The authors have declared that no competing interests exist.
Conceived and designed the experiments: HI KD. Performed the experiments: HI FM-N. Analyzed the data: FF HI. Contributed reagents/materials/analysis tools: PD AL SH KD. Wrote the paper: HI KD FF.
Current address: Nottingham Trent University, Nottingham, United Kingdom
Current address: Centre for Population Health, Burnet Institute, Melbourne, Australia
Haptoglobin related protein (Hpr) is a key component of trypanosome lytic factors (TLF), a subset of high-density lipoproteins (HDL) that form the first line of human defence against African trypanosomes. Hpr, like haptoglobin (Hp) can bind to hemoglobin (Hb) and it is the Hpr-Hb complexes which bind to these parasites allowing uptake of TLF. This unique form of innate immunity is primate-specific. To date, there have been no population studies of plasma levels of Hpr, particularly in relation to hemolysis and a high prevalence of ahaptoglobinemia as found in malaria endemic areas.
We developed a specific enzyme-linked immunosorbent assay to measure levels of plasma Hpr in Gabonese children sampled during a period of seasonal malaria transmission when acute phase responses (APR), malaria infection and associated hemolysis were prevalent. Median Hpr concentration was 0.28 mg/ml (range 0.03–1.1). This was 5-fold higher than that found in Caucasian children (0.049 mg/ml, range 0.002–0.26) with no evidence of an APR. A general linear model was used to investigate associations between Hpr levels, host polymorphisms, parasitological factors and the acute phase proteins, Hp, C-reactive protein (CRP) and albumin. Levels of Hpr were associated with Hp genotype, decreased with age and were higher in females. Hpr concentration was strongly correlated with that of Hp, but not CRP.
Individual variation in Hpr levels was related to Hp level, Hp genotype, demographics, malaria status and the APR. The strong correlations between plasma levels of Hp and Hpr suggest that they are regulated by similar mechanisms. These population-based observations indicate that a more dynamic view of the relative roles of Hpr and Hpr-Hb complexes needs to be considered in understanding innate immunity to African trypanosomes and possibly other pathogens including the newly discovered
Haptoglobin-related protein (Hpr) is found within trypanosome lytic factor (TLF), the component of human plasma (and that of some old world primates) which kills
Hpr is found predominantly as a 45 kD heterodimer consisting of one α- and one β-chain covalently bound together to form (αβ)-dimers
Hpr and Hp are primarily expressed in the liver, however, Hpr transcript levels were determined to be only 6% of those of Hp due to the presence of a retroviral-like element in intron 1
Hpr levels have not been studied at a population level and the association between Hpr, Hp and the APR is unknown, particularly in relation to malaria-induced hemolysis. Consequently, we developed a specific ELISA that distinguishes between Hp and Hpr. Hpr levels were measured, together with Hp and the acute phase proteins, C-reactive protein (CRP) and albumin, in a large cohort of children living in a malaria-endemic area of Gabon. We investigated the association of Hpr levels in relation to these acute phase proteins as well as Hp genotype, demographics and malariometric indices. As a comparison, levels of Hpr were also measured in healthy children from the United Kingdom who do not experience such a chronic burden of infectious disease and were not exposed to malaria.
Details of the Gabon study design, population and laboratory methods have been published previously
Ethical clearance was obtained from the “Comité d’Ethique Pour La Recherche En Médecine Humaine” as well as the Ministry of Public Health and the Governor of the Province. After consultation with the villages’ authorities regarding an acceptable procedure, a village-wide information meeting was held. In Bakoumba, a public convocation was addressed to all inhabitants where those present included medical authorities (Director of the Hospital, Head of the medical laboratory), local authorities (Sous-Préfet, school headmaster, chief of SODEPAL: Société D’Exploitation du Parc de la Lékédi and traditional leaders) as well as leaders of the project at CIRMF (Centre International de Recherches Médicales de Franceville). The research project was presented orally and discussed. Additional information was provided through the director of SODEPAL (Jean Bourgeais) and the medical staff of the hospital. In Dienga, information was provided orally by the staff of the health center (a nurse and a laboratory worker recruited among villagers). People understood that samples would be used for research on malaria, in relation to red blood cell polymorphisms. The verbal consent of parents/guardians was sought and was given, after allowing time for appropriate consultation, via the representatives of parent associations. Parents that were unwilling for their children to participate were identified and their children subsequently excluded, without prejudice, from study surveys. The Comité d’Ethique Pour La Recherche En Médecine Humaine was informed that an individual verbal consent was obtained and agreed to this, literacy levels being low. Ethical approval was also granted by the Institutional Review Board of the New York University School of Medicine.
