Total and functional parasite specific IgE responses in Plasmodium falciparum-infected patients exhibiting different clinical status

Venous blood was collected on EDTA in sterile vacutainers from each patient on the day of hospitalization (day 0, before any treatment), and seven (day 7) and thirty days later (day 30). Plasma was obtained by centrifuging the blood samples at 5000 rpm for 15 min. Plasma samples were stored at -80°C until use.

Erythrocytic stages of the P. falciparum malaria parasite line FAN 5HS (source: NCCS, Pune, India) and 3D7 were cultured using candle jar dessicators as previously described [19]. The culture medium was RPMI 1640 (Gibco-BRL), supplemented with 0.5% AlbuMix (Gibco BRL). The cultures were maintained in six-well or 24-well tissue culture plates (NUNC). Parasitaemia was 5% at the start of culture and reached 25% after six days. Culture medium and fresh RBCs were added every other day.

Parasite soluble antigen was prepared from synchronous cultures containing more than 20% mature trophozoites; more than 6% rings and more than 5% schizonts were used. The cultures were pooled and centrifuged at 3,000 rpm at 4°C, and the pRBC pellet was kept and the supernatant discarded. The pRBC pellet was suspended in 10 ml sterile PBS 1 × (0.15 M, pH7.2) and then centrifuged. The parasitized red blood cell (pRBC) pellet was washed five times and then lysed by adding 15 ml of 0.1% saponin. The saponin treatment frees the parasites from the infected RBCs. This was centrifuged at 6,000 rpm for 30 min at 4°C. The supernatant was discarded and the parasite pellet was washed five or six times with sterile cold PBS. The parasite pellet was resuspended in 1 ml protein isolation buffer containing a cocktail of protease inhibitors. This was briefly sonicated and the tube was kept at 4°C for between four and five hours. The contents of the tube were agitated by cyclo-mixing and then centrifuged at 6,000 rpm for 30 min at 4°C. The clean supernatant was collected in a separate tube and the pellet was discarded. The contents were sterilized by passing through 0.22 μm-pore filters. Aliquots of the antigen were frozen at -70°C until use. Parasite proteins were quantified by the Bradford method. The concentration of the parasite line FAN 5HS and 3D7 were 1.2 and 2.6 mg/ml, respectively.

Normal red blood cell (RBC) extract was prepared from the same batch of RBCs used for culturing the parasites, and followed the same procedure as previously described for pRBCs. Briefly, the RBCs were washed with PBS and the buffy coat was removed. After centrifugation, the RBC pellet was suspended in 1 ml protein isolation buffer containing a cocktail of protease inhibitors. This was briefly sonicated and the tube was kept at 4°C for between four and five hours. The contents of the tube were agitated by cyclo-mixing and then centrifuged at 6000 rpm for 30 min at 4°C. The clean supernatant was collected in a separate tube and the pellet discarded. The protein contents were estimated using a protein determination kit (BCATM protein assay Kit, Pierce, France).

An ELISA method was used to detect total IgE plasma levels in samples corresponding to day 0, day 7 and day 30. ELISA plates (96 microwell plates, reacti-bind 96 EIA Plate 100/PKG, Pierce) were coated with 50 μl/well of purified sheep polyclonal anti-human IgE solution at 5 μg/ml (The Binding Site, Birmingham UK) by incubation overnight at 4°C. The plasma samples were diluted 1:5 and incubated for two hours at 37°C. Bound IgE was detected using a peroxydase-conjugated polyclonal anti-human-IgE (The Binding Site, Birmingham UK). Binding was revealed using the OPD substrate (Sigma) and the product was quantified from the optical density (OD) at 450 nm. Serial dilutions ranging from 2 μg/ml to 0.0019 μg/ml of IgE solution (human monoclonal IgE provided by Dr Thierry Batard – Stallergenes, Anthony, France) gave the standard curve. The median of each optical density value was fitted into the sigmoidal standard curve using a specific ELISA programme running in Igor version 3.16 (Wavemetrics, Lake Oswego, OR).

A new rat mast cell line RBL-2H3 transfected with a human Fcε RI α-chain that triggers degranulation upon human IgE cross-linking was used [20]. Cells were maintained in Dulbecco medium (Gibco BRL, Eragny, France) containing 10% foetal bovine serum (FCS), 100 U/ml penicillin and 100 U/ml streptomycin (GIBCO BRL, France). Cells were expanded by incubation at 37°C for three to four days in complete Dulbecco medium supplemented with G418 (GIBCO BRL, France).

