| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasmodium vivax Duffy binding protein II (PvDBPII) is an important protein in the P. vivax life cycle. The interaction between the PvDBPII and the human erythrocyte Duffy antigen receptor for chemokines (DARC) is required for blood stage P. vivax infections; alternative erythrocyte invasion pathways into human erythrocytes have not been identified.4–8 The PvDBPII is highly polymorphic within the region required for binding.9,10 We have observed differences in the binding of different recombinant PvDBPII variants to Duffy-positive erythrocytes as measured with a fluorescence-activated cell binding assay (Martin KK and others, unpublished data). This enhanced binding may increase parasite virulence by greater invasion into erythrocytes as reflected by increased parasitemia levels and/or greater frequency in the parasite population. In support of this hypothesis, we previously observed that Papua New Guinean children infected with P. vivax containing PvDBPII with a lysine at residue 447 were more likely to be blood smear positive and had higher parasitemia levels compared with children infected with P. vivax parasites with PvDBPII with a serine at residue 447.11 The present study examines the hypothesis that certain PvDBPII alleles and/or haplotypes, particularly those with enhanced binding in vitro, are more common and/or are associated with higher parasitemia levels in children.
To examine whether PvDBPII alleles influence P. vivax parasitemia, we conducted an in-depth examination of P. vivax infections in a cohort of 206 Papua New Guinean children 5–14 years of age (mean ± 9.4 years) who were followed every two weeks over a six-month study. We recently described the study design and epidemiology of malaria infection in the cohort in depth.12 At the beginning of the study, the children were treated to clear blood stage malaria infections. In this longitudinal cohort study, P. vivax parasitemia was assessed by several methods including blood smear, a semi-quantitative post–polymerase chain reaction (PCR) ligation assay, and a quantitative real-time PCR (QRT-PCR). Plasmodium vivax PCR-positive samples were then genotyped to determine the predominant PvDBPII allele.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study design. In May 2004, 206 children 4–14 years of age (mean = 9.4 years, 95% confidence interval [CI] = 9.2–9.6 years) attending the Mugil/Megiar community schools were enrolled in the study after informed parental consent was obtained. A baseline peripheral 10 mL of venous blood was collected into EDTA-Vacutainer® tubes (Becton Dickinson, Franklin Lakes, NJ). Blood smears from all children were obtained and a short physical examination was conducted. All children irrespective of blood smear positivity were then treated with a seven-day course of artesunate according to Papua New Guinea national treatment guidelines (i.e., 4 mg/ kg on day 1, 2 mg/kg at days 2–7; 5 treatments were directly observed) for the eradication of all malaria blood stage infections. Artesunate does not clear liver stage Plasmodium infections. Two weeks after the start of treatment, all children were examined for treatment success with a finger prick of 250-µL blood collected into an EDTA-Microtainer® tube (Becton Dickinson) for blood smear and PCR analysis. Every two weeks thereafter for six months, enrolled children were followed-up at the community schools or in their village (when possible) if they were not present in school that day. Blood smear readings and a 250-µL blood sample were obtained at each follow-up and a short physical examination was conducted. Samples were also collected through the community health center or at school when a child presented with symptoms of malaria illness. Only five clinical cases of infection with P. vivax were observed in the study on the basis of the following criteria: P. vivax positive with a parasite density greater than 1,000 parasites/µL and blood smear–negative for P. falciparum, P. malarie, and P. ovale. Further details of field procedures and levels of infection and disease with all four types of malarial infections have been described elsewhere.12 Study protocols and consent forms were reviewed and approved by the Veteran’s Affairs Research Service and the Papua New Guinea Medical Research Advisory Committee institutional review boards.
