Comparative analysis of ELISAs employing repetitive peptides to detect antibodies to Plasmodium falciparum sporozoites.

Comparative analysis of ELISAs employing repetitive peptides to detect antibodies to Plasmodium falciparum sporozoites G. Del Giudice,' A. Douglas,2 J.P. Verhave,3 R.A. Wirtz,4 & F. Zavala5 In the last few years, a number of different recombinant and synthetic peptides consisting of the repetitive sequence of the Plasmodium falciparum circumsporozoite protein (NANP), have been produced and used to develop immunoassays for the detection of antibodies against P. falciparum sporozoites in human sera. A comparative study of three enzyme-linked immunosorbent assays (ELISAs) that employed different (NANP), peptides (the synthetic peptides (NANP)3 and (NANP)40 as well as the recombinantpeptides R32tet32 and R32LR) was carried out using serum samples from individuals who were living in different malaria-endemic areas. The results obtained for these peptide-based ELISAs were compared with those obtained foran immunofluorescence assay (IFA) that used glutaraldehyde-fixed sporozoites. All the methods tested exhibited 100% specificity on sera from persons not exposed to malaria, good reproducibility (coeffients of variation ranged from 3% to 15% for peptide-based ELISAs), and good sensitivity. Reproducibility and sensitivity were lower for the IFA than for the peptide-based ELISAs, perhaps because of the subjective element in the interpretation of the results which is inherent in the IFA method. ELISAs based on peptides that contain a higher number of (NANP) repeats, i.e., (NANP)40 and R32tet32 or R32LR, gave results which correlated better with each other than with those obtained with the ELISA that employed a shorter (NANP)3 peptide. (NANP),-based ELISAs are relatively simple and inexpensive methods for the detection of anti-P. falciparum sporozoite antibodies and can readily be used in epidemiological research in the field. These assays could contribute to a better understanding of the natural history of the host-parasite relationship in malaria research. Malaria sporozoites possess a major surface antigen, the circumsporozoite (CS) protein, which uniformly surrounds their external coat. Despite their short persistence in circulating blood, sporozoites induce a strong immune response that is characteristically species- and stage-specific (1). Plasmodial CS proteins consist of tandem repeats of a number of amino acids, flanked by nonrepetitive sequences (2). All the mono- clonal antibodies so far produced against sporozoites appear to recognize this repetitive sequence in the CS protein (3), as do naturally acquired antibodies from I WHO Immunology Research and Training Centre, Department of Pathology, University of Geneva, rue Michel Servet 1, 1211 Geneva 4, Switzerland. Requests for reprints should be sent to this address. 2 Department of Epidemiology and Population Sciences, London School of Hygiene and Tropical Medicine, London, England. 3Department of Medical Parasitology, University of Nijmegen, Nijmegen, Netherlands. 4Department of Entomology, Walter Reed Army Institute of Re- search, Walter Reed Army Medical Center, Washington, DC, USA. " Department of Medical and Molecular Parasitology, New York University, New York, NY, USA. Reprint No. S00s malaria-endemic areas (4). The repetitive domain of the Plasmodiumfalciparum CS protein consists of four amino acids (Asn-Ala-Asn-Pro =NANP) repeated several times (5, 6), which is well conserved in all of the isolates from different geographical regions so far investigated (7-9). This repetitive sequence, which has been produced by both genetic engineering methods (10) and chemical synthesis (4, 11), has been employed as the basis for the development of prototype malaria vaccines (12-14) and of assays for the detection of antisporozoite antibodies in human populations exposed to infection with falciparum malaria. Until recently, circumsporozoite precipitation and immunofluorescence assay (IFA) were the only tech- niques available for the detection of antisporozoite antibodies in serum (15,16). These techniques require a regular supply ofsporozoites, which in turn demands mosquito breeding facilities-which are available in very few laboratories. A number of different recom- binant and synthetic (NANP). peptides have been produced recently and used to develop immunoassays for the detection of antisporozoite antibodies in human sera. These peptide-based