Samples taken from 3–4 year old Caucasian children allowed comparison of Hpr levels in two different populations. These serum samples were obtained from stored serum previously collected from healthy children as part of studies conducted by the Oxford Vaccine Group. The use of such samples for further studies was approved by the Oxford Clinical Research Ethics Committee as C02.013 ‘Use of stored serum samples held by the Oxford Vaccine Group to further vaccine related research’.
Age and sex were recorded for each child. Parasite densities of
Hpr was detected using a capture enzyme-linked immunosorbent assay (ELISA) which was developed using a specific anti-Hpr monoclonal antibody and polyclonal anti-Hp which cross-reacts with Hpr
Initially various concentrations of polyclonal anti-Hp were used as capture antibody, followed by varying concentrations of pooled plasma (containing Hp and Hpr) (neat to 1 in 1,000,000 dilutions in 3%BSA/PBS) or purified Hp1-1 (Sigma) (100 mg/ml to 0.1 ng/ml), then anti-Hpr (diluted to 10 µg/ml PBS) and finally anti-mouse-conjugate. If polyclonal anti-Hp was used as capture antibody, the ELISA was non-specific, as cross-reaction with Hp occurred. Attempts to remove Hp by means of Hb bound to 1 µm aliphatic amine latex beads (Interfacial Dynamics Corporation, USA) were unsuccessful, only removing just over half of Hp. Alternative methods were attempted using anti-apolipoprotein A-I as capture antibody and anti-Hpr as detector antibody and vice versa but no significant OD values were obtained.
The method which was found to be sensitive and relatively specific was to use the monoclonal anti-Hpr as capture antibody and the polyclonal anti-Hp as detection antibody. The anti-Hpr antibody was diluted in PBS (10 µg/ml) and 100 µl aliquots used to coat 96-well plates (Nunc, Hereford, UK) overnight at 4°C. Following washes (PBS/0.05% Tween 20) and blocking (3% BSA in PBS), the samples, diluted 1∶1000 in 3%BSA/PBS, were added and the plate incubated at room temperature for 60 minutes. Detection was by rabbit anti-human Hp (Sigma, Poole, UK), diluted to 3.33 µg/mL, followed by goat anti-rabbit IgG alkaline phosphatase conjugate (Sigma, Poole, UK), diluted to 46 ng/mL. The substrate was p-nitrophenyl phosphate (Sigma, Poole, UK) at 1 mg/ml in 10% diethanolamine, containing 0.5 mM MgCl2, pH 9.8. The reaction was allowed to proceed for 1 hour and the OD at 405 nm was measured. The ELISA was calibrated using purified Hpr as a standard prepared by detergent solubilization of human HDL followed by affinity chromatography using a matrix of anti-Hpr (SF14.11) coupled to Affi-gel 10 beads (Bio-Rad)
The reproducibility of the ELISA was investigated by taking 10 plasma samples from random UK blood-donors (courtesy of John Radcliff Infirmary) and running the samples on the same and different ELISA plates. The intra-plate coefficient of variation (CV) was found to be 4.2%. The inter-plate CV was 21%, but for the majority of samples the CV was <15% with only a few samples with low Hpr concentrations giving higher CV values.
The similarity in structure between Hpr and Hp might have caused cross-reactivity within the ELISA, particularly since it has previously been shown that 0.9% of plasma Hp associates with HDL
To investigate whether the Hpr ELISA was specific for Hpr rather than Hp, it was performed using various dilutions pooled Hp (Sigma) (stock concentration = 100 µg/ml). To determine whether the ELISA detected Hpr in HDL, it was performed using various dilutions of a purified HDL fraction (stock concentration = 22.5 mg/ml total protein of which 300 µg/ml was found to be Hpr as determined using the Hpr standard). To show that the ELISA could detect Hpr in plasma, it was performed with dilutions of pooled sera (Courtesy of John Radcliffe blood bank). All results are means of duplicate readings and are minus readings from blank wells.
Plasma Hp levels were determined by ELISA
Levels of CRP were determined using a commercial ELISA kit (American Laboratory Products Company, Windham, NH, USA). The cut off value to define an APR is ≥10 µg/ml
Albumin concentrations were determined by mixing 280 µl bromocresol green reagent (Abbott Laboratories Ltd, UK) and 2.4 µl sample in duplicate and reading optical density at 630 nm. Blanks (water), standards (25, 57 g/l) and a quality control (29 g/l) were also prepared. The mean blank value was subtracted from all mean values. A graph was drawn using the two calibration points (straight line up to 60 g/l) and sample values read off. Hypoalbuminemia is defined as levels <35 mg/ml
Differences between categorical variables were assessed using chi-square tests. Differences between categorical variables and continuous variables were assessed using Mann Whitney
A general linear model was used to examine the effect of variables of interest on Hpr levels in Gabon. Since Hpr, Hp and CRP levels showed heteroscedasticity, the levels were log transformed before analysis. Variables of interest included CRP concentration (continuous and 2 categories), Hp (continuous and 2 categories), albumin (continuous and 2 categories),
Levels of Hpr and Hp were determined by ELISA in children from Gabon (aged 1–12 years, 53% male) and Caucasian children (aged 3–4, 47% male) from the United Kingdom. The median concentration of Hpr in Gabon was 0.28 mg/ml which was significantly higher than healthy Caucasian controls (0.049 mg/ml), overall and with age-specific comparisons (
Scatterplot of raw Hpr values (mg/ml) plotted against raw Hp values (mg/ml). Associations were assessed by Spearmans rho. Hpr was positively correlated with Hp in A) Gabonese children (rs = 0.29,
Gabon | Caucasians | ||||
n | Hpr mg/ml | n | Hpr mg/ml |
|
|
Overall | 563 | 0.28 [0.17–0.46] | 36 | 0.049 [0.041–0.069] | <0.001 |
Age (years) | |||||
1–4 | 173 | 0.36 [0.21–0.5] | 36 | 0.049 [0.041–0.069] | <0.001 |
5–9 | 318 | 0.26 [0.17–0.44] | |||
10–14 | 61 | 0.22 [0.15–0.34] | |||
Sex | |||||
Female | 259 | 0.32 [0.18–0.5] | 19 | 0.054 [0.041–0.08] | <0.001 |
Male | 294 | 0.25 [0.16–0.4] | 17 | 0.047 [0.041–0.061] |
Values represent median [interquartile range]. P-values represent comparisons between Caucasians and Gabonese children.
Variable | All children | |
(n = 553) | ||
C-reactive protein (µg/mL) | 50.3 | [13.4–187.3] |
C-reactive protein (>10 µg/mL) |
444 | (80.3) |
Haptoglobin (mg/mL) | 0.13 | [0.03–0.44] |
Hypohaptoglobin |
296 | (53.5) |
Albumin (mg/mL) | 40.61 | (11.5) |
Low Albumin |
83 | (15.1) |
296 | (53.6) | |
Density/µL |
800 | [286–3314] |
Haptoglobin genotype | ||
1F-1F | 92 | (16.6) |
1F-2 | 184 | (33.3) |
1S-1F | 91 | (16.5) |
1S-1S | 22 | (4.0) |
1S-2 | 85 | (15.4) |
2-2 | 79 | (14.3) |
Continuous data are shown as median value [inter-quartile range] or mean (standard deviation) and categorical data are n (%). Note for C-reactive protein, albumin and plasmodium prevalence n = 550.
C-reactive protein >10 µg/mL was the cut-off used to define an acute phase response
The cut off value for hypohaptoglobinaemia is <0.18 mg/mL
The cut off value for low albumin is <35 mg/mL
The frequency of species in this population has been published previously
We wished to identify variables that were associated with Hpr levels at the population level. Variables chosen for investigation were APR proteins (Hp, CRP and albumin),
Variable |
|
(95% CI) |
|
Acute Phase Proteins | |||
Haptoglobin | 0.13 | (0.09, 0.16) | <0.001 |
C-reactive protein | 0.001 | (−0.04, 0.04) | 0.96 |
Albumin | 0.015 | (0.005, 0.025) | 0.004 |
Haptoglobin genotype | |||
1–1 | 0.31 | (0.14, 0.49) | 0.001 |
2–1 | 0.17 | (0, 0.34) | 0.05 |
2–2 | Reference | ||
Haptoglobin subtype | |||
1F-1F | 0.41 | (0.2, 0.61) | <0.001 |
1F-1S | 0.31 | (0.1, 0.51) | 0.004 |
1S-1S | −0.05 | (−0.37, 0.27) | 0.75 |
2-1F | 0.2 | (0.02, 0.38) | 0.03 |
2-1S | 0.11 | (−0.1,0.32) | 0.3 |
2-2 | Reference | ||
Demographics | |||
Age | −0.04 | (−0.06, −0.02) | <0.001 |
Males | −0.19 | (−0.3, −0.08) | 0.001 |
Females | Reference | ||
|
|||
Positive | −0.16 | (−0.28, −0.05) | 0.006 |
Negative | Reference | ||
Log10(Plasmodium+1) | −0.06 | (−0.09, −0.02) | 0.003 |
Coefficients are from univariate general linear models investigating the association between variables of interest and log(Haptoglobin-related protein) values.
A multivariate general linear model was then constructed to examine the effect of multiple variables on Hpr levels. The resulting fitted model predicts the median Hpr level for the population controlling for relevant variables. Variables chosen for the final model included age, sex, Hp genotype and Hp concentration. The median Hpr level in the population, after controlling for relevant variables, was 0.27 mg/ml (95% CI 0.25, 0.29). Hpr levels decreased with age (
Data shown represent the first large-scale population study of Hpr levels in humans. Previous laboratory investigations, using a small number of samples, reported lower Hpr levels in the range of 0.02–0.05 mg/ml
Hpr levels were found to decrease with age in the Gabonese children. This may be dependent upon intrinsic factors related to the age of the host. It is interesting to note that some acute phase proteins show age-related profiles, e.g. levels of Hp decrease with age, whereas others are unrelated to age, e.g. CRP
Univariate analysis showed that Hpr was correlated with the acute phase protein Hp, but not CRP or albumin in Gabonese children. Hp levels are determined by several factors, including IL-6-dependent APR, Hp genotype, parasite density and hemolysis
Hpr levels were inversely correlated with parasitemia. Hp levels also decrease with increasing parasite density due to intravascular hemolysis and subsequent clearance of Hb-Hp complexes
Hpr levels were also associated with Hp genotype; Hp2-2 individuals had significantly lower levels of Hpr compared to Hp1-1 and Hp 2-1 individuals. This may be related to the differential affinity of Hp genotypes for binding Hb and subsequent patterns of clearance. Whilst the affinity of the higher molecular weight polymers for Hb is generally lower than that of Hp1-1, their plasma levels are lower, reflecting the clearance of a greater molecular weight per molecule. Thus levels are in the phenotype order: Hp1-1>Hp2-1>Hp2-2
Our results show that levels of Hpr vary between and within populations from the United Kingdom and Gabon. Since all humans are refractory to infection with
Our population-based observations indicate that a more dynamic view of the relative roles of Hpr and Hpr-Hb complexes needs to be considered in understanding innate immunity to African trypanosomes and possibly other pathogens that cause ahaptoglobinemia, such as the newly discovered
Caucasian samples were kindly provided by AJ Pollard and D Kelly, Oxford Vaccine Group and the Oxford Biomedical Research Centre, Department of Paediatrics, University of Oxford.