Diagnostic testing and quantification of Plasmodium infections. Plasmodium infections were identified by two methods: blood smear and a post-PCR oligonucleotide ligation assay (OLA). For each of the study participants, a blood smear and a 250-µL blood sample were obtained at each follow-up. Thick and thin blood smears were stained with 4% Giemsa for 20 minutes and examined under microscopy at 100x magnification with an oil immersion lens. The number of parasites per 200 leukocytes was counted for each sample in two independent examinations by trained microscopists. All slides showing discrepant results for the two readings (different malaria species or densities observed) were read a third time to verify the results. Parasitemia for each sample was determined from a geometric mean of two or more non-discrepant readings and converted to parasites per microliter of blood using a standard leukocyte count of 8,000/µL of whole blood.12 Parasitemia levels were normalized by log transformation of n + 1. Thus, parasitemia used in this study is defined as the geometric mean of the parasites/µL + 1.
The post-PCR OLA was based on amplification of the small subunit ribosomal RNA gene as described and was used for Plasmodium species diagnosis.12–15 The infection intensity of P. vivax was quantified using a QRT-PCR as described.16 Only P. vivax-positive samples identified by OLA were quantified using the real-time PCR. All other samples were assumed to have a P. vivax parasitemia of zero.
Plasmodium vivax Duffy binding protein genotyping. The polymorphic PvdbpII gene was used as a marker for P. vivax strains. PvdbpII was genotyped at 14 polymorphic residues using a PvdpbII post-PCR OLA as described.17 Haplotypes for single P. vivax infections were constructed on the basis of the genotyping results for the 14 single nucleotide polymorphisms (SNPs) at 13 amino acids; 8 of the 27 haplotypes were observed previously in the same population. On the basis of the observed haplotypes for single infections, haplotypes for mixed P. vivax infections (more than two PvDBPII variants were found) were inferred using the predominant allele at each SNP as described.17 Novel haplotypes not observed in single infections were listed as unclassified (5.3%). A subset of samples (n = 76) was further sequenced to verify genotyping results. Phylogenic analysis was used to investigate the relationship between the observed PvdbpII genotypes. The phylogenetic tree was estimated using the neighbor-joining method and maximum parsimony as implemented in the MEGA 3.1 program.18,19 The reliability of the neighbor- joining tree was assessed by 1,000 bootstrap pseudo-replications.
Statistical analysis. All statistical analysis was performed in Statistical Analysis System version 9.1 (SAS Inc., Cary, NC). For parasitemia outcomes that incorporated repeated measures on the same individual, a linear mixed model was used in SAS 9.1 (PROC MIXED). The model was based on a restricted maximum likelihood model that estimates the applied maximum likelihood estimation techniques to the likelihood function associated with a set of error contrasts rather than to the original observations as described.20 The repeated option of PROC MIXED was used to specify the covariance structure of the error term to incorporate repeated measures on the same individual. PROC MIXED allowed us to incorporate subjects with incomplete data into the analysis. The compound symmetry covariance structure was used as this structure gave the best fit based on Akaike’s Information Criteria, the Corrected Akaike’s Information Criteria, and the Bayesian Information Criteria as provided in the SAS 9.1 output.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sequence analysis of these same 76 samples identified three additional non-synonymous SNPs in more than one sample that have not been previously identified: T404R (2 of 76 samples [2.6%]), K411I (2 of 76 samples [2.6%]), and I430T (5 of 76 samples [6.6%]).11,21–23 We also observed another synonymous mutation at residue R378R (C to A transversion at the third position of the codon); we have also observed a C to T transition in this same position in this population.11 In addition, three other SNPs were observed with each SNP likely representing PCR errors.24
Diversity of P. vivax in the population on the basis of PvDBPII.
To assess the diversity and relative frequency of PvDBPII variants in the population, 751 (95%) of the P. vivax PCR and/or blood smear–positive samples (n = 788) were amplified using the PvdbpII-specific OLA assay and genotyped. A total of 712 samples were assigned to 1 of at least 27 different PvDBPII haplotypes and only the dominant haplotype was assessed in mixed infections (Table 1
). There were 39 (5.3%) samples for which we could not reliably ascribe a haplotype. The PvDBPII haplotypes clustered into three main groups on the basis of phylogenic analysis using both neighbor-joining and maximum parsimony methods that gave similar trees. Two haplotypes did not fit into the same families with both methods: PNG-26 and PNG-27. The condensed neighbor-joining tree is shown in Figure 1. At least eight of these haplotypes were present in this same population four years prior to this study (PNG-AH, PNG-C, PNG-U, PNG-P, PNG-T, PNG-O, PNG-E, and Sal-1), which indicated that polymorphisms at these residues are stable over time.11
|
|
). These same haplotypes were most prevalent in the children at each of the two-week follow-up time points with remarkably consistent frequency throughout the study (Figure 2).The mean number of PvDBPII haplotypes observed in each child over the course of the study was 2.6 (95% CI = 2.3–2.8, range = 0–7 genotypes).