methods, mainly enzyme- linked immunosorbent essays (ELISAs), have been Bulletin of the World Health Organization, 67 (5): 515-523 (1989) © World Health Organization 1989 515 G. Del Gludlce et al. used in epidemiological research, and the results obtained are similar to those obtained with IFA (17-23). Since these assays employ recombinant or syn- thetic peptides that consist of different numbers of (NANP) repeats, it is interesting to determine whether all such peptides detect antisporozoite antibodies in the same manner. Here, we report the results of a comparative analysis of ELISAs that use different (NANP). peptides to detect anti-P. falciparum sporo- zoite antibodies in serum samples from individuals who were living in several malaria-endemic areas. The comparative study was carried out during a workshop held in Geneva under the auspices of the UNDP/ World Bank/WHO Special Programme for Research and Training in Tropical Diseases. Materials and methods Serum samples A total of 253 serum samples were tested by each ELISA method. The set of sera consisted of the following groups: (a) sera from individuals who were living in malaria-endemic areas, i.e., Brazil, Burkina Faso, Cameroon, Colombia, Honduras, Kenya, Mexico, Thailand, and West Africa; (b) sera from 16 healthy blood donors from Geneva; and (c) sera from five patients with systemic lupus erythematosus and five with rheumatoid arthritis (Table 1). Sera in groups (b) and (c) were included to test the specificity of each method. To test the sensitivity of the assays, one serum sample that was known to be positive for anti-P. falciparum sporozoite antibodies (obtained from P.C., a European who acquired acute P.falciparum malaria in the Cote d'Ivoire) was tested undiluted and predilu- ted at 1:2,1:4, 1:8, and 1:16 in a human serum that was unexposed to malaria. To test the reproducibility of the assays, a pool of 30 serum samples was prepared from individuals who were living in Nigeria and known to have anti-P. falciparum antibodies. Serum from this pool was divided between five different sample tubes, and the same was done for a pool of serum from healthy blood donors from Geneva. Pools were used since insufficient serum was available from individuals to use single donor sources. Each sample tube was coded by Dr L. Martinez in WHO, and aliquots of 100 Iu were prepared and stored at -70 'C until used. Methods for detecting anti-P. falclparum sporozolte antibodies The methods outlined below for the detection of anti-P. falciparum sporozoite antibodies were tested. Table 1: List of the serum samples that were assayed In the study No. of Origin Provided by: sera Malaria-endemic areas Brazil M. Arruda 20 Burkina Faso J.-B. Ouedraogo 22 J.-P. Verhave 41 Cameroon R. Moyou 6 Colombia M. Herrera 36 Honduras I.G. Gallo 19 Kenya G.M. Marangalla 16 Mexico L. Gonzales-Ceron 12 Thailand C. Khamboonruang 16 R. Rosenberg 10 West Africa" F. Zavala 14 Nigeriab G. Del Giudice 5 Positive controic G. Del Giudice 5 Malaria-nonendemic area (Geneva) Blood donors G. Del Giudice 16 Negative poold G. Del Giudice 5 SLE' patients G. Del Giudice 5 RA' patients G. Del Giudice 5 Total 253 ' The exact origin of these serum samples was not known. bA pool of 30 serum samples that were positive for anti-P/as- modium falciparum sporozoite antibodies, and assayed five times. c Positive control from a European who contracted P. falciparum malaria, assayed undiluted, and prediluted 1:2,1:4,1:8, and 1:16 in normal human serum. dA pool of 30 serum samples from healthy blood donors from Geneva, and assayed five times. I SLE =systemic lupus erythematosus. f RA = rheumatoid arthritis. e The IFA, which used glutaraldehyde-fixed P. falci- parum sporozoites (NF54 strain) from membrane-fed Anopheles stephensi mosquitos, was carried out as previously described (15). Sera that fluoresced at a 1:40 dilution were considered to be positive. The remaining methods were ELISAs that em- ployed different synthetic or recombinant peptides consisting ofvariable numbers ofthe repetitive (NANP) epitope of the P. falciparum CS protein, as discussed below. * The (NANP)3-BSA ELISA, in which the synthetic peptide (NANP)3 conjugated to bovine serum albu- min (BSA) [(NANP)3-BSA], was carried out as pre- viously described (18), the only modification being that a horseradish peroxidase anti-human IgG antibody was employed as a probe. Serum samples were tested at a dilution of 1:40 and the results determined at A=414 nm