Abstract
Infants born in areas of stable malaria transmission are relatively protected against severe morbidity and high density Plasmodium falciparum blood-stage infection. This protection may involve prenatal sensitization and immunologic reactivity to malaria surface ligands that participate in invasion of red cells. We examined cord blood T and B cell immunity to P. falciparum merozoite surface protein-1 (MSP-1) in infants born in an area of stable malaria transmission in Kenya. T cell cytokine responses to the C-terminal 19-kDa fragment of MSP-1 (MSP-119) were detected in 24 of 92 (26%) newborns (4–192 IFN-γ and 3–88 IL-4-secreting cells per 106/cord blood lymphocytes). Peptide epitopes in the N-terminal block 3 region of MSP-1 also drove IFN-γ and/or IL-13 production. There was no evidence of prenatal T cell sensitization to liver-stage Ag-1. A total of 5 of 86 (6%) newborns had cord blood anti-MSP-119 IgM Abs, an Ig isotype that does not cross the placenta and is therefore of fetal origin. The frequency of neonatal B cell sensitization was higher than that indicated by serology alone, as 5 of 27 (18%) cord blood samples contained B cells that produced IgG when stimulated with MSP-119 in vitro. Neonatal B cell IgG responses were restricted to the Q-KNG allele of MSP-119, the major variant in this endemic area, whereas T cells responded to all four MSP-119 alleles evaluated. In utero sensitization to MSP-1 correlated with the presence of malaria parasites in cord blood (χ2 = 20, p < 0.0001). These data indicate that prenatal sensitization to blood-stage Ags occurs in infants born in malaria endemic areas.
Pregnancy is associated with an increased risk of maternal Plasmodium falciparum malaria and predisposes the developing fetus to stunted growth (1). The pathogenesis of both conditions is thought to involve cytoadherence and sequestration of infected RBCs in the placenta (2). Accumulation of infected red cells at this interface between the maternal and fetal circulation may result in congenital malaria and exposure of the fetal immune system to malaria parasites and/or soluble malaria Ags. Although direct examination of fetal immune responses to malaria Ags before birth is not feasible or ethically acceptable in humans, clinical and epidemiologic observations suggest that immunity to blood-stage P. falciparum is engendered in utero. Infants born in areas of stable malaria transmission frequently have enlarged spleens (3), and autopsy studies of African newborns indicate that their spleen weights are greater than those of infants born in areas in which malaria is not transmitted (4). Infants younger than ∼6 mo of age also appear to be relatively protected against severe malaria morbidity, mortality, and high density asexual parasitemia compared with older children (5). The mechanism of this resistance is unclear. From an immunologic perspective, attention has been focused on the possible contribution of maternal IgG Abs acquired by the fetus during gestation (6, 7, 8). In contrast, there is relatively little knowledge of whether or how prenatal T or B cell sensitization to P. falciparum may contribute to malaria susceptibility during the perinatal period, or if boosting through natural infection in the several months after birth affects susceptibility later in infancy.
Results of studies of rodent malaria models are consistent with the notion that immunologic experience with malaria Ags during fetal development affects susceptibility during infancy. Offspring of Plasmodium berghei-infected pregnant rats become sensitized to blood-stage Ags in utero, and lymphocytes from newborn animals transfer malaria-specific immunity to naive rats (9, 10). Study of analogous responses in humans is logistically challenging and limited to examination of humoral and cellular immune responses in cord blood, which reflects the pool of circulating fetal lymphocytes at birth. The presence of IgM or IgE Abs to malaria Ags in cord blood constitutes prima facie evidence of in utero sensitization since, unlike IgG, these Ig isotypes do not cross the placenta from the maternal circulation (8, 11, 12). There have been relatively fewer studies of cord blood T cell responses to malaria Ags (12, 13, 14) and to our knowledge, no descriptions of cord blood B cell responses. These investigations, which reported lymphocyte proliferation, IL-2, IL-4, and IFN-γ responses to lysates of infected red cells (i.e., schizont extracts), recombinant proteins, or synthetic antigenic peptides, suggest that a small proportion of infants become sensitized to malaria Ags during gestation. For example, Fievet et al. (14) reported that 9.8% of 164 Cameroonian newborns had cord blood lymphocyte (CBL)3 IFN-γ responses to native P. falciparum ring-infected erythrocyte surface Ag (Pf155/RESA). Response rates to synthetic peptides corresponding to RESA ranged from 1.3 to 6.3%. CBL from control neonates born in areas in which malaria was not transmitted were not used to evaluate whether the relatively weak responses reported were nonspecific.