|
).
|
). This mean, however, was not adjusted to account for differences in age, multiple strain P. vivax infections, differing number of follow-up observations for each child, or the increased P. vivax positivity observed in some of the children. Using a mixed linear model that adjusted for these confounding factors, we found no association between PvDBPII haplotype and P. vivax parasitemia as measured by blood smears (F = 1.28, P = 0.16) or QRT-PCR (F = 0.64, P = 0.92).
Because we did not observe a single PvDBPII haplotype that affected P. vivax parasitemia, we clustered the haplotypes into three groups as defined by the main branches of the neighbor-joining phylogenic tree (Figure 1). As shown in Table 2, parasitemia in the Sal-I group tended to be higher than the other two groups but this difference did not reach statistical significance for parasitemia as determined by blood smears (F = 2.01, P = 0.1381) or for parasitemia determined by QRT-PCR (F = 0.54, P = 0.5835).
|
We recently observed that a recombinant fragment including subdomain 3 of PvDBPII binds better to DARC-positive erythrocytes than the full PvDBPII, which indicated that this portion of the molecule contains the minimum essential binding motif (Martin KK and others, unpublished data). In addition, we observed that recombinant proteins for the corresponding PNG-AH and PNG-C haplotypes bind more DARC-positive erythrocytes than other haplotypes (Martin KK and others, unpublished data). We examined whether haplotypes of subdomain 3 (amino acids 406–528) are associated with higher parasitemias. All haplotypes that contained the same subdomain 3 were grouped together for analysis (Table 3). We found a wide range of parasitemia levels associated with different subdomain 3 haplotypes based on blood smears (3.2–92.0 parasites/µL; F = 2.57, P = 0.0083). There was no association with haplotypes and parasitemia levels, however, based on QRT-PCR (F = 0.90, P = 0.5250) (Table 3).
|
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). These observations are consistent with previous studies using PCR-based oligonucleotide probes for allele determination and recent findings using a different high-throughput genotyping method for the P. falciparum merozoite surface protein (MSP-1).32–35 At least eight of the PvDBPII haplotypes (PNG-AH, PNG-O, PNG-P, PNG-T, PNG-E, PNG-C, PNG-U, and Sal-1) were observed four years earlier in the same population.11 We also identified 19 additional haplotypes in samples with single P. vivax infections; an additional 39 samples (5.2%) could not be inferred based on the PvDBPII-specific OLA. These unclassified PvDBPII haplotypes may represent newly emerging or rare alleles that are at a low frequency in the population. A few of these unclassified samples had more that one PvDBPII haplotype with similar concentrations, which made their definitive haplotype assignment impossible. This study represents the first large-scale genotyping of the PvDBPII using the OLA. Although 98.4% of the 14 SNPs examined by sequencing matched the OLA results, haplotype reconstruction proved challenging because we were examining 14 SNPs, many of the infections contained multiple P. vivax strains, and the OLA reaction at residue 503 was problematic. The error rate we observed (0.14), however, is similar to other error rates for haplotype reconstruction.36 We may be able to improve our haplotype reconstruction of the most abundant clone in multiple strain P. vivax infections by using an algorithm for reconstructing haplotyes such as PHASE.37
The PvDBPII OLA reactions for the 14 SNPs were chosen because these SNPs were present in the population with a frequency greater than 5% in 2000.11,17,24 Sequence analysis of 76 samples showed one other novel SNP at residue I430T that occurred at a frequency of 6.6% that should be included for future genotyping studies.