using a Multiskan Titertek micro-ELISA reader. * The R32tet32 ELISA, in which the recombinant peptide R32tet32 (consisting of the sequence (NANP)15- NVDP (NANP)1 NVDP, plus 32 amino acids that are 516 Comparison of ELISAs employing repetitive peptides encoded by a tetracycline-resistant gene read out-of- frame (tet32)), or the R32LR ELISA, in which the R32 LR recombinant lacks the last 30 amino acids of the tet32 moiety, but contains the first two, leucine and arginine (LR), was carried out as described previously (17, 22). Serum samples were tested at a dilution of 1:40, and the results determined at A =414 nm using a Multi- skan Titertek micro-ELISA reader. * The (NANP)40 ELISA, which uses the synthetic peptide (NANP)40, was carried out as described previously (19,20). Serum samples were tested at a dilution of 1: 200 and results determined at A=492 nm using a Multiskan Titertek micro-ELISA reader. Definition of titration units In order to standardize the results obtained with the different methods, a serum sample that was known to be positive for anti-P.falciparum sporozoite antibodies (from P.C.) was used as a positive control in each method. The absorbance obtained with this serum in each test was assigned a value of 100 titration units (TU), and all the absorbances obtained with the coded sera were proportionally adjusted to this value. Definition of the cut-off value After the code had been broken, the cut-off value for "positive" sera was defined as the mean +3 standard deviations of the value for the TU obtained for the 16 normal human sera included in the set of 253 samples tested. According to this procedure, the cut-off values were as follows: - IFA= 1:40 dilution; - (NANP)3-BSA ELISA= 15.14 TU; - R32tet32 ELISA = 13.88 TU; - R32LR ELISA=9.13 TU; and - (NANP)40 ELISA= 12.58 TU. Values at or above these cut-offs were considered to be positive for anti-P. falciparum sporozoite anti- bodies. Results Specificity Specificity was tested with a panel of sera from healthy blood donors from Geneva and from European patients with systemic lupus erythematosus or rheumatoid arthritis. As shown in Fig. 1, all sera from healthy blood donors gave results below the cut-offvalues with each method. Similarly, negative results were always obtained with sera from patients with systemic lupus erythematosus or rheumatoid arthritis, despite the presence of high titres of rheumatoid factors and other auto-antibodies. It can be concluded that the IFA and Fig. 1. Plots showing the specificity of the (NANP)-based ELISAs and the immunofluorescence assay (IFA) for the detectlon of antl-Plasmodlum alciperum sporozoite anti- bodies. Serum samples from 16 healthy blood donors from Geneva (a), from five patients with systemic lupus ery- thematosus (A) and from five patients with rheumatoid arthritis (o) were tested. Horizontal lines represent cut-off values. 20 C F-Z c 10 - 0 - -10 o - ;% °0 0 L LA 5H IL (NANP) 40 (NANP) 3 R32t et 32 R32LR ELISA ELISA ELISA ELISA 40 20 10 c. CA 0 IFA the peptide-based ELISAs for the detection of anti-P. falciparum sporozoite antibodies appear to be specific, since no false-positive results were obtained. Sensitivity Sensitivity was tested using a serum that was known to be positive for anti-P.falciparum sporozoite antibodies and which was progressively diluted in normal human serum. Fig. 2 shows that the (NANP)0-based ELISAs Fig. 2. Plots showing the sensitivity of the Immunofluor- escence assay (IFA) and the (NANP)-based ELISAs for the detection of anti-Plasmodlum fakiperum sporozolte anti- bodies. A serum sample (P.C.) that was known to be positive for antisporozolte antibodies was tested undiluted and pre- diluted 1:2, 1:4, 1:8 and 1:16 In normal human serum. *: (NANP). ELISA; A: (NANP),-BSA ELISA; 0: R32et32 ELISA; E: R32LR ELISA; and O: IFA. 120 -1 F80 100 - . 