Association of malaria parasite population structure, HLA, and immunological antagonism

Science

Volumn 279, Issue 5354, 1998, Pages 1173-1177

  Gilbert, Sarah C ^a   Plebanski, Magdalena ^b   Gupta, Sunetra ^c   Morris, Joanne ^d   Cox, Martin ^a   Aidoo, Michael ^b   Kwiatkowski, Dominic ^e   Greenwood, Brian M ^e   Whittle, Hilton C ^e   Hill, Adrian V S ^a,b  

^a UNIVERSITY OF OXFORD   (United Kingdom)

^b UNIVERSITY OF OXFORD   (United Kingdom)

^c UNIVERSITY OF OXFORD   (United Kingdom)

^d LONDON SCHOOL OF HYGIENE AND TROPICAL MEDICINE   (United Kingdom)

^e MEDICAL RESEARCH COUNCIL LABORATORIES   (Gambia)

Author keywords

[No Author keywords available]

Indexed keywords

EPITOPE; HLA B ANTIGEN;

ARTICLE; CELLULAR IMMUNITY; CYTOTOXIC T LYMPHOCYTE; GAMBIA; HLA TYPING; HOST PARASITE INTERACTION; HOST RESISTANCE; HUMAN; MAJOR CLINICAL STUDY; MALARIA; PLASMODIUM FALCIPARUM; POPULATION STRUCTURE; PRIORITY JOURNAL;

ALLELES; ANIMALS; ANTIGENS, PROTOZOAN; CHILD; EPITOPES; EVOLUTION; EVOLUTION, MOLECULAR; GAMBIA; GENES, PROTOZOAN; HLA-B35 ANTIGEN; HUMANS; LIGANDS; MALARIA, FALCIPARUM; MODELS, BIOLOGICAL; PLASMODIUM FALCIPARUM; PROTOZOAN PROTEINS; T-LYMPHOCYTES, CYTOTOXIC; VARIATION (GENETICS);

PLASMODIUM FALCIPARUM;

EID: 7144227964     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.279.5354.1173     Document Type: Article

References (33)