The current study presents evidence for prenatal T and B cell sensitization to the malaria vaccine candidate Ag merozoite surface protein-1 (MSP-1) in infants born in a malaria holoendemic area of Kenya. Paired samples of maternal and CBL were evaluated for cytokine and Ig responses to four alleles of the C-terminal 19-kDa fragment of the Ag (MSP-119). Two synthetic peptides were also used to determine whether this immunity extended to the N-terminal region of MSP-1. Neonatal responses to MSP-1 were compared with those stimulated by liver-stage Ag-1 (LSA-1), a P. falciparum Ag expressed in infected hepatocytes following exposure to the bites of sporozoite-containing mosquitoes.
Materials and Methods
Study population
Paired cord and maternal blood samples were collected at Msambweni Hospital (Kwale District, Coast Province, Kenya). Pregnant women were recruited from the antenatal clinic. Malaria is holoendemic in this area of Kenya. At two points separated by 6 mo during the year the current study was conducted, the prevalence of P. falciparum infection determined by thick blood smears from 1105 residents was 39 and 55% in 5–12 year olds and 12 and 24% in subjects >16 years old. P. falciparum was the dominant species, although Plasmodium malariae infection is also present (15).
Maternal venous blood obtained within 24 h of parturition and cord blood from full-term newborns of uncomplicated pregnancies was anticoagulated with heparin. Control cord blood was obtained from healthy North American newborns and adults who had never traveled outside the United States. Plasma was stored at −70°C until Ab assays were performed.
Ethical approval was obtained from the Human Investigations Institutional Review Boards at the Kenya Medical Research Institute and University Hospitals (Case Western Reserve University, Cleveland, OH). All mothers provided witnessed oral consent for participation.
Ags and mitogens
Recombinant P. falciparum MSP-119 (kindly provided by D. Kaslow, Vical, San Diego, CA) corresponding to four variants designated as the Q-KNG, E-KNG, Q-TSR, and E-TSR alleles was expressed in Saccharomyces cerevisiae, affinity purified, and demonstrated to have the correct conformation, as described previously (16). This nomenclature refers to constructs containing the amino acid Q or E at codon 1644, and KNG or TSR at codons 1691, 1700, and 1701. The concentration of endotoxin in preparations used for evaluation of T and B cell responses was <0.5 ng/ml, which is 5- to 50-fold less than that required for stimulation of cytokine production by human lymphocytes.
To assess T cell responses to the N-terminal region of MSP-1, peptides designated P2 (GYRKPLDNIKDNVGKMEDYIKK; codons 250–271) and P3 (KLNSLNNPHNVLQNFSVFFNK; codons 1101–1121) were used (17). LSA-1 peptides corresponding to aa residues 84–107 (LTMSNVKNVSQTNFKSLLRNLGVS) and 1888–1909 (DNEILQIVDELSEDITKYFMKL) (NF54 strain of P. falciparum, GenBank X56203) (18) were used to evaluate whether CBL responded to malaria antigenic peptides not expressed in red cells. Peptides were synthesized and purified by HPLC to greater than 85% purity (Sigma Genosys, St. Louis, MO). PMA plus ionomycin (Calbiochem, La Jolla, CA) were used in parallel cultures as the positive mitogen control. Only preparations that responded to the mitogen are reported (19).
Isolation of PBMC and culture conditions for in vitro T cell cytokine production
All studies were performed using maternal PBMC and CBL separated from whole blood by density gradient centrifugation on Ficoll-Hypaque. The cells were resuspended in RPMI 1640 supplemented with 10% FCS, 4 mM l-glutamine, 25 mM HEPES, and 80 μg/ml gentamicin (C-RPMI; BioWhittaker, Walkersville, MD). Duplicate aliquots of lymphocytes at varying concentrations were incubated at 37°C in 5% CO2 in 0.5 ml C-RPMI under the following conditions: 1) media alone; 2) 10 μg/ml MSP-119 Q-KNG, E-KNG, Q-TSR, or E-TSR; and 3) 50 ng/ml PMA plus 1 μg/ml ionomycin. In separate studies, cells were stimulated with 10 μg/ml MSP-1 P2 or P3 peptide.