80- r 60- P 40- 20 - 0- - 60 -40 > C, cn -20 -0 1 2 4 8 16 Reciprocal of serum dilutions 517 0ol so 0 a_r,0 CPO Z.' 0 Ep a G. Del Gludice et al. gave titration curves with similar slopes. The results for the R32tet32 ELISA were always greater than the cut-off value. In contrast, the (NANP)3-BSA ELISA and the (NANP)40 ELISA gave positive results for the first four serum dilutions, while the fifth dilution was just less than the cut-off value. Finally, serum that was undiluted or prediluted 1:2 and 1:4 was positive by IFA, but negative when prediluted 1:8 or 1:16. Thus, IFA was less sensitive than the peptide-based ELISAs in detecting anti-P. falciparum sporozoite anti- bodies. The serum used to estimate the sensitivity of the assays was the same as that used as a positive control in each test. Fig. 2 shows that the ELISAs gave absorbances of approximately 100 TU for the undilu- ted sample, thus showing that the reproducibility of the results was good (see next paragraph). Reproducibility Reproducibility was determined with pools of sera that were positive or negative for anti-P.falciparum sporo- zoite antibodies, and the measurements repeated five times blind. The five serum samples from the positive pool and the five samples from the negative pool were always positive and negative, respectively, in the IFA and ELISAs (Fig. 3). The coefficient of variation was 15% for the R32LR ELISA, 7.4% for the (NANP)40 ELISA, and 2.7% for the (NANP)3-BSA ELISA, as determined with the pool of positive sera. The coef- ficient of variation could not be calculated for the R32tet32 ELISA, since the five serum samples from the positive pool always gave absorbances that were Fig. 3. Plots showing the reproducibility of the Immuno- fluorescence assay (IFA) and the (NANP)-based ELISAs for the detection of antI-Plasmodlum taiciperum sporozolte antibodies. Assays on one positive pool (e) from Nigerian and one negative pool (c) from Geneva subjects were repeated five times. 200 180 - 160 - 140- 120- 100 - 80- 20 1J0 10- ; 1. j i r§a - 640 - 320 160 - 80 - 40 - 20 - 10 - 0 IFA off-scale (> 2.0). Finally, the IFA exhibited a coefficient of variation of 31.3% (two of the five samples were scored as positive (1:640) and the other three as positive (1:40)), and this method therefore had the lowest reproducibility of those tested. Correlations between the results of the assays Comparison of the results of the assays (based on the measured absorbances) using a Spearman rank corre- lation indicated that all the correlations were statis- tically significant (P<0.001) (Table 2). The strongest correlation was observed between the R32tet32 ELISA and the (NANP)40 ELISA (R=0.87) and the weakest between the (NANP)3-BSA ELISA and the (NANP)40 ELISA (R= 0.61). It is interesting to note that the correla- tion between the R32tet32 ELISA and the (NANP)40 ELISA was similar to that between the R32tet32 ELISA and the R32LR ELISA (R = 0.86). Similar results were observed when the assays were compared in terms of their ability to agree or disagree, i.e., give "positive" or "negative" results (Table 3). As expected, the best agreement was between the R32tet32 ELISA and the R32LR ELISA (82.7%), followed by that between the R32tet32 ELISA and the (NANP)40 ELISA (79.8%); the lowest was between the R32tet32 ELISA and the (NANP)3-BSA ELISA (59.1 %). Comparison of the results of the ELISA methods with those of the IFA indicated that the best agreement was with the (NANP)3-BSA ELISA (80.8%), and the lowest was with the R32tet32 ELISA (63%). The R32tet32, R32LR, and (NANP)40 ELISAs detected as positive (and in some instances, as highly positive) those sera from malaria-endemic areas that were scored as negative by IFA and the (NANP)3-BSA ELISA. In one case, only the R32tet32 ELISA gave a high TU for a serum that was negative by the (NANP)40 ELISA. These results suggest that the ELISAs with the longer peptides, i.e., the R32tet32, R32LR, and (NANP)40 ELISAs, may have a higher sensitivity than the ELISA with the shorter peptide, i.e., the (NANP)3 ELISA. Detection of anti-P. falciparum sporozolte anti- bodies In paired mother-cord serum samples Sixteen paired mother-umbilical cord serum samples from Burkina Faso were included in the set of coded sera that were tested. As shown in Table 4, there was good overall agreement between the results obtained with the ELISAs. For 16 of the 32 sera, the results obtained (either positive or negative) agreed with all the assays. However, for 23 (72%) of the sera a mixture of positive and negative results were found in at least four assay methods. The mean difference in the TU 518 (NANP) 40 (NANP) 3 R32t et 32 R32LR ELISA ELISA ELISA ELISA I, "O" w Z c *X Comparison of ELISAs employing repetitive peptides Table 2: Results of Spearman rank correlations (R) for the ELISAs and IFA that were usd to detect anti-Plasmodlum faiciperum sporozoite antibodies In the sudy ELISA method Method (NANP)3-BSA R32tet32 (NANP)40 R32LR IFA n=249; n=249; n=249; n=249; R= 0.65; R=0.68; R= 0.70; R= 0.70; (P< 0.001) (P< 0.001) (P< 0.001) (P< 0.001) (NANP)3-BSA ELISA - n= 249; n= 253; n= 