins (24). This suggests that an anti- body-oriented vaccine is definitely worth pursuing. The inclusion of a panel of immunodominant do- mains from different known CS sequences may prime T cells for exposure to various strains. However, there is an important difference between natural exposure followed by in vitro challenge (as in our experiment) and immunization followed by natural exposure. In the former case, lymphocytes exposed to parasites WHO Bulletin OMS: Supplement Vol. 68 199082 Immunogenicity of the circumsporozolte (CS) protein over a lifetime may be recalled in vitro. Nothing is known about the history of strain exposure in any endemic region. In the latter case, the effectiveness of the vaccine will depend on the ability of lymphocytes primed by a (finite) panel ofknown variant peptides to respond to parasites currently present in the endemic region. Knowledge of the pattern of variation would also have major ramifications on the design of a vaccine to stimulate cellular immunity. Any degree of variation of CS protein sequence within a given locale would almost preclude a role for CS-specific cytotoxic T lymphocyte immunity in vaccine design, unless the CTL epitopes on the different proteins (if indeed those particular CS proteins contained CTL epitopes) cross-reacted. Even then, such a vaccine would have a protective effect only on those individuals within the population whose CTL recognized those particular CTL epitopes. It would seem that the role of CTL in vaccine design could be augmented by the identifica- tion of CTL epitopes on many other sporozoite- or hepatic-stage proteins. Then it may be possible to use a large panel of different CTL epitopes that could stimulate a protective response in most people. How- ever, it is sobering to realize that in endemic popula- tions natural immunity to sporozoites may only rarely occur (4). This may indicate that a large number of .protective CTL epitopes do not exist on the sporo- zoite- or the hepatic-stage forms. Other factors, how- ever, may be responsible for the failure of natural sporozoite immunity to develop, including antigen dosage, malaria-induced immunosuppression, emer- gence of new antigenic strains in a given region, etc. Thus it is very important to continue to study the immune response to sporozoites, the geographical and dynamic pattern of variation, and to search for other important antigens which may play a role in protection. Acknowledgement We are very grateful to Smith Kline & French (Swedeland, PA) for providing R32tet32 for use in the ELISA assays. References 1. Clyde, D.F. et al. Immunization of man against sporo- zoite-induced falciparum malaria. Am. j. med. sci., 266: 169-177 (1973). 2. Rleckmann, K.H. et al. Use of attenuated sporozoites in the immunization of human volunteers against falci- parum malaria. Bull. Wld Hith Org., 57 (Suppl. 1): 261-265 (1979). 3. Nussenzwelg, R.S. & Nu.senzwelg, V. Development of sporozoite vaccines. Phil. trans. Roy. Soc., London, Ser. B, 307: 117-128 (1984). 4. Hoffman, S.L. et al. Naturally acquired antibodies to sporozoites do not prevent malaria: vaccine develop- ment implications. Science, 237: 639-642 (1987). 5. Sadoff, J.C. et al. Oral Salmonella typhimurium vac- cine expressing circumsporozoite protein protects against malaria. Science, 240: 336-338 (1988). 6. Chen, D.H. et al. Immunity to sporozoite-induced mal- aria infection in mice. I. The effect of immunization of T and B-cell-deficient mice. J. immunol., 118:1322-1327 (1977). 7. Wei", W.R. et al. CD8+ T cells (cytotoxic/suppres- sors) are required for protection in mice immunized with malaria sporozoites. Proc. Natl Acad. Sc., USA, 85: 573-576 (1988). 8. Schofield, L. et al. Gamma interferon, CD8 T cells and antibodies required for immunity to malaria sporo- zoites. Nature, 330: 664-666 (1987). 9. 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: 1504-1513 (1980). 10. Kumar, S. et al. Cytotoxic T cells specific for the circumsporozoite protein of Plasmodium falciparum. Nature, 334: 258-260 (1988). 11. Dontfrald, F. et al. Human and murine CD4 T-cell epitopes map to the same region of the malaria cir- cumsporozoite protein: limited immunogenicity of sporozoites and circumsporozoite protein. Mol. biol. med., 5: 185-196 (1988). 12. Good, M.F. et al. Limited immunological recognition of critical malaria vaccine candidate antigens. Science, 242: 574-577 (1988). 13. Bennick, J.R. & Yewdell, J.W. Murine cytotoxic T lymphocyte recognition of individual influenza virus proteins: high frequency of nonresponder MHC Class I alleles. J. exp. med., 168: 1935-1939 (1988). 14. Good, M.F. et al. The real difficulties for malaria sporozoite vaccine development: nonresponsiveness and antigenic variation. Immunology today, 9: 351-355 (1988). 15. Young, J.F. et al. Expression of Plasmodium falcipa- rum circumsporozoite proteins in Escherichia coli for potential use in a human malaria vaccine. Science, 228: 996-999 (1985). 16. Ballou, W.R. et al. Safety and efficacy of a recombin- ant DNA Plasmodium falciparum sporozoite vaccine. Lancet, 1: 1277-1281 (1987). 17. 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). 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. Genetic control of the immune re- sponse in mice to a Plasmodium falciparum sporo- zoite vaccine. Widespread nonresponsiveness to single malaria T epitope in highly repetitive vaccine. J. exp. med., 164: 655-660 (1986). 20. Del Gludlce, G. et al. The antibody response in mice to WHO Bulletin OMS: Supplement Vol. 68 1990 83 M.F. Good et al. carrier-free synthetic polymers of Plasmodium falci- parum circumsporozoite repetitive epitope is l-Ab_ restricted: possible implications for malaria vaccines. J. immunol., 137: 2952-2955 (1986). 21. 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). 22. Berzeosky, J.A. et al. Protein antigenic structures rec- ognized by T cells: potential applications to vaccine design. Immunol. rev., 98: 9-52 (1987). 23. Rothbard, J.B. & Taylor, W.R. A sequence pattern common to T-cell epitopes. Europ. Mol. Biol. Org. j., 7: 93-100 (1988). 24. Da Groot, A.S. et al. Human T-cell recognition of polymorphic epitopes from malaria circumsporozoite protein. J. immunol., 142: 4000-405 (1989). 25. SInigagila, F. et al. A malaria T-cell epitope recog- nized in association with most mouse and human MHC Class 11 molecules. Nature, 336: 778-780 (1988). 26. de Ia Cruz, V.F. et al. Lack of cross-reactivity between variant T-cell determinants from malaria circum- sporozoite protein. J. immunol., 141: 2456-2460 (1988). 84 WHO Bulletin OMS: Supplement Vol. 68 1990