ELISPOT and ELISA for cytokine production
ELISPOT assays were similar to those previously described (19). T-spot plates (Athersys, Cleveland, OH) were coated with capture Abs in sterile PBS overnight at 4°C and blocked with C-RPMI with 10% FCS. Plates were then washed three times with sterile PBS. Cell suspensions were prepared, counted, and incubated at 2 × 106/ml for 48 h in 0.5 ml with Ag and for 24 h with mitogen. The supernatants were removed, fresh media added, and cells transferred to ELISPOT plates coated with anti-IFN-γ (M-700A, at 4 μg/ml; Endogen, Cambridge, MA) or anti-IL-4 (18651-D at 4 μg/ml; BD PharMingen, Palo Alto, CA) to capture the cytokine. The number of cells added to each well for Ag-specific cytokine production was 0.5 × 106 and 0.1 × 105 for mitogen-driven production. Cells were then cultured for 12–16 h at 37°C in a 5% CO2 in air. The plates were finally washed three times with PBS, followed by three washes with PBS-Tween (0.05%) to remove lymphocytes. Paired detecting mAbs for each cytokine were subsequently added as follows: 1) biotinylated anti-IFN-γ (M-701B, 2 μg/ml; Endogen) and 2) anti-IL-4 (18502-D, 2 μg/ml; BD PharMingen, Palo Alto, CA). The plates were incubated overnight at 4°C and then washed again three times in PBS-Tween. Streptavidin-HRP diluted 1/2000 in PBS (DAKO, Carpenteria, CA) was added, and incubation continued for 2 h at room temperature. After washing, substrate 3-amino-9-ethylcarbazol (Sigma) was added, and the resulting spots were enumerated by image analysis (ImmunoSpot Analyzer; CTL Technology, Cleveland, OH). ELISA for quantification of IFN-γ and IL-13 in culture supernatants incubated for 72 h in media alone or with P2 or P3 peptide was performed as described (20).
Depletion and enrichment of CD4+ and CD8+ T cells
CBL and PBMC were washed once in cold RPMI with 2% FCS. CD4+ and CD8+ cell depletion/enrichment was performed using magnetic beads directly conjugated to anti-CD4 or anti-CD8 Abs (Dynal, Lake Success, NY). Immunomagnetic depletion was performed according to the manufacturer’s instructions and routinely removed >88% CD4+ cells and >85% CD8+ from nonfractionated PBMC. The purity of positively selected CD4+ or CD8+ T cells was >95% (data not shown). Autologous nonfractionated cells (105/ml) were added to preparations enriched for CD4+ or CD8+ T cells (2 × 106/ml) as APCs. Preparations depleted of CD4+ and CD8+ cells were suspended to 2 × 106/ml under identical conditions.
MSP-119-driven IgG production by B cells
CBL were incubated in Iscove’s DMEM supplemented with 10% FCS, 4 mM l-glutamine, 25 mM HEPES, 80 μg/ml gentamicin, and ITS (insulin, transferrin, and selenium; BioWhittaker) at a density of 2 × 106 cells/ml in a total volume of 1 ml. Cultures containing media alone, 1 and 10 μg/ml MSP-119, or a 1/200 dilution of pokeweed mitogen were incubated at 37°C in 5% CO2 for 7–10 days (21). Supernatants were immediately frozen at −70°C for subsequent determination of polyclonal and MSP-1-specific IgG. Data expressed are peak Ag-driven IgG measured by ELISA. Only experiments in which there was significant production of pokeweed mitogen-driven IgG were included in the analysis.
MSP-119-specific Ab levels
Ab to MSP-119 was measured by ELISA (16). Briefly, Immulon 4 plates were coated with 200 ng/ml rMSP-119 (a separate plate was used for each of the four variants) in borate buffer (pH 7.4). After blocking and washing, plasma diluted 1/200 was added and incubation continued for 1 h at room temperature. The plates were washed, and alkaline phosphatase-conjugated anti-human IgG, IgM, or IgE (Jackson ImmunoResearch, Malvern, PA) was added at 1 μg/ml for 1 h at 37°C. Substrate p-nitrophenyl phosphate (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was added after the final wash. The reaction was stopped by adding 5% EDTA, and absorbance was read at 405 nm with an ELISA reader. A pool of plasma from adult Kenyans with high titers of Abs to MSP-119 was used as standard. The value obtained with a 1/200 dilution of this pool was arbitrarily designated as 1 U. Both high and low internal controls were run on each plate. Plasma obtained from 9 North American cord blood samples and 11 North American adults without a history of travel to malaria endemic areas were run in an identical fashion and served as negative controls. All negative samples were repeated at a dilution of 1/50. A positive response wasdefined as an OD greater than the mean plus 3 SDs of plasma from nine North American newborns (for cord blood) or eleven North American adults.
PCR diagnosis of P. falciparum and microsatellite genotyping
DNA was extracted from red cell pellets of maternal and cord blood and amplification of small subunit ribosomal DNA conducted using nested PCR for P. falciparum-specific sequence, as described previously (22, 23). PCR was performed using a Peltier Thermal Cycler, PTC-100 (MJ Research, Watertown, MA). Negative and positive controls were run for all nest 1 and nest 2 PCR reactions. Agarose gel electrophoresis was performed to visualize the amplicons. Negative controls always lacked amplicons. DNA was extracted from CBL and maternal lymphocytes, and microsatellite genotyping was performed, as described.
Sequencing
PCR products for MSP-119 using previously described primers (24) were purified from low melting agarose gel, cloned, and sequenced.