249; R=0.66; R= 0.61; R= 0.74; (P<0.001) (P< 0.001) (P< 0.001) R32tet32 ELISA - n=249; n=249; R= 0.87; R= 0.86; (P<0.001) (P<0.001) (NANP)4,0 ELISA - - n= 249; R=0.76; (P< 0.001) Table 3: Percentage of tests that agreed or dlagreed (as expresed by whether they gave "positive" or "negative" result) for four of the assays In the study ELISA method (NANP)3-BSA R32tet32 (NANP)40 Method +ve -ve +ve -ve +ve -ve IFA f +ve 26 13 37 1.9 35.1 3.81-ve 6.3 54.8 35.1 26 27.4 33.7 (NANP)3-BSA ELISA +{ve 31.7 0.5 31.1 0.51-ye 40.4 27.4 30.2 38.2 R32tet32 ELISA {+ve 57.2 14.9 -ve 5.3 22.6 R32LR ELISA +ve 30.8 1.4 56.3 15.9 48.1 14.4R3-ve 26.9 40.9 1.4 26.4 9.6 27.9 The results for tests that agreed (either both positive or both negative) are shown in italics. values of the sera from the mother-umbilical cord pairs was not statistically significant for the R32tet32 ELISA, the R32LR ELISA, the (NANP)3-BSA ELISA, and the IFA (paired Student's t-tests); however, this difference was statistically significant for the (NANP)40 ELISA (Table 5). This ELISA employs an anti-IgG, IgA, IgM antibody probe, whereas the other ELISAs use anti-IgG antibody probes, and it is possible that the difference could arise because antisporozoite IgM antibodies are detected by the (NANP)40 ELISA in the serum samples from mothers, but not in those from umbilical cords. Discussion Although Vandenberg et al. demonstrated in 1969 the existence of an antibody response to malaria sporo- zoites by using the CS precipitation assay (24), it was not until the introduction of the more sensitive IFA by Nardin et al. (15) that anti-P. falciparum sporozoite antibodies could be detected in sera from subjects living in areas that are hyperendemic for malaria. However, because of the difficulty in obtaining the antigen used in this IFA, this technique was of limited use for epidemiological studies (15,16,25,26). The recent availability of synthetic and recom- binant peptides from the repetitive domain of the P. falciparum CS protein has, however, made it possible to develop relatively simple, sensitive assays for the detection of antibodies against sporozoites. Unlike IFA, such assays require neither a source of sporo- zoites nor sophisticated equipment and can therefore be used readily in epidemiological research on indivi- duals naturally exposed to malaria infection. They are 519 G. Del Gludlce et al. Table 4: Results for the detetIon of antl-Plasmodlum faklparum sporozolte antibodies In the 16 pairs of mother-umbilical cord serum samples assayed In the study IFA' ELISA methodb (reciprocal (in TU) serum dilution) (NANP)3-BSA R32tet32 (NANP)4, R32LR Mc 20 2.61 23.94 16.75 17.68 8 Ud 20 9.98 23.94 6.27 29.27 M 20 -17.26 11.27 24.22 9.76 28 U 0 -1.49 2.82 6.51 3.05 M 1280 122.20 140.85 148.55 122.00 29 U 1280 148.51 140.85 117.23 122.00 M 20 2.61 33.10 30.36 20.12 38 U 0 1.49 7.04 5.66 5.49 M 0 0.84 14.08 11.08 3.05 50 U 40 1.21 7.04 5.06 3.66 M 40 9.51 61.97 30.48 52.44 55 U 160 21.50 50.70 28.31 37.20 M 40 16.42 36.62 23.37 22.56 116 U 160 29.10 34.51 18.67 25.00 M 40 13.62 35.92 11.33 21.34 131 U 20 12.50 28.87 12.05 22.56 M 20 4.29 26.06 19.28 12.20 155 U 40 6.06 21.13 8.67 14.02 M 640 72.11 77.46 41.93 85.37 164 U 160 64.83 73.94 41.93 70.12 M 160 32.46 21.83 66.99 7.93 178 U 160 47.29 53.52 27.47 46.95 M 40 -5.22 24.65 18.07 13.41 290 U 0 -1.40 6.34 8.67 5.49 M 40 17.44 40.14 29.86 30.49 300 U 40 15.67 40.14 24.70 25.61 M 40 22.57 47.89 36.75 41.46 319 U 80 28.17 59.15 19.28 51.83 M 0 60.17 76.06 61.33 78.05 321 U 40 56.62 97.18 59.40 78.05 M 20 18.84 78.87 47.11 32.32 498 U 0 5.32 19.72 14.34 -1.83 a Positive results are shown in italics. bPositive results (those above the cut-off values) are shown in italics. c M=serum sample from the mother; dU=serum sample from the umbilical cord. also likely to have a valuable role in trials of antisporo- zoite vaccines. One of the aims of the study was to compare the results obtained with ELISAs based on (NANP)n constructs as well as with those obtained with IFA, which was until recently the only method available for the detection of anti-P. falciparum sporozoite anti- bodies. All the assays tested showed 100% specificity, based on serum samples from Geneva, since no false-positive results were obtained. It is noteworthy that the presence of high titres of rheumatoid factors and other autoantibodies in sera from individuals with systemic lupus erythematosus or rheumatoid arthritis did not interfere with the specificity of