Background
Severe
Plasmodium falciparum malaria has exerted an enormous mortality toll over the course of human evolution and as of today, the disease still causes approximately 400,000 annual deaths [
1].
Severe malaria (SM) is a heterogeneous disease state with a multifaceted and poorly understood pathogenesis [
2]. Besides a wide range of parasite factors, several human host genetic factors have been shown to influence the susceptibility to SM, including the sickle-cell trait, α
+-thalassaemia and the histo-blood group (Bg) ABO system [
3‐
6]. Consequently, many of these have been evolutionary selected for in human populations in malaria endemic regions, which also highlights the influence this disease has on the trade-off between risk and benefit for human adaptation [
7].
Epidemiological and genome-wide association studies have provided robust evidence of a protective effect of BgO as compared to non-O Bgs with regards to development of SM [
5,
6,
8,
9]. SM has also been strongly linked to the ability of
P. falciparum iRBCs to bind to uninfected RBCs and the vascular endothelium, cytoadhesive events known as rosetting and cytoadherence, respectively [
10,
11]. These phenomena collectively contribute to blockade of oxygen transport in the microvasculature and endothelial activation with an augmented inflammatory response, all hallmark features of the disease pathobiology [
10]. Rosetting is mediated by PfEMP1, RIFIN and STEVOR, parasite ligands that interact with a plethora of RBC receptors such as BgA, complement receptor 1 (CR1) and glycophorin C (GYPC) [
12‐
15]. BgA rosetting has been shown to be more prominent compared to BgO, with augmented rosetting rates (RR) and rosette sizes (RS) [
6,
13,
16]. Further, BgA-mediated rosetting has been suggested to confer protection from the immune system of the host, by preventing exposure of important immunogenic parasite-derived epitopes at the RBC surface [
17].
BgA has more extensive, genetically driven, variation in erythrocytic antigen expression as compared to BgB, with A1 RBCs possessing approximately five times more A antigen than A2 RBCs [
18]. Similarly, BgAB can be divided into A1B and A2B depending on the A-transferase allele [
19]. Although there is a general consensus of A antigens promoting cytoadhesive events, the data on rosetting in weak versus strong BgA subgroups is scarce and based on investigations of relatively few donors [
13,
16,
20]. In addition, besides host cell resident ABO antigens, differences in serum component levels, such as soluble ABO antigens, von Willebrand Factor, ICAM-1 and E-selectin exist and their relative contribution in the context of Bgs is not well understood [
21‐
23].
A better understanding of how host factors are involved in the cytoadhesive events implicated in SM pathogenesis is warranted. The current paper explores the effect of all four ABO Bgs on P. falciparum rosