Genetic polymorphisms associated with anti-malarial antibody levels in a low and unstable malaria transmission area in southern Sri Lanka

Ethical clearance for this study was granted by the Ethics Review Committee, Faculty of Medicine, University of Colombo. One thousand and eleven individuals over 14 years of age and who gave written consent to participate in the study were recruited to the study. Proxy consent was obtained for the young participants (aged 14–18 years) from their parents or the guardian/s.

This study was conducted in eight adjacent villages in Kataragama Medical Officer of Health (MOH) division in the district of Moneragala [1]. Kataragama is an area in the dry lowland coastal plains of south-east Sri Lanka, where the malaria situation is considered “unstable” and low within the last decade.

The number of houses in each village and number of individuals in each house were listed, based on previous census records maintained by the field research facility at the Malaria Research Station, Kataragama [1]. Each house and each individual living in that house was given a unique number for identification. The study subjects were visited during four consecutive visits to the area between December 2006 and May 2007 in order to collect the relevant data and blood samples for DNA extraction, sera and thin/thick blood smears.

Five mL of blood was collected from all study subjects to normal tubes for serum samples and for EDTA–coated tubes for DNA extraction. Thin and thick blood smears were prepared for examination for the presence of parasites in the blood. Each tube and the corresponding slides were labeled according to the serial number given for each individual. Data on age, sex, history of previous clinical malaria during the past 10 years and information on the use of bed nets were recorded.

Serum was separated from the clotted blood samples by centrifugation (12,000 rpm for eight minutes) and analysed for six anti-malarial antibodies (ie, anti-AMA1, anti-MSP1, anti-MSP2, anti-NANP, for P. falciparum, anti-AMA1 and anti-MSP1 for Plasmodium vivax) and total IgE with the levels determined using standard ELISA technique. Fifty μL of each antigen at a dilution of 0.5 μg/mL (for AMA1 (3D7), MSP2 (3D7) and IgE) or 1 μg/mL (for MSP1 (Wellcome genotype) and NANP [(NANP)4] were coated on ELISA plates (Immulon4 ELISA plates/Dynatech) and were incubated at 4°C overnight. The plates were washed with a solution of PBS and 0.05% Tween20 (PBS/T) and 200 μL of blocking solution (2% skimmed milk in PBS/T) were added before incubating the plates for three hours in ambient temperature. Serum samples were added in duplicates after washing the incubated plates with PBS/T. These were incubated overnight at 4°C. Plates were washed in PBS/T and 50 μL of horseradish peroxidase – conjugated rabbit anti-human IgG (DAKO) diluted 1/5,000 in PBS/T were added and incubated for three hours at room temperature. The plates were again washed with PBS/T and OPD substrate solution (100 μL/well) was added and left at room temperature for 10–15 minutes for the assay to develop. 2 M H₂SO₄ (25 μL/well) was added to stop the reaction and the plates were read at 492 nm in an ELISA reader ( Additional file 1 OD values). The cut-off value of the assay was determined by calculating the arithmetic mean of the absorbance of negative control samples obtained from European individuals who had never been exposed to malaria and adding three standard deviations to that value (mean OD + 3SD). Using standard positive and negative controls, positive–negative threshold baseline was constructed using OD values obtained upon ELISA and was used for calculation of the observed antibody titres as described by Snyder et al. in 1983 [9].

DNA was extracted from whole blood (2.5 mL) collected in to EDTA tubes using Nucleon BACC2 commercial DNA extraction kit [Gen-Probe Life Sciences, Tepnel Research Products & Services, Manchester, UK]. Five ng of gDNA was whole-genome amplified by primer-extension pre-amplification (PEP) using N15 primers (Sigma, UK) and Biotaq (Bioline, UK) polymerase as previously described by Zhang et al.[10]. Single nucleotide polymorphisms (SNPs) were assayed on the Sequenom® iPLEX platform according to manufacturer’s instruction using diluted PEP DNA (1:10). Genotype calls were made using the Sequenom® Typer v4.03 software [11].

Blood samples were screened for malaria infection by microscopic examination of stained thin and thick blood smears by trained microscopists.

Antibody levels were analysed in relation to age, gender and the history of malaria. For the genetic analysis a total of 170 SNPs in 62 genes were genotyped ( Additional File 2 SNPs). This included 65 SNPs with known associations with malaria infection/disease severity. SNPs were filtered according to the following criteria: SNPs with less than 95% genotype call, not in Hardy-Weinberg Equilibrium (p < 0.05), or monomorphic were excluded from further analysis. Individuals who had less than 95% of the SNPs genotyped were also excluded from further analysis. After quality control of SNPs and samples, 118 SNPs in 1,008 individuals were selected for further analysis.

Data were analysed using SPSS V 13.0, R software package for statistical analysis (2.6.1) and MINITAB 14.0. Kruskall–Wallis test and Mann–Whitney U tests were used to compare and study associations and Fisher’s Exact and Chi-squared tests were also used for genotype analysis. Significance of SNPs was confirmed by generation of q values using the software QVALUE (V1.0) [12]. Linkage disequilibrium (LD) plots for the 19 genes where the 23 significant SNPs were located were generated using Haploview (V4.2).

Ages ranged from 14 to 89 years with approximately equal numbers of males and females (514:497). Over 99% of the population belonged to one ethnic group (Sinhala) with only seven belonging to Tamil ethnicity. The majority of individuals (>95%) used bed nets. Only 18.4% of individuals (186/1,011) have had clinical malaria within the past 10 years and none within the past five years (Table 2). At the time of sample collection, blood smear examinations were negative for malaria parasites in all subjects.

Greater than 80% and >97% of participants were seropositive for P. falciparum and P. vivax respectively (Table 2). The study population was divided into three groups based on the participants’ past history of clinical malaria within the past 10 years (Table 2). Anti-AMA1 (Pf) (p = 0.004), anti-MSP2 (Pf) (p = 0.027), anti-NANP (Pf) (p = 0.002) and anti-MSP1 (Pv) (p = 0.003) levels were significantly higher in people who have had one or more clinical malaria episodes within the past 10 years, when compared to those who gave no evidence of clinical malaria during past 10 years (Table 3).

Total study population was categorized into five age-groups based on their ages at the time of collection of serum (Table 2). The three age groups comprised 14–29, 30–44, and 45–59 years old and were almost equally represented (37.6%; 30.4%, 23.1% respectively) while a relatively lower proportion of subjects represented the >60-year-old age group. A high level of anti-MSP1 (Pf) (p = 0.001) and of anti-AMA1 (Pv) (p < 0.001) was observed in the 45–59 year old age group (when compared to other age groups) (Table 4).

The levels of all six anti-malarial antibodies were comparable between males and females (Table 3). However, the total IgE concentration between males and females were seen to be significantly different (p < 0.001) with the median concentration of total IgE in males (5,719.66 ng/mL) being significantly higher when compared to females (3,298.72 ng/mL).

Antibody levels were classified into a low-high binary trait using a cut-off of 5,000 IU/mL by ploting a histogram of antibody levels transformed in to logarithm base 10. The threshold was set at 10 for the histogram, as there was a clear separation between two peaks at this point.

The residents of Akkarawissa consistently showed overall higher antibody levels compared to the residents in all other villages (Figure 1A-1B) for not only all the anti-malarial antibodies but IgE as well. These high overall levels were contributed by >60% of the village particpants who had high antibody levels (>5,000 IU/ml), in contrast to <10% of individuals from the other villages.

Genetic data was analysed against the categorical classification of antibody data using a cut-off of 5,000 IU/ml to define the low and high antibody groups. Significant association was found with four SNPs and anti-AMA1(Pf); two SNPs and anti-MSP1(Pf) (none for anti-MSP2(Pf)); eight SNPs and anti-NANP(Pf); two SNPs and anti-AMA1(Pv); six SNPs and anti-MSP1(Pv); and seven SNPs and IgE (Table 5). Eleven SNPs out of these 28 were located in genes that code for interleukins, ie, IL4, IL10, IL13 and IL22. Most of these markers within the interleukin genes were seen to be in association with high levels of P. falciparum anti-malarial antibodies and five out of seven host genetic markers, which were significantly associated with high IgE levels, were located in IL4 and IL10 genes (with the other two being located in GBP7 and DERL3 genes).

The only interleukin-associated host genetic marker that was significantly associated with high levels of P. vivax antibodies was rs848 for anti-MSP1 (Pv), which is a coding SNP in IL13. The SNP coding for the sickle-cell variant (HbS or rs334), which is situated in HBB was significantly associated with high antibody levels against both AMA1 and NANP of P. falciparum antigens. It was interesting to note that the majority of SNPs were segregated on a few selected chromosomes with eight and 11 SNPs out of 28 found within genes located on chromosome 1 and chromosome 5 respectively. However, taken together, none of the SNPs had significant association with elevated levels of all tested antibodies (Chi-Squared test, Regression analysis, p < 0.05). Nine SNPs, which had significant associations with elevated antibody levels were found to be synonymous in their function (Table 4), which included all significant SNPs located in interleukin genes. However, some of the studied SNPs, e.g. those located in FCGR2a, SLC22A4, TLR4, ICAM1 and HBS genes, were non-synonymous in function (Table 5). Allele frequencies in two SNPs (rs20541 and rs7935564) in two genes, ie, IL13 and TRIM5 were found to be significantly different (p = 0.038) between those with past history of malaria (one or more clinical malaria attacks during last 10 years, i.e. malaria-susceptible individuals) and people without past history, i.e. apparently protected from malaria (Chi-Squared test p = 0.036).

Significance of allele frequencies of SNPs was further confirmed by the calculation of q values for each p value using the software QVALUE (QVALUE Version 1.0; Alan Dabney and John D Storey; pi0 =1, FDR level 0.05).

Two SNPs (rs1800872 and rs1800871), which were significant for elevated levels of anti-NANP (Pf) and total IgE, were in linkage disequilibrium (D’ = 1.00, LOD = 334.61, r² = 1.00)(Figure 2a-2c). These two SNPs were also in LD with rs1518110 (D’ = 1.00, LOD = 278.77, r² = 0.87) which was significant for raised levels of both anti-NANP(Pf) and total IgE. In chromosome 5, high linkage disequilibrium could be seen between SNPs rs1801033 and rs1881457 (D’ = 0.858, LOD = 164.96, r² = 0.646), rs18811457 and rs2243250 (D’ = 0.93, LOD = 206.85, r² = 0.754), rs25882 and rs2706348 (D’ = 0.86, LOD = 146.5, r² = 0.667). Similarly high LD could also be seen between rs1801033 and rs22443250 (D’ = 0.689, LOD = 107.26, r² = 0.472), rs2706348 and rs272867 (D’ = 0.662, LOD = 54.79, r² = 0.279).