β-Hexosaminidase is known as a component of the basophil and the mast cell specific granule, and is released during degranulation of these cells [21]. Degranulation was monitored after antigen stimulation by measuring the level of released β-hexosaminidase. Fcε RI α-chain RBL-2H3 transfected rat mast cell line cultures (5 × 10⁵ cells per well) were incubated with the different serum samples at a non-cytotoxic dilution (previously determined) for 48 hours at 37°C in the absence of the G418 antibiotic. The upregulated receptors were saturated by incubation at 4°C for 30 minutes with the same samples diluted 1:10. The cells were then washed with PBS 1X, centrifuged and resuspended in 1 ml Tyrode buffer before being centrifuged again. Finally, the cell pellet was resuspended in 450 μl of D2O (50%) and Tyrode buffer (50%) solution and each culture sample was distributed to 10 ELISA plate wells. Different controls were carried out for each sample. Control cells on lane 1 and 2 were subjected to Triton disruption (Triton 5%) and represented 100% enzyme release. Cells on lanes 3 and 4 were incubated with 50 μl of complemented Dulbecco medium without serum and represented the background enzyme release. Lanes 5 and 6, 7 and 8, 9 and 10 were incubated with 50 μl of different duplicated concentrations of parasite extract (1,000, 100 and 10 ng/ml) for 30 minutes at 37°C. After centrifugation of each well sample, 50 μl of each supernatant was collected and incubated with 50 μl of PNAG substrate solution for 90 minutes at 37°C. The level of released β-hexosaminidase was estimated from the OD at 405 nm using a spectrophotometer. All results are expressed as the percentage of total β-hexosaminidase in the cells after correcting for spontaneous release in unstimulated cultures, calculated as following: (experimental β-hexosaminidase – background β-hexosaminidase)/(total β-hexosaminidase – background β-hexosaminidase) × 100.

FACS analysis was performed after incubating RBL-2H3-D12.8 cells with several dilutions of serum samples to follow the induction of the high affinity receptor (Fcε RI) expression after stimulation by IgEs in the patient's sera. Cells were incubated for 48 hours at 37°C with the different serum samples optimally diluted to avoid cytotoxicity. A saturation step with the same sera diluted 1:10 was done by incubation at 4°C for 30 minutes. Cells were washed with PBS 1X and incubated with FITC-labelled anti-IgE (Tebu, Le Perray en Yvelines, France) (1/100) for 30 minutes. Cells were washed again, centrifuged, resuspended in PBS 1X and analysed by cytofluorometry using Cellquest software (Beckton Dickinson, USA). 10,000 cells were acquired per tube.

The levels of cytokines in the plasma (IL-4, TNF, INF-γ, and IL-10) were estimated by Opti-ELISA kits (Pharmingen, San Diego, CA.USA) used following the manufacturer's instructions.

Due to a non-normal distribution of the scores in each group, non-parametric tests were performed, using the median to compare the different clinical groups. The Mann Whitney test was used for comparisons between two groups and the Kruskall Wallis test to compare three or more groups. Spearman's correlation was used to check for correlations between parameters. P values less than 0.05 were considered as significant. Chi-squared test was used to compare qualitative variables.

Total IgE levels were analysed in endemic controls and in cohorts of P. falciparum-infected patients with different clinical forms of malaria, ranging from asymptomatic to cerebral disease, from Gabonese and Indian endemic areas to study the association between the IgE response and disease severity. Total IgE levels were measured by ELISA in individual sera before drug administration (corresponding to day 0) and determined the general distribution in the studied populations from Gabon (Figure 1A) and India (Figure 1B). Total IgE concentrations were found to be much higher in patients from India (mainly adults) than in patients from Gabon (children). In both populations independent of the different levels of IgE in each population, the median IgE levels within each clinical group tended to increase upon infection (mainly in UM and SM groups), although the difference between the groups was only significant in the Indian population (Kruskall Wallis, p = 0.0005). As only Indian patients showed a significant difference, the Mann Whitney test was used to compare the different groups in this population only. There was a significant increase in IgE levels in the EC group compared to the NEC group (p = 0.042). The most significant increase in IgE levels (versus the EC group) occurred in the UM patients (p = 0.015) and in the SM patients (p = 0.013). No significant difference between the EC group and the CM and Ex-CM groups was observed.

A range of values of IgE levels was defined enabling the analysis of the frequency of normal, moderate and high IgE levels in each clinical group of patients. The so-called normal values were adjusted to the studied population because the Gabonese and Indian groups had different plasma total IgE ranges. Therefore, the normal value (N) was defined by the median IgE levels in the endemic controls of each study population. Consequently, all values between N and 2N were considered as low/moderate IgE levels and those between 2N and 3N as moderate/high IgE levels, with the highest levels being above 3N (Figures 1C and 1D). Even in Gabonese patients, for whom the increase of IgE in the disease groups was not significant, a higher percentage of patients with clinical disease had higher IgE levels than controls and asymptomatic patients. These differences were more marked in the UM, SM and CM Indian patients (Figure 1D). In the Indian population, the NEC group did not have moderate/high IgE levels, although a high percentage of patients exhibited normal IgE levels (Figure 1D). Also, no significant change was detected in IgE levels over time in the UM, SM and CM groups of the Gabonese cohorts when tested seven days and 30 days after treatment (Table 2). No significant association of malaria and IgE levels with sex in the two studied populations. However, a significant increase in IgE levels with age (p = 0.00034) was observed in the Gabonese subjects (Figure 2A) but not in the Indian subjects.