Statistics
Results are expressed as the mean ± SEM using log-transformed data, unless otherwise stated. Differences between experimental groups were evaluated by Student’s t test of log-transformed data or by paired t test, where appropriate. Differences between proportions were examined by χ2 analysis. Correlation coefficients were calculated using Pearson’s rank correlation test.
Results
Admixture of maternal and fetal circulation in cord blood samples
Because the objective of this research was to determine whether the fetus is sensitized to malaria Ags in utero, it was important to confirm that cord blood had not become mixed with maternal blood at the time of vaginal delivery. Two approaches were taken to assess this issue. First, polyclonal IgE levels in paired samples of maternal and fetal blood were compared. Since IgE does not cross the placenta, elevated levels in cord blood would suggest that admixture of the maternal and fetal circulation had occurred and/or that the fetus had produced significant amounts of IgE in utero. This criterion is a particularly sensitive indicator of admixture in this population since elevated IgE levels secondary to chronic helminthic infections are common in adults living in this area of Kenya. IgE levels in maternal blood ranged from 1,432 to 29,256 ng/ml (the range of age-matched North Americans is 50–329 ng/ml) (25). None of the 92 cord blood samples reported in the studies described below contained polyclonal IgE (sensitivity of the assay is >0.1 ng/ml). Second, DNA was extracted from lymphocytes of eight maternal-cord blood pairs, and microsatellite alleles of the autosomal D19S1600 locus were compared. Cord blood genomic DNA produced had only two alleles in all cases. One was identical to that of maternal DNA as template, and the other was discordant. The discordant allele was presumably paternal, although we did not attempt to obtain DNA from fathers. If admixture had occurred, cord blood would be expected to contain three alleles, assuming the parents were heterozygous for different alleles.
CBL cytokine responses to rMSP-119 and N-terminal P2 and P3 peptides
CBL from nine healthy North American newborns did not produce IFN-γ or IL-4 when incubated in culture medium alone or in media containing recombinant MSP-119 or LSA-1 peptides. CBL from 81 of 92 Kenyan newborns incubated in culture medium alone produced no IFN-γ or IL-4 by ELISPOT; 11 samples produced low levels of IFN-γ (2–4 cytokine-secreting cells/106 lymphocytes). A positive response to MSP-119 in the former group of 81 subjects was defined as >4 IFN-γ or IL-4 cytokine-secreting cells/106 lymphocytes. In the latter group, a positive response was defined as a net value for MSP-119-stimulated cultures that was at least 2-fold greater than constitutive production.
Twenty-six percent (24 of 92) of Kenyan newborns had CBL that produced IFN-γ or IL-4 in response to one or more MSP-119 variants (Table I⇓). The proportion of IFN-γ responses was greater than IL-4 responses (22 vs 8 of 92 subjects). The strength of the IFN-γ- and IL-4-positive responses ranged from 4 to 218 and 4 to 132 cytokine-secreting cells/106 lymphocytes, respectively. There were no differences in the proportion of newborns with CBL that responded to the E-KNG, Q-KNG, E-TSR, or Q-TSR variant of MSP-119. A sufficient number of lymphocytes to evaluate responses to all four variants were available from 19 newborns. There were no differences in the frequency of responses to each variant, and CBL from most subjects responded to two or more variants (data not shown). None of the 55 CBL preparations examined produced cytokine when incubated with LSA-1 84–107 or 1888–1909 peptides (Table I⇓).
Cord blood lymphocyte cytokine responses to MSP-119 and LSA-1 peptides
CBL were obtained from an additional group of 12 newborns. IFN-γ and IL-13 responses to MSP-1 P2 and P3 peptides were evaluated by ELISA, and the effects of CD4+ and CD8+ depletion and enrichment were evaluated (Table II⇓). A total of 5 of 12 subjects (42%) produced IFN-γ or IL-13. CD4 cell depletion studies were performed using CBL from two subjects. IFN-γ and IL-13 responses to MSP-1 peptides were completely eliminated following depletion of CD4+ cells from CBL of subject number 4. Whereas CBL from subject number 5 produced IL-13 when incubated with P2 or P3 peptide, no cytokine production was observed following depletion of CD4 cells. Enrichment for CD4 cells by positive selection was performed using CBL from subject numbers 6, 7, and 9. Two of the three enriched preparations produced IFN-γ, whereas there was no response when parallel aliquots of nonfractionated CBL were incubated with the MSP-1 peptides. Enrichment for CD8 and non-CD4/non-CD8 cells from subject number 9 had no effect on cytokine production. MSP-1 peptides failed to stimulate IFN-γ or IL-13 production by CBL from any of the 9 North American newborns examined.