the assays. The reproducibility of the (NANP)n-based ELISAs that were tested (the (NANP)3-BSA, R32tet32, R32LR, and (NANP)40 ELISAs) was good, with a coefficient of variation for positive pools of sera that ranged from 3% to 15%. However, the IFA was much less repro- ducible than the peptide-based ELISAs. This can be ascribed to the element of subjectivity in interpreting the result of the IFA method. When sensitivity was tested by progressively diluting a positive serum, the IFA had the lowest sensitivity, since it detected as positive only the first three dilutions of the positive control serum. Only 63% and 69% of the results obtained with 520 Comparison of ELISAs employing repetftive peptides Table 5: Statistical comparison (paired Student's f-tests) of the result obtained for the assay of the 16 pairs of mother-umbilical cord serum samples In the study ELISA method (in TU) Statistical IFA (NANP)3-BSA R32tet32 (NANP)40 R32LR parameter (logio titre) Moan value { Ms 1.53 23.33 46.92 38.59 35.63Mean valueUb 1.43 27.84 41.68 25.26 33.65 Standard deviation fM 0.80 34.81 33.42 33.59 33.22 l.U 0.96 38.28 37.49 28.69 33.63 t value 0.38 -1.81 1.05 4.21 0.51 Two-tail probability 0.711 0.091 0.311 0.001 0.621 M =serum sample for the mother. b U=serum sample from the umbilical cord. the ELISAs that used the R32tet32 or (NANP)40 peptides, respectively, agreed with those obtained with the IFA, whereas those for the ELISA that used the (NANP)3 peptide there was 81% agreement with those for the IFA. The R32tet32 ELISA and (NANP)40 ELISA exhibited the highest degree of agreement (80%), and both gave results that agreed less with those of the (NANP)3-BSA ELISA (59% for the R32tet32 ELISA, 69% for the (NANP)40 ELISA, and 71% for the R32LR ELISA). Several interpretations of these results are possible. First, it appears that peptide-based ELISAs are more sensitive than the IFA in detecting anti-P. falciparum sporozoite antibodies in human sera. This could be due to the following technical problems inherent in the IFA procedure: sporozoite fixation with glutaraldehyde, storage of the slides, subjective interpretation of the results, etc. It is not known whether IFAs that use unfixed living sporozoites (26) give results that exhibit closer agreement with those obtained with peptide-based ELISAs. Alternatively, it cannot be excluded that synthetic and recombinant peptides express new or modified epitopes which may recognize larger, and in some cases non-specific, antibody populations that are not detected by IFA. Second, the results of the ELISAs based on peptides that consist of a large number of (NANP) repeats (R32tet32, R32LR, and (NANP)40) correlated more closely with each other than with the results of the ELISA based on the shorter (NANP)3 peptide, which failed to detect as positive some sera that were positive in the other ELISAs. This could arise because the R32tet32 ELISA and (NANP)40 ELISAs are more sensitive for the reasons outlined below. e The R32tet32, R32LR, and (NANP)40 peptides contain a higher number of(NANP) sequences, which could increase the sensitivity of the corresponding assays. However, in this case, the sensitivity of the (NANP)3-BSA ELISA should be increased by raising the concentration of peptide used to coat the micro- titration plates. e The R32tet32 and R32LR peptides contain two (NVDP) repeats, which may detect specific antibodies that are not detected by synthetic (NANP)n-based ELISAs. However, the results obtained in the ELISAs that use R32tet32 or (NANP)40 (which does not contain (NVDP) repeats) are very similar. Only one serum (from Colombia) was highly positive in the R32tet32 ELISA and negative in the (NANP)40 ELISA. * The R32tet32, R32LR, and (NANP)40 peptides may express conformational epitopes, i.e., epitopes not simply arising from the primary structure of the amino acid sequence, which may be absent in the shorter (NANP)3 peptide. This would seem to be the most likely explanation for the discrepancies in the results between the (NANP)3-BSA ELISA and the R32tet32 and (NANP)40 ELISAs. However, more research is required to define such "conformational" epitopes in the (NANP) sequence and to determine the presence and relevance (ifany) ofanti4NVDP) antibodies in the sera of subjects living in areas where P.falciparum malaria is endemic. (NANP)n-based ELISAs for the detection of anti-P.falciparum sporozoite antibodies are relatively simple and inexpensive methods that are readily applicable