1. A. V. S. Hill et al., Nature 352, 595 (1991).
2. M. R. Thursz et al., N. Engl. J. Med. 332, 1065 (1995).
3. P. Romero et al., Nature 341, 323 (1989).
4. A. V. S. Hill et al., ibid. 360, 434 (1992).
5. A. Bertoletti et al., ibid. 369, 407 (1994).
6. P. Klenerman et al., ibid., p. 403.
7. M. Aidoo et al., Lancet 345, 1003 (1995).
8. Peripheral blood mononuclear cells (PBMCs) from malaria-exposed individuals determined to be responders to cp26 or cp29 12 months before (7) were stimulated in vitro with peptides cp26 (KPKDELDY) (20) and cp29 (KSKDELDY). Binding of the peptides cp26 and cp29 was assessed in HLA assembly assays using the T2 cell line transfected with HLA-B35, as described [J. Elvin, V. Cerundolo, T. Elliott, A. Townsend, Eur. J. Immunol. 21, 2025 (1991)]. The cp27 and cp28 peptides failed to show binding at 100 μM; cp26 and cp29 showed 50% maximal assembly at 30 μM and 2 μM, respectively. The induction of primary CTL responses with these peptides has also been explored in detail (13). In brief, the generation of primary CTL lines from malaria-naïve donors by peptide cp26 was inhibited 91% in the presence of an equimolar dose of cp29 (n = 4), and the generation of lines by peptide cp26 was inhibited 64% by an equimolar dose of cp29 (n = 4). The cp29 peptide induced short-term CTL lines from malaria-naïve donors threefold more efficiently than did cp26 at the same peptide concentration. Methods are described in (27).
9. D. J. Conway, B. M. Greenwood, J. S. McBride, Parasitology 1, 1 (1991).
10. R. E. L. Paul et al., Science 269, 1709 (1995); H. A. Babiker et al., Parasitology 109, 413 (1994); D. Walliker et al., Science 236, 1661 (1987).
11. R. E. L. Paul et al., Science 269, 1709 (1995); H. A. Babiker et al., Parasitology 109, 413 (1994); D. Walliker et al., Science 236, 1661 (1987).
12. R. E. L. Paul et al., Science 269, 1709 (1995); H. A. Babiker et al., Parasitology 109, 413 (1994); D. Walliker et al., Science 236, 1661 (1987).
13. M. J. Lockyer, K. Marsh, C. I. Newbold, Mol. Biochem. Parasitol. 37, 275 (1989).
14. A shown above. The latter assumes that mixing between vectors and hosts occurs in a homogeneous fashion. Spatial heterogeneities may create differences in the force of infection terms between A and B such that the proportionate contribution from the same genotype is larger; in this case, we may expect to see a stronger effect of antagonism on the parasite distribution within A (that is, the host genotype in whom antagonism occurs) than within B. In the limit of no mixing, the force of infection terms will be entirely separate, and the effects of antagonism will only be seen in the hosts with HLA-B35.
15. M. Plebanski et al., in preparation.
16. In reality, they will also be competing by virtue of other common immune responses (such as blood-stage immunity); within this theoretical framework, this competition is mediated by the non-strain specific component of immunity within the non-HLA-B35 population. When two strains are in competition, even slight differences in reproductive success can result in the elimination of one or the other and can precipitate large differences in frequency.
17. Strong APL antagonism is observed for polyclonal CTL populations at both induction and effector levels (Fig. 1) (13).
18. S. Gupta, K. Trenholme, R. M. Anderson, K. P. Day, Science 263, 961 (1994).
19. M. P. Davenport, Immunol. Today 16, 432 (1995).
20. P. Klenerman, U. C. Meier, R. E. Phillips, A. J. McMichael, Eur. J. Immunol. 25, 1927 (1995); K. Snoke et al., J. Immunol. 151, 6815 (1993).
21. P. Klenerman, U. C. Meier, R. E. Phillips, A. J. McMichael, Eur. J. Immunol. 25, 1927 (1995); K. Snoke et al., J. Immunol. 151, 6815 (1993).
22. A. Lalvani et al., Res. Immunol. 145, 461 (1994).
23. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.
24. 51Cr release assays were performed as described (4, 7), using HLA-matched B lymphoblastoid cell lines as targets. Targets were prepulsed with 10 mM peptide. Based on previous studies in this population on malaria-specific CTL (1, 7), a threshold level for a positive response is taken as 10% specific lysis, a level further validated by limiting dilution analysis [M. Plebanski, M. Aidoo, H. C. Whittle, A. V. S. Hill, J. Immunol. 158, 2849 (1997)]. Antagonism was investigated in two ways. For donor Z22 (cp26 responder), labeled targets were prepulsed for 1 hour with cp26, washed, and incubated (1 hour) alone or with varying concentrations of cp29 (putative antagonist) or FluB35 {an unrelated HLA-B35-binding peptide ASCMGLIY [T. Dong et al., Eur. J. Immunol. 