There are several potential explanations that could account for the dominant PvDBPII haplotypes that were observed in this population. First, certain haplotypes may bind better to the DARC, which may enhance parasite invasion and thus fitness as reflected by higher peripheral blood parasitemias. We found that recombinant PvDBPII corresponding to PNG-AH, the most common haplotype, bound more erythrocytes than recombinant protein corresponding to other variants (Martin KK and others, unpublished data). However, recombinant protein corresponding to the PNG-O and PNG-P variants, the next most frequent variants in the population, did not bind better to erythrocytes compared with less common variants (PNG-T and Sal-I). The relationship, therefore, of in vitro binding with haplotype frequency in the population is complex. Also, there was no detectable relationship between levels of parasitemia and PvDBPII haplotype, which rejected our initial hypothesis that certain haplotypes may invade more successfully than others. The measurement of parasitemia levels, however, is an imprecise measure of parasite fitness. Factors such as varying levels of protective antibodies could affect parasitemia levels in this population of children with significant protection against clinical disease; only five children had clearly confirmed clinical disease caused by P. vivax malaria.12 It is still possible that slight differences exist in parasite invasion efficiency that we were unable to detect. Ultimately, the isolation of parasites from different strains and evaluation of their ability to invade and replicate in vitro might better substantiate this hypothesis. Such experiments are now feasible because of greater success with short-term culture of P. vivax.38–40
A second explanation to account for dominant haplotypes is that certain PvDBPII haplotypes may be less immunogenic, thereby escaping the host immune response, which results in a greater frequency in the population. Although less immunogenic PvDBPII haplotypes may not result in an increased frequency in those children with little immunity because they are more susceptible to all strains, a less immunogenic haplotype may be able to escape immune detection in those individuals with stronger immunity towards P. vivax, which leads to higher levels of this haplotype in the human population. Overall, this seems less likely because plasma antibodies from these 206 children prior to treatment recognized recombinant PvDBPII corresponding to the dominant PNG-AH variant just as well and if not better than recombinant proteins corresponding to less abundant haplotypes, e.g., PNG-T, PNG-O, PNG-P, and Sal-I. To better understand the relationship of haplotypes with immunity, we need to examine their distribution in younger, less immune children relative to older children. We appreciate, however, that plasma recognizing recombinant proteins may not reflect the development of functional antibody responses in vivo. Thus, we are currently investigating functional antibody assays. Additionally, if PvDBPII were under strong immune selection with escape mutants, we might expect to see a more dynamic change in PvDBPII haplotypes in parasite populations. Certain variants may be eliminated with immune pressure and others could emerge that may not have been previously recognized by the host. This has been observed with P. falciparum erythrocyte membrane protein-1.41 Instead, we observed a fairly consistent proportion of the different variants over the period of observation. Six months, however, may be an insufficient period of time to observe such changes. Of note, more than one-third of the children lacked any detectable antibody response to any of PvDBPII variants tested, even though these children appeared to have developed a high degree of immunity to clinical P. vivax malaria, as indicated by the almost complete absence of clinical P. vivax disease.12 This suggests that the PvDBPII is not highly immunogenic and significant naturally acquired immunity conferred by antibodies to DB-PII may develop in only some individuals.
Alternatively, the haplotype frequencies may simply be stochastic. Parasites with certain PvDBPII haplotypes may have been introduced into this population by chance and persist by relapsing and enhanced transmission with occasional escape mutants because of relatively weak immune selection pressure. This possibility would be supported by showing spatial heterogeneity of the PvDBPII variants in the study populations in Papua New Guinea and by measuring the relative abundance of these dominant PvDBPII haplotypes from archival samples from the same populations in previous years.