in epidemiological research in the field. Similar methodology is now also being used to detect antibodies against P. vivax sporozoites (27). Further- more, the results of the recent analysis of the amino acid sequence of the P. malariae CS protein (28) should soon permit the development of assays for the detec- tion of anti-P. malariae sporozoite antibodies. Such assays could contribute to a better understanding of the natural history ofthe host-parasite relationships in 521 G. Del Gludice et al. the epidemiology of malaria. Finally, the detection of antisporozoite antibodies is of potential value to malaria control programmes, since it provides an index of the degree of malaria transmission (26). Acknowledgements This work was supported by the UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases. We would like to thank Smith Kline & French, USA, for the R32tet32 and R32LR peptides and Eniricerche, Italy, for the (NANP)40 peptide; W. Roeffen, C. Tougne, H. Wang, and H.C. Wynn for superb technical assistance; Dr L. Martinez, Dr L. Molineaux, Dr A. Muir, and Dr P. Smith for suggestions and advice; Dr M. Arruda, Dr J.-B. Ouedraogo, Dr R. Moyou, Dr L.-X. Zhang, Dr M. Herrera, Dr I.G. Gallo, Dr S.K. Subbarao, Dr G.M. Marangalla, Dr 0. Doumbo, Dr L. Gonzales-Ceron, Dr P.J. Herath, Dr Y.G. Matola, Dr C. Khamboonruang, and Dr H.N. Doan who contributed to the establishment of the panel of sera that were tested. Professor P.-H. Lambert for encouragement and advice; Dr L. Martinez for helpful suggestions in the preparation of the manuscript; and Mrs F. Adam for secretarial assistance. Resume Evaluation comparative de trols ELISA em- ployant des peptides rep6titifs pour detecter les anticorps dirlges contre les sporozoites de Plasmodlum falciparum Ces dernieres ann6es, on a prepare par g6nie gen6tique ou synthese un certain nombre de pep- tides constitues de la s6quence repetitive (NANP)n de la prot6ine circumsporozoftaire de Plasmodium falciparum. Ces peptides ont servi i mettre au point des m6thodes immunologiques de d6tection des anticorps dirig6s contre les sporozoftes de P. falci- parum dans le serum humain. Trois m6thodes de titrage immuno-enzymatique (ELISA) employant dif- ferents peptides (NANP)n (les peptides synthetiques (NANP)3 et (NANP)40 et les peptides recombinants R32tet32 et R32LR) ont fait l'objet d'une etude comparative au cours de laquelle on a analyse des echantillons de serums d'individus vivant dans differentes regions ou' le paludisme est endemique. Les resultats de ces ELISA ont ete compares a ceux obtenus par une technique d'immunofluorescence (IFA) utilisant des sporozoites fixes par le glutaralde- hyde. Toutes les methodes ont montre une specifi- cite de 100% sur les serums de personnes n'ayant jamais ete exposees au paludisme, une bonne reproductibilite (coefficient de variation compris entre 3% et 15% pour les ELISA) et une bonne sensibilit6. La reproductibilite et la sensibilit6 ont ett plus faibles pour l'IFA que pour les ELISA, peut-6tre en raison de l'6l6ment subjectif inherent A I'interpretation des r6sultats de la m6thode IFA. La correlation des r6sultats a et6 meilleure entre les ELISA fond6s sur l'utilisation de peptides compre- nant un grand nombre de sequences (NANP), c'est-a- dire (NANP)40, R32tet32 et R32LR, qu'entre ceux-ci et l'ELISA fonde sur l'utilisation du peptide plus court (NANP)3. Les methodes ELISA fond6es sur l'utilisation des peptides (NANP)n constituent un moyen relative- ment simple et peu couteux de detecter les anti- corps diriges contre les sporoz6iles de P. falciparum et sont facilement applicables a la recherche epidemiologique sur le terrain. Ces essais pour- raient aider c mieux comprendre l'histoire naturelle de la relation h6te-parasite dans I'etude du paludisme. References 1. Cochrane, A.H. et al. Immunization against sporo- zoites. In: Kreier, J.P., ed. Malaria, vol. 3. New York, Academic Press, 1980, pp. 163-202. 2. Nuesnzwelg, V. & Numenzwelg, R.S. Circumsporo- zoite proteins of malaria parasites. Cell, 42: 401-403 (1985). 3. Zavala, F. et al. Circumsporozoite proteins of malaria parasites contain a single immunodominant region with two or more identical epitopes. Journal of experi- mental medicine, 157: 1947-1957 (1983). 4. Zavala, F. et al. Rationale for the development of a synthetic vaccine against Plasmodium falciparum malaria. Science, 228: 1436-1440 (1985). 