26, 335 (1996)] derived from influenza matrix protein}. Inhibition was calculated as 100 × percent of specific lysis (% SL) [cp26 alone - cp26 + (cp29 or FluB35)]/cp26 alone. Donors Z22 and Z25 (both cp26 responders) were tested in a subsequent separate experiment in an antagonism assay where the antagonist is present on a different target cell from the index peptide. The assay was set up with d14 effectors from the cp26-stimulated PBMC cultures. Briefly, radiolabeled, or "hot," targets were prepulsed with or without cp26 and unlabeled, or "cold," targets were prepulsed with either cp29 or FluB35. Targets were then used together at a cold:hot ratio of 1:1. Assays were performed in duplicate, and differences between wells were always < 10% of the averaged counts. Data are presented as % SL by subtracting the corresponding background lysis controls, for example, percent of lysis (% lysis) [cp26 pulsed targets (hot) with FluB35 pulsed (cold)] - % lysis [no peptide targets (hot) with FluB35 (cold)]. A similar assay was set up for the cp29 responder Z87, but in this case the hot targets were pre-pulsed with either cp29 or in the absence of peptide and cold targets with either cp26 or FluB35. A fourth HLA-B35 donor studied showed an early cross-reactive response without antagonism (22). Subtracted background was in all cases % lysis [no peptide targets (hot) and cp26-pulsed targets (cold)]. For lines derived from the malaria-naïve donors (BR and PTE), antagonism was investigated by using as targets autologous cells prepulsed for 1 hour with 10 μM cp26 or cp29, washed, and then incubated for a further 1 hour in the absence or presence of 10 μM of the corresponding antagonist. These lines were not tested on recombinant vaccinia-infected targets, but other malaria-specific CTL lines from this population have been shown to lyse target cells infected with recombinant vaccinia virus [M. Aidoo, A. Lalvani, H. C. Whittle, A. V. S. Hill, K. J. H. Robson, Int. Immunol. 9, 731 (1997)].
25. 51Cr release assays were performed as described (4, 7), using HLA-matched B lymphoblastoid cell lines as targets. Targets were prepulsed with 10 mM peptide. Based on previous studies in this population on malaria-specific CTL (1, 7), a threshold level for a positive response is taken as 10% specific lysis, a level further validated by limiting dilution analysis [M. Plebanski, M. Aidoo, H. C. Whittle, A. V. S. Hill, J. Immunol. 158, 2849 (1997)]. Antagonism was investigated in two ways. For donor Z22 (cp26 responder), labeled targets were prepulsed for 1 hour with cp26, washed, and incubated (1 hour) alone or with varying concentrations of cp29 (putative antagonist) or FluB35 {an unrelated HLA-B35-binding peptide ASCMGLIY [T. Dong et al., Eur. J. Immunol. 26, 335 (1996)] derived from influenza matrix protein}. Inhibition was calculated as 100 × percent of specific lysis (% SL) [cp26 alone - cp26 + (cp29 or FluB35)]/cp26 alone. Donors Z22 and Z25 (both cp26 responders) were tested in a subsequent separate experiment in an antagonism assay where the antagonist is present on a different target cell from the index peptide. The assay was set up with d14 effectors from the cp26-stimulated PBMC cultures. Briefly, radiolabeled, or "hot," targets were prepulsed with or without cp26 and unlabeled, or "cold," targets were prepulsed with either cp29 or FluB35. Targets were then used together at a cold:hot ratio of 1:1. Assays were performed in duplicate, and differences between wells were always < 10% of the averaged counts. Data are presented as % SL by subtracting the corresponding background lysis controls, for example, percent of lysis (% lysis) [cp26 pulsed targets (hot) with FluB35 pulsed (cold)] - % lysis [no peptide targets (hot) with FluB35 (cold)]. A similar assay was set up for the cp29 responder Z87, but in this case the hot targets were pre-pulsed with either cp29 or in the absence of peptide and cold targets with either cp26 or FluB35. A fourth HLA-B35 donor studied showed an early cross-reactive response without antagonism (22). Subtracted background was in all cases % lysis [no peptide targets (hot) and cp26-pulsed targets (cold)]. For lines derived from the malaria-naïve donors (BR and PTE), antagonism was investigated by using as targets autologous cells prepulsed for 1 hour with 10 μM cp26 or cp29, washed, and then incubated for a further 1 hour in the absence or presence of 10 μM of the corresponding antagonist. These lines were not tested on recombinant vaccinia-infected targets, but other malaria-specific CTL lines from this population have been shown to lyse target cells infected with recombinant vaccinia virus [M. Aidoo, A. Lalvani, H. C. Whittle, A. V. S. Hill, K. J. H. Robson, Int. Immunol. 9, 731 (1997)].
26. 51Cr release assays were performed as described (4, 7), using HLA-matched B lymphoblastoid cell lines as targets. Targets were prepulsed with 10 mM peptide. Based on previous studies in this population on malaria-specific CTL (1, 7), a threshold level for a positive response is taken as 10% specific lysis, a level further validated by limiting dilution analysis [M. Plebanski, M. Aidoo, H. C. Whittle, A. V. S. Hill, J. Immunol. 158, 2849 (1997)]. Antagonism was investigated in two ways. For donor Z22 (cp26 responder), labeled targets were prepulsed for 1 hour with cp26, washed, and incubated (1 hour) alone or with varying concentrations of cp29 (putative antagonist) or FluB35 {an unrelated HLA-B35-binding peptide ASCMGLIY [T. Dong et al., Eur. J. Immunol. 