Cord blood lymphocyte cytokine responses to MSP-1 P2 and P3 peptidesa
MSP-1-driven IL-10 production was measured to assess whether cells depleted by enrichment for CD4+ may suppress IFN-γ production in vitro. IL-10 was measured in three samples, numbers 5, 7, and 9 (see Table II⇑). Supernatants from unfractionated CBL of subject number 7 contained 193 and 168 pg/ml, and subject number 9 contained 224 and 291 pg/ml IL-10 following stimulation with MSP-1 P2 and P3 (net values after subtracting constitutive production), respectively. Parallel CD4 cell-enriched preparations contained no IL-10 (subject number 7) and 105 and 192 pg/ml IL-10 (subject number 9). No net MSP-1-induced IL-10 was detected in subject number 5. Therefore, MSP-1 stimulated IL-10 production by non-CD4+ cells in unfractionated cultures in which no MSP-1-driven IFN-γ was detected. This contrasted to MSP-1-induced IFN-γ production in parallel cultures enriched for CD4+ cells, supporting the possibility that non-CD4+ cells may suppress Ag-induced IFN-γ by IL-10.
MSP-119 variants in the study population
Nonsynonymous mutations in MSP-119 codons 1644, 1691, 1700, and 1701 have been described in several malaria endemic regions and P. falciparum clones (26, 27, 28, 29). We therefore assessed the MSP-119 variants that existed in our study population. PCR amplicons encoding MSP-119 were obtained from eight blood smear-positive subjects over a period of 2 mo (four from cord blood and four from maternal blood). Three to five clones were sequenced from each isolate. Eleven unique clones with nonsynonymous mutations in codons 1642 (K and R), 1699 (S and N), and 1716 (L and F) were identified. With respect to codons 1644, 1691, 1700, and 1701, only the Q-KNG variant was detected.
IgM, IgE, and IgG Abs to MSP-119 in cord and maternal blood
Five of eight-six (5.8%) cord blood samples contained IgM Ab to MSP-119 Q-KNG. The proportion of maternal plasma samples with IgM Abs to MSP-119 Q-KNG was greater (49 of 81 subjects, 60.5%; p < 0.001 compared with cord blood). The levels of IgM Ab in maternal and cord blood were not significantly correlated (r2 = 0.1, p = 0.6). IgE Ab to MSP-119 was not detected in any of the cord or maternal blood samples.
Unlike IgM and IgE, maternal IgG crosses the placenta in utero. Thus, when Abs of this Ig isotype are present in cord blood, it is not possible to distinguish whether they are of maternal or fetal origin. A total of 85% (61 of 81) of maternal samples and 78% (67 of 86) of cord blood samples contained IgG Abs to MSP-119 Q-KNG. The geometric mean IgG Ab level (±SEM) in maternal blood was higher than those of cord blood (43 ± 8 vs 30 ± 9 U/ml). There was a significant correlation between the level of Ab in maternal and cord blood (r2 = 0.413, p < 0.0001).
Ab levels to the other MSP-119 variants were generally lower than to the Q-KNG allele. The geometric mean IgG (±SEM) in maternal and cord blood samples, respectively, were as follows: E-KNG, 41 ± 9 and 25 ± 6; Q-TSR, 32 ± 8 and 24 ± 8; E-TSR, 19 ± 6 and 11 ± 5 U/ml.
Polyclonal IgG and MSP-119-specific Ab production by cord blood B cells
CBL IgG production in response to rMSP-119 was examined in 27 Kenyan and 6 North American newborns (Fig. 1⇓). Spontaneous IgG production (i.e., CBL incubated in culture medium alone) by the Kenyan samples was ∼4-fold higher than that of North Americans (geometric mean ± SEM = 4.1 ± 1.2 vs 1.1 ± 0.3 ng/ml, p < 0.001; Fig. 1⇓, A and B). The Q-KNG variant stimulated IgG production by 5 of 27 (18.5%) Kenyan CBL samples (Fig. 1⇓, C and D, paired lines with filled circles). Kenyan newborns with P. falciparum in cord blood by PCR were more likely to have B cells that responded to Q-KNG MSP-119 (Fig. 1⇓C) than those without detectable infection (Fig. 1⇓D, 4 of 8 vs 1 of 19, p = 0.02). The E-TSR variant of MSP-119 failed to stimulate IgG production in parallel cultures (Fig. 1⇓B). CBL from North American newborns did not respond to Q-KNG (Fig. 1⇓A) nor the other three MSP-119 variants (data not shown).
IgG production by cord blood B cells stimulated with rMSP-119. CBL (2 × 106) were incubated for 7–10 days with MSP-119, and total IgG in culture supernatants was measured as described in Materials and Methods. All preparations produced a net increase of >20 ng/ml IgG when incubated with a 1/200 dilution of pokeweed mitogen. •, A significant (p < 0.05) increase in total IgG production for MSP-119 Q-KNG-stimulated cells vs those incubated in culture medium alone. ○, There was not a statistically significant difference between MSP-119-stimulated and control wells. A, Presents results of experiments in which CBL from full-term healthy North American newborns were incubated with MSP-119 Q-KNG. B, Describes results of studies in which Kenyan CBL were incubated with MSP-119 E-TSR. C and D, Separate Kenyan newborns according to whether or not P. falciparum small subunit rDNA was detectable in cord blood. The asterisk denotes culture supernatants that contained IgG Abs to MSP-119 Q-KNG (6 and 14 U/ml). Note differences in the scale of the y-axis for IgG production by North American vs Kenyan CBL.