5. Dame, J.B. et al. Structure of the gene encoding the immunodominant surface antigen on the sporozoite of the human malaria parasite Plasmodium falciparum. Science, 225: 593-599 (1984). 6. Enea, V. et al. DNA cloning of Plasmodium falciparum circumsporozoite gene: amino acid sequence of repetitive epitope. Science, 225: 628-630 (1984). 7. Lockyor, M.J. & Schwarz, R.T. Strain variation in the circumsporozoite gene of Plasmodium falciparum. Molecular and biochemical parasitology, 22: 101-108 (1987). 8. De Ia Cruz, V.F. et al. Sequence variation in putative functional domains of the circumsporozoite protein of Plasmodium falciparum. Implications for vaccine development. Journal of biological chemistry, 262: 11935-11939 (1987). 9. Del Portillo, H.A. et al. Circumsporozoite gene of a Plasmodium falciparum strain from Thailand. Mole- cular and biochemical parasitology, 24: 289-294 (1987). 10. Young, J.F. et al. Expression of Plasmodium falci- parum circumsporozoite protein in Escherichia coli for potential use in a human malaria vaccine. Science, 228: 958-962 (1985). 11. Ballou, W.R. et al. Immunogenicity of synthetic pep- 522 Comparison of ELISAs employing repetitive peptides tides from circumsporozoite protein of Plasmodium falciparum. Science, 228: 996-999 (1985). 12. Ballou, W.R. et al. Safety and efficacy of a recombinant DNA Plasmodium falciparum sporozoite vaccine. Lancet, 1: 1277-1281 (1987). 13. Herrlngton, D.A. et al. Safety and immunogenicity of a synthetic peptide malaria vaccine against Plasmo- dium falciparum sporozoites. Nature, 328: 257-259 (1987). 14. Etllnger, H.M. et al. Assessment of a synthetic peptide- based vaccine against the sporozoite stage of the human malaria parasite Plasmodium falciparum. Journal of immunology, 140: 626-633 (1988). 15. Nardin, E.H. et al. Antibodies to sporozoites: their frequent occurrence in individuals living in an area of hyperendemic malaria. Science, 206: 597-599 (1979). 16. Tapchal.rl, P. etal. Anti-sporozoite antibodies induced by natural infection. American journal of tropical medicine and hygiene, 32: 1203-1208 (1983). 17. Hoffman, S.L. et al. Immunity to malaria and naturally acquired antibodies to the circumsporozoite protein of Plasmodium falciparum. New England journal of medicine, 315; 601-606 (1986). 18. Zavala, F. et al. Synthetic peptides for the detection of humoral immunity to Plasmodium falciparum sporo- zites. Journal of immunological methods, 93: 55-61 (1986). 19. Del Gludlce, G. et al. Detection of human antibodies against Plasmodium falciparum sporozoites using synthetic peptides. Journal of clinical microbiology, 25: 91-96 (1987). 20. Del Gludlce, G. et al. Antibodies to the repetitive epitope of Plasmodium falciparum circumsporozoite protein in a rural Tanzanian community: a longitudinal study of 132 children. American journal of tropical medicine and hygiene, 36: 203-212 (1987). 21. Campbell, G.H. et al. Detection of antibodies in human sera to the repeating epitope of the circumsporozoite protein of Plasmodium falciparum using the synthetic peptide (NANP)3 in an enzyme-linked immunosorbent assay (ELISA). American journal of tropical medicine and hygiene, 37: 17-21 (1987). 22. Webster, H.K. et al. Development of immunity in natural Plasmodium falciparum malaria: antibodies to the falciparum sporozoite vaccine 1 antigen (R32tet32). Journal of clinical microbiology, 25: 1002-1008 (1987). 23. Hoffman, S.L. et al. Naturally acquired antibodies to sporozoites do not prevent malaria: vaccine develop- ment implications. Science, 237: 639-642 (1987). 24. Vandenberg, J.P. et al. Protective immunity produced by the injection of X-irradiated sporozoites of Plas- modium berghei. V. In vitro effects of immune serum on sporozoites. Military medicine, 134 (suppl.): 1183-1190 (1969). 25. Tapchalsrl, P. et al. Antibodies against malaria sporo- zoites in patients with acute uncomplicated malaria and in patients with cerebral malaria. American jour- nal of tropical medicine and hygiene, 34: 831-836 (1985). 26. Drullhe, P. et al. Levels of antibodies to Plasmodium falciparumsporozoite surface antigens reflect malaria transmission rates and are persistent in the absence of reinfection. Infection and immunity, 53: 393-397 (1986). 27. Lee, M. et al. Interaction of Malaysian sera with Plasmodium vivax sporozoite antigen. American jour- nal of tropical medicine and hygiene, 39: 535-539 (1988). 28. Lal, A.A. et al. Structure of the circumsporozoite protein gene of Plasmodium malaria. Molecular and biochemical parasitology, 30: 291-294 (1988). 523