The correlation between IgE levels and the parasite load was tested. Although the general trend was different in the Indian and Gabonese population, there was no significant correlation between IgE concentration and parasite load for all groups together. In Gabonese cohorts, a negative correlation for all groups was observed, except for the UM patients where the correlation showed a positive tendency. In the Indian cohorts, a positive correlation was observed between IgE levels and parasite load, mainly in the UM and SM groups (Figure 2B).

Previous studies have used ELISA to quantify specific IgE present in the serum [7, 8, 10, 11]. The functionality of specific IgEs present in the serum was studied by evaluating the ability of these IgEs to induce mast cell degranulation in the presence of the parasite antigen. A rat mast cell line transfected with the human α-chain of Fcε RI was used [20]. Human Fcε RI expression was induced after incubation with all serum samples at non-cytotoxic dilutions. FACS was used to detect the presence of FcåRI receptors on the mast cells surface induced by IgE present in serum samples. Although the fluorescence intensity revealing human Fcε RI expression by the mast cells varied between patient samples, IgE receptors were upregulated in all the samples tested. No correlation between total IgE levels in the serum and the up-regulation of mast cell receptors was found. There was no significant correlation between IgE levels and receptor upregulation and the level of mast cell degranulation. Sensitized cells were then incubated with different concentrations of a P. falciparum blood-stage antigen extract. Mast cell degranulation was measured by quantifying the release of β-hexosaminidase. Control cells not exposed to any serum gave a maximum mast cell degranulation of 3%. Therefore, serum samples giving an enzyme release greater than 5% in the presence of at least one of the antigen concentrations were considered positive for functional IgE. In the Gabonese cohorts (Figure 3A), there were functional P. falciparum IgE in all clinical groups. However, the percentage of patients with functional specific anti-parasite IgE was higher in asymptomatic and uncomplicated malaria patients than in other groups. Also, the percentage of patients displaying parasite-specific IgE was lower in the group exhibiting severe disease. The distribution of patients per group releasing between 5 and 10%, 10 and 30% and above 30% β-hexosaminidase induced by specific anti-parasite IgE revealed that one patient in CM group had a degranulation level above 30%, being the highest induced response among all the tested individuals. The same assay was carried out on the Indian population. Although there was no significant difference between groups, the percentage of patients having functional IgE recognizing the parasite extract was slightly higher in EC and UM groups than in SM and CM groups (Figure 3B). All positive patients had an enzyme release of between 10 and 30%. No significant correlation was found between P. falciparum-specific IgE-induced mast cell degranulation levels and sex, age and parasitaemia.

IgE production is influenced by cytokines produced by activated T cells. These cytokines are also involved in pathophysiological mechanisms associated with severe malaria [4, 22, 23]. Therefore, the relationship between the cytokine profile, the IgE levels and the clinical manifestation was investigated. IFN-γ, TNF and IL-10 levels were measured in the sera of the Indian and Gabonese patients. IL-4 levels were measured only in Indian patients as it was not different between the Gabonese groups. IFNγ concentrations were highest in the Gabonese AI and CM groups (Table 3). This cytokine is significantly higher in the AI and CM groups than in EC (p = 0.02 and p = 0.009 respectively). The plasma TNF concentration was similar in the severe SM and CM groups and in disease-free EC and AI groups. TNF levels were clearly higher in SM and CM groups than in the UM group (p < 0.001). Surprisingly, EC and AI also exhibited higher TNF levels when compared to UM group. No association between IFN-γ or TNF and IgE levels were found. However, a significant positive correlation was found between the concentration of total IgE and IL-10 in the UM group (p = 0.02) and a significant negative correlation in the AI group (p = 0.02) (Figure 4). The median levels of the different cytokines in the plasma of Indian patients are given in Table 4. IL-10 and TNF levels were higher in CM patients than in controls and other P. falciparum-infected patients. The plasma concentrations of these cytokines were moderate in cured CM patients (ExCM). Their levels of IL-10 and TNF were slightly higher in endemic controls than in non-endemic controls. Levels of IFN- were lower in the CM group than in AI group. No difference was found for the levels of IFN-γ between the uncomplicated and severe disease. Although, there was a significant correlation between IgE levels and IFN-γ (Figure 5A), TNF (Figure 5B) and IL-10 (Figure 5C) levels when looking at all the groups combined. Most diseased groups had high cytokine levels, whereas control groups had lower levels (Figure 5A, B and 5C).