MSP-119-specific IgG Ab was measured in culture supernatants from the five CBL preparations that produced polyclonal IgG. Ab to MSP-119 Q-KNG was detected in the two supernatants that contained >200 ng polyclonal IgG (values of 6 and 14 U/ml; marked with asterisks in Fig. 1⇑C). IgG Ab to the other three variants of MSP-119 was not detected in these or the other three samples that produced polyclonal IgG.
Four of the five newborns with MSP-119-specific IgM Ab in plasma from cord blood also had MSP-119-reactive B cells. These subjects were positive for P. falciparum by PCR (Fig. 1⇑C).
Maternal T cell cytokine responses to MSP-119 and LSA-1
Maternal PBMC were obtained within 24 h of parturition, and cytokine production in response to MSP-119 and LSA-1 was evaluated. Criteria for a positive response were the same as those described for CBL. Thirty-six percent (29 of 81 subjects) had PBMC that produced IFN-γ or IL-4 when stimulated with MSP-119 (Table III⇓). The proportion with IFN-γ responses was greater than IL-4 responses (36% vs 9% for any variant of MSP-119). There were no differences in the percentage of responses to the E-KNG, Q-KNG, E-TSR, or Q-TSR variants. There was a sufficient number of PBMC from 17 women to evaluate responses to all four MSP-119 variants. Analysis of this subset indicated there were no differences in the frequency of responses, and that CBL from many of the subjects responded to two or more MSP-119 alleles (data not shown).
Maternal peripheral blood mononuclear cell cytokine production in response to recombinant MSP-119 and LSA-1 peptidesa
The strength of maternal IFN-γ responses tended to be stronger than CBL responses. For example, the geometric mean for maternal responses to all variants of MSP-119 was 57 IFN-γ-secreting cells/106 lymphocytes (Table III⇑) vs 15 for CBL (Table I⇑). There was neither a positive nor negative association between maternal and CBL cytokine responses.
A total of 5 of 37 (14%) maternal PBMC preparations made IFN-γ when incubated with LSA-1 84–107, and 13 of 35 (37%) responded to LSA-1 1888–1909. These data contrast those for CBL, which uniformly failed to respond to LSA-1 (Table I⇑). PBMC from seven North American adults who had never traveled to a malaria endemic area did not respond to any variant of MSP-119 or LSA-1.
The mitogens PMA + ionomycin stimulated more IL-5 and similar or lower levels of IFN-γ and IL-10 from Kenyan CBL compared with CBL obtained from North American newborns (data not shown). These results are similar to observations reported previously from the same population of Kenyan newborns (30).
Malaria infection status and MSP-119-driven cytokine responses
P. falciparum detected by PCR in cord blood correlated with CBL MSP-119-driven IFN-γ and/or IL-4 production (Table IV⇓, p < 0.0001). Similarly, infection of the mother correlated with the ability of her PBMC to respond to MSP-119.
Relationship between malaria infection status determined by PCR and T cell cytokine responses to MSP-119
Discussion
MSP-1 is a glycoprotein expressed by P. falciparum late-stage trophozoites and schizonts. The C-terminal MSP-119 fragment produced by a series of protease-mediated reactions most likely participates in the process of red cell invasion by merozoites (31, 32). Several lines of evidence suggest that MSP-1 and MSP-119 may be efficacious in a malaria vaccine. mAbs to MSP-1 have been found to inhibit in vitro growth of parasites (33, 34). Ab levels to MSP-119 correlate with age-related changes in the level of asexual parasitemia in some, but not all malaria endemic areas (35, 36, 37, 38, 39, 40, 41). Primate vaccine trials (42, 43) and studies of mouse malaria models indicate that immunization with MSP-119 (44, 45) protects against blood-stage infection. Given that children younger than 3–4 years old suffer the bulk of malaria morbidity and those born in areas in which malaria is endemic will ultimately be involved in clinical trials of malaria vaccines, there is great need for a better understanding of the mechanisms of naturally occurring immunity to MSP-1 in this age group. Results of this study show that some infants born in an area of coastal Kenya in which malaria transmission is stable are sensitized to MSP-1 in utero. This immunity is manifest by both T and B cell responses to MSP-119 and MSP-1 N-terminal peptides detectable at the time of birth. In contrast, there was no evidence of prenatal immunity to LSA-1, which is not expressed in infected red cells, but only in hepatocytes of individuals infected with P. falciparum sporozoites (46, 47). Neonatal T and B cell responses to MSP-1 were not due to admixture of fetal and maternal blood since: 1) concurrent maternal immune responses to MSP-1 and LSA-1 did not correlate with Ag-specific immunity in the newborn, and 2) comparison of polyclonal IgE levels and alleles of a polymorphic autosomal locus in paired samples of maternal and cord blood showed no evidence of mixture between the two circulations.