26, 335 (1996)] derived from influenza matrix protein}. Inhibition was calculated as 100 × percent of specific lysis (% SL) [cp26 alone - cp26 + (cp29 or FluB35)]/cp26 alone. Donors Z22 and Z25 (both cp26 responders) were tested in a subsequent separate experiment in an antagonism assay where the antagonist is present on a different target cell from the index peptide. The assay was set up with d14 effectors from the cp26-stimulated PBMC cultures. Briefly, radiolabeled, or "hot," targets were prepulsed with or without cp26 and unlabeled, or "cold," targets were prepulsed with either cp29 or FluB35. Targets were then used together at a cold:hot ratio of 1:1. Assays were performed in duplicate, and differences between wells were always < 10% of the averaged counts. Data are presented as % SL by subtracting the corresponding background lysis controls, for example, percent of lysis (% lysis) [cp26 pulsed targets (hot) with FluB35 pulsed (cold)] - % lysis [no peptide targets (hot) with FluB35 (cold)]. A similar assay was set up for the cp29 responder Z87, but in this case the hot targets were pre-pulsed with either cp29 or in the absence of peptide and cold targets with either cp26 or FluB35. A fourth HLA-B35 donor studied showed an early cross-reactive response without antagonism (22). Subtracted background was in all cases % lysis [no peptide targets (hot) and cp26-pulsed targets (cold)]. For lines derived from the malaria-naïve donors (BR and PTE), antagonism was investigated by using as targets autologous cells prepulsed for 1 hour with 10 μM cp26 or cp29, washed, and then incubated for a further 1 hour in the absence or presence of 10 μM of the corresponding antagonist. These lines were not tested on recombinant vaccinia-infected targets, but other malaria-specific CTL lines from this population have been shown to lyse target cells infected with recombinant vaccinia virus [M. Aidoo, A. Lalvani, H. C. Whittle, A. V. S. Hill, K. J. H. Robson, Int. Immunol. 9, 731 (1997)].
27. M. Plebanski, data not shown.
28. Expected values were calculated for three different parasite rates assuming a binomial distribution with the proportion of the total population uninfected being 1-PR. Pairing of cp27 with cp26 or cp29 was found significantly less often than expected. Linkage disequilibrium with variants of another polymorphic region of the CS protein, Th2R (11) was significant for only 3 of the 12 combinations assessed: Cp38-cp28 (D = 0.041), cp39-cp26 (D = 0.09), and cp39-cp29 (D = 0.064). Correcting for allele frequency variation by calculating the ratio of D to its maximum possible value (D′) [M. Nei, Molecular Evolutionary Genetics (Columbia Univ. Press, New York, 1987)] yielded a mean D′ of 0.28 for the 12 combinations, representing relatively weak linkage disequilibrium for loci only 26 amino acids apart.
29. 2, and 0.5 mM spermidine, with deoxynucleotide triphosphates (0.2 mM) and a 0.5 mM concentration of each primer, for an initial denaturation step of 99°C for 5 min, followed by 35 cycles each of 1 min 30 s at 94°C, 2 min at 55°C, and 2 min at 72°C. The results from all 20 samples tested by allele-specific PCR agreed with the dot-blot results. PCR products were also cloned into pGEM-T (Promega) and sequenced to directly confirm the dot-blot results. This showed that all cp29 sequences had the codon GGT immediately before the B35 epitope [N. Yoshida et al., Exp. Parasitol. 71, 386 (1990)], whereas cp26 sequences had AAT. A combination of allelle-specific PCR sequencing was used to confirm the presence of both cp26 and cp29 sequences in 10 randomly selected samples that were positive for both cp26 and cp29 on dot-blots. There was complete concordance between the dot-blot, allele-specific PCR, and sequencing results. Linkage disequilibrium with the polymorphic TH2R region (amino acids 334 through 342) close to the Cp26-cp29 variants (amino acids 368 through 390) was assessed by typing 478 samples for three TH2R variants [cp36 (YLKTIQNSL), cp38 (YLQKIKNSL), and cp39 (YLNKIQNSL)], using sequence-specific oligonucleotides.
30. 2 analysis. This was the primary comparison because of the functional difference between these strains: cp26 and cp29 are HLA-B35 epitopes, cp27 and cp28 are not. Confounders were allowed for by log-linear regression analysis using SPSS 6.0. The significant increase in the number of cp26 or cp29 infections, or both, in the HLA-B35 control group was found to be independent of age, malaria condition, ethnic group, hospital site, and area of residence. Further analysis of the influence of two class I alleles (A2 and B53) and 10 class II haplotypes on strain distribution showed no significant association.
31. M. Plebanski, C. E. M. Allsopp, M. Aidoo, H. Reyburn, A. V. S. Hill, Eur. J. Immunol. 25, 1783 (1995).
32. 51Cr release assays after 7 days of culture. Limiting dilution analysis performed on donors Z22 and Z58 confirmed the existence of low levels of CTL to cp26 in these donors in the absence of cp29 CTL precursors. However, consistent with data on antagonistic HIV variants (6), a low level of killing of cp29 pulsed target could be detected for cp26-stimulated CTLs if the assay was harvested after 18 hours instead of 4 hours (22).
33. We thank the Gambian children and their parents and guardians, and the villagers of Brefet for their participation. This study was approved by the Gambian government-U.K. Medical Research Council (MRC) joint ethical committee. We thank C. Allsopp, A. Gallimore, and A. Jepson for assistance, and A. McMichael, A. McLean, R. Andersen, and C. Newbold for advice. A.V.S.H. is a Wellcome Trust Principal Research Fellow. Funded by the Wellcome Trust and the MRC.