In the present study, we observed that children infected with two or more strains of P. vivax had parasitemia levels at least two times higher than those children with only one detectable strain of P. vivax. Previously, we have shown that the prevalence of P. vivax infections, the intensity of infection (as measured by parasitemia), and the percent of infections with more than one strain of vivax by both PvDBPII and P. vivax merozoite surface protein 3
(PvMSP3) peaks in children 5–9 years of age in this community and begins to decrease in children more than 9 years of age.11 The same trends were observed in the current study. An explanation for this observation is that as children acquire immunity to P. vivax malaria, they more readily clear blood stage infections and thus are less likely to acquire multiple parasite strains detectable at any particular time. Measuring PvDBPII haplotyes will underestimate the complexity of parasite strains in an individual. In a subset of children, 5–14 different P. vivax micro-satellite markers were measured, which showed a higher complexity of infection than we were able to detect with the PvDBPII-specific OLA genotyping (King CL, Karunaweera, unpublished data).42 PvMSP3 restriction fragment length polymorphism analysis performed on samples collected from this same population four years earlier also support the observation that the complexity of P. vivax infections is higher than we are able to detect with the OLA genotyping.24
In conclusion, we observed many PvDBPII haplotypes present in this population with three dominant haplotypes. It is unlikely that certain haplotypes increased efficiency of blood stage infection as measured by increased peripheral blood parasitemia levels. Studies that evaluate the association of the different haplotypes with cellular and humoral immunity to PvDBPII and their spatial and temporal pattern of distribution in humans and mosquitoes will provide further insights to explain the distribution of PvDBPII variants in parasite populations. If certain haplotypes are predominant across different age groups and locations, this would simplify development of a PvDBPII subunit vaccine.
Received March 23, 2007. Accepted for publication June 22, 2007.
Acknowledgments We thank the study participants for their time; the teachers at the Mugil and Megiar Schools for their participation; the nurses of the New Guinea Institute for Medical Research (IMR) at the Mugil Health center for assistance; the IMR staff for collecting samples during the follow-ups and preparing microscopic sections for analysis; Daniel Tisch for statistical advice; and Arlene Dent for helpful comments on the manuscript.
Financial support: This study was supported by a grant from the Veteran’s Affairs Research Service and funds from the Papua New Guinean Government. Jennifer L. Cole-Tobian was supported by the ID/GeoMed Training Grant (T32-AI-07024) during this study.
* Address correspondence to Jennifer L. Cole-Tobian, or Christopher L. King, Center for Global Health and Disease, Case Western Reserve University, Wolstein Research Building 4-101, 10900 Euclid Avenue, Cleveland, OH 44106. E-mail: jtobian{at}gmail.com or cxk21{at}case.edu 
Authors’ addresses: Jennifer L. Cole-Tobian, Center for Global Health and Disease, Case Western Reserve University, Wolstein Research Building 4-101, 10900 Euclid Avenue, Cleveland, OH 44106. Pascal Michon, Elijah Dabod, and Ivo Mueller, Papua New Guinea Institute of Medical Research, PO Box 378, Madang 511, Papua New Guinea. Christopher L. King, Center for Global Health and Disease, Case Western Reserve University, Wolstein Research Building 4-132 10900 Euclid Avenue, Cleveland, OH 44106 and Louis Stokes Veterans Affairs Medical Center, 10701 East Boulevard, Cleveland, OH 44106.
Reprint requests: Christopher L. King, Center for Global Health and Disease, Case Western Reserve University, Wolstein Research Building 4-132, 10900 Euclid Avenue, Cleveland, OH 44106 and Research Department, Veterans Affairs Medical Center, 10701 East Boulevard, Cleveland, OH 44106, Telephone: 216-368-4817, Fax: 216-368-4825, E-mail: christopher.king{at}case.edu.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. L. Cole-Tobian, P. Michon, M. Biasor, J. S. Richards, J. G. Beeson, I. Mueller, and C. L. King Strain-Specific Duffy Binding Protein Antibodies Correlate with Protection against Infection with Homologous Compared to Heterologous Plasmodium vivax Strains in Papua New Guinean Children Infect. Immun., September 1, 2009; 77(9): 4009 - 4017. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. L. King, P. Michon, A. R. Shakri, A. Marcotty, D. Stanisic, P. A. Zimmerman, J. L. Cole-Tobian, I. Mueller, and C. E. Chitnis Naturally acquired Duffy-binding protein-specific binding inhibitory antibodies confer protection from blood-stage Plasmodium vivax infection PNAS, June 17, 2008; 105(24): 8363 - 8368. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