IFN-γ production was used as the primary indicator of prenatal T cell sensitization since this cytokine is secreted by cord blood T cells (48, 49). As judged by IFN-γ as well as IL-4 and IL-13 production, CBL from 26% (n = 92; Table I⇑) and 42% (n = 12; Table II⇑) of Kenyan newborns responded to rMSP-119 and the MSP-1 P2 or P3 peptides. These responses were not due to nonspecific reactivity with yeast-derived rMSP-119 since CBL from American newborns (or adults) failed to produce cytokine when incubated with the same constructs. Keitel and coworkers (50) reported that 30 of 39 malaria naive American adults had lymphocyte proliferation responses (stimulation index >3) to rMSP-119. It is not clear why this was not the case in the current study. It is possible that the two preparations of MSP-119 differed in purity or mixture with ill-defined fermentation products that have different effects on lymphocyte proliferation and cytokine production. The proportion of Kenyan newborns with cord blood T cell responses to MSP-1 reported in this work was similar to that of the only other comparable study of African newborns (12), which used p190 to evaluate recall responses by infants born in Gambia. This report indicated that ∼40% of 90 newborns had weak CBL proliferation responses (stimulation index >2) when incubated with p190N or p190L. Cytokine responses were not described in the Gambian study, and we did not examine lymphocyte proliferation. The overall frequency of responses by maternal PBMC in the present study (36%) was similar to that of healthy Gambian and Kenyan adults (12, 38, 51).
The role of CD4 and CD8 T cell subsets in cytokine responses to the MSP-1 P2 and P3 peptides was examined in 12 newborns (we were unable to obtain a sufficient amount of rMSP-119 to perform similar experiments with this preparation). Depletion of CD4, but not CD8 cells abrogated peptide-driven cytokine responses in two preparations, suggesting that former T cell subset is responsible for Ag-driven production of IFN-γ and IL-13 (Table II⇑). In two subjects, enrichment for CD4 cells demonstrated the presence of MSP-1-reactive lymphocytes that were not detected in cell cultures before enrichment (subject numbers 7 and 9). This result may be due to the fact that 2–3 times as many cytokine-producing CD4 T cells were present after enrichment. Alternatively, counterregulatory cytokines such as IL-10 produced by cells depleted by enrichment for CD4 cells may suppress Ag reactivity in vitro (52). Although detailed assessment of this problem will require examination of additional CBL preparations, more IL-10 was measured in unfractionated compared with CD4+-enriched cell culture supernatants. This suggests cross-regulation by IL-10 may suppress MSP-1-driven IFN-γ production by CD4+ cells in some samples.
T cell responses to MSP-1 are presumably involved in maintaining protection against blood-stage malaria by providing cytokine help to B cell for Ab production and Ig switching. The primary role of Ab in mediating protection is exemplified by the demonstration that Abs to the Plasmodium yoelii MSP-119 homologue confer protection against lethal blood-stage infection in mice, and mAbs to P. falciparum MSP-119 inhibit merozoite invasion in vitro (53, 54). Evidence for a role in anti-MSP-1 Ab in protection against human malaria under conditions of natural transmission is less clear. Some, but not all reports have demonstrated a statistically significant association between the level of anti-MSP-119 Abs and clinical morbidity in children, at least at the population level (36, 37, 38, 39, 55). There is also a general correlation between the level of passively acquired maternal IgG Abs at birth and the time to first malaria infection in some studies (56), but not in others (7). It is, however, difficult to extrapolate these results to a direct role for Ab to MSP-1 in mediating resistance against high density parasitemia or clinical morbidity for several reasons. The amount of maternal IgG Ab acquired by newborns most likely differs among infants within a population, and the rate of Ab decay in vivo cannot be directly measured. Maternal Abs are also directed against blood-stage Ags other than and in addition to MSP-1. These Abs may have specificities for different regions and alleles of a given Ag. Aside from the expected finding that nearly 78% of cord blood samples from newborns investigated in this study had IgG Abs to MSP-119, the current data indicate that 5 of 86 subjects also had IgM Abs. This Ig isotype originates from the fetus since maternal IgM is not transported across the placenta. This observation may be germane to a recent prospective study of infants born in a malaria holoendemic area of western Kenya (57). Several infants had a marked rise in anti-MSP-119 Ab following malaria infection detected within 3–6 mo of birth. Heterogeneity of responses in the study cohort was interpreted to be a consequence of functional immaturity to generate Abs during the neonatal period. An alternative explanation consistent with the existence of IgM Ab at the time of birth is that infants with the most robust Ab responses may have been primed to MSP-1 in utero.
In addition to evidence of neonatal B cell sensitization demonstrated by the presence of IgM Abs in cord blood, we also observed that CBL from 5 of 27 newborns produced IgG following stimulation with MSP-119. All of these subjects also had MSP-119-reactive T cells, as might be expected since Ig isotype switching by B cells requires T cell help. There was, however, no correlation between the presence of MSP-119-specific IFN-γ and/or IL-4-secreting CBL and in vitro IgG production or MSP-119-specific IgM (i.e., many newborns had MSP-119-reactive T-cells, but lacked evidence of B cell sensitization). This discordance may have several explanations. T cell priming may have occurred late in gestation so that insufficient time had elapsed for Ig isotype switching and/or expansion of B cells to occur. Alternatively, MSP-1-specific T cells generated in utero may not provide optimal B cell help in all cases. Some MSP-1-specific T cells in Kenyan mothers and their fetuses may have arisen in response to other organisms. Prior studies have isolated malaria-specific T cell clones from nonexposed donors that responded to environmental organisms (58). MSP-1-specific T cells described in this work produced primarily IFN-γ, and not IL-4 or IL-13, indicative of a predominantly Th1 phenotype. This cytokine profile is consistent with the absence of MSP-1-specific IgE in maternal or fetal plasma and low or absent IgG4 Ab (personal observations) reported in other studies (59). Contrary to theories that prenatal sensitization favors a Th2 cytokine bias (60, 61), this was not the case for MSP-1-driven cytokine profiles in the present study.
The availability of constructs corresponding to four variants in the two epidermal growth factor domains of MSP-119 enabled examination of allele-specific responses. The MSP-119 Q-KNG was the only variant detected in P. falciparum genomic DNA isolated from eight individuals in the current study. IgG Abs in mothers and their newborns recognized the Q-KNG as well as the E-KNG, Q-TSR, and E-TSR alleles, although Ab to the E-TSR variant was less than that observed to the other variants. This is consistent with the notion that Abs to MSP-119 bind to conformational epitopes conserved in the two epidermal growth factor motifs, i.e., binding is not constrained by changes in primary amino acid sequence that do not affect folding through disulfide bonds in cysteine residues (33, 39, 62, 63, 64, 65). Similarly, cytokine responses in mothers and their newborns were stimulated equally well by each of the four constructs. This finding suggests that the Th cell epitopes included in the recombinant Ags were processed properly and conserved among the four constructs. In contrast, cord blood IgM Ab and in vitro B cell IgG production were restricted to MSP-119 Q-KNG. This result and previous studies of African adults (66) indicate that Ig receptors for MSP-119 expressed by clonal populations of neonatal B cells are constrained by subtle alterations in Ag conformation. Moreover, the data suggest that B cell sensitization to the various alleles of MSP-119 is related to exposure to the variants represented in a given geographic region. In this context, it will be of interest in future studies to determine whether the temporal pattern of B cell responses to MSP-119 alleles during infancy parallels the production of Abs with inhibitory activity specific for parasites expressing the same or different alleles. The recent description of Plasmodium chabaudi blood-stage parasites in which the genomic sequence for MSP-119 can be replaced by specific alleles from P. falciparum will enable examination of the functional role of such Abs (67, 68).
In conclusion, these findings firmly establish that fetal exposure to malaria infections and/or their antigenic materials occurs during gestation. The ability of the fetus to develop an immune response to malarial Ags may affect the efficacy of blood-stage malaria-based vaccines administered during the perinatal period. Therefore, it will be important to more fully characterize the immune response of the fetus and young infant with natural malaria exposure.
Acknowledgments
We acknowledge and appreciate the help of the nurses at the Msambweni District Hospital and the cooperation of the subjects who participated in this study.
Footnotes
-
↵1 This work was supported by grants from the National Institutes of Health (AI33061, AI45446, and AI45473).
-
↵2 Address correspondence and reprint requests to Dr. Christopher L. King, Division of Geographic Medicine, Case Western Reserve University School of Medicine, Room W147, Harlan Wood Building, 10900 Euclid Avenue, Cleveland, OH 44106-4983. E-mail address: cxk21{at}po.cwru.edu
-
↵3 Abbreviations used in this paper: CBL, cord blood lymphocyte; LSA-1, liver-stage Ag-1; MSP-1, merozoite surface protein-1; MSP-119, 19-kDa C-terminal portion of MSP-1.
- Received August 31, 2001.
- Accepted October 25, 2001.
- Copyright © 2002 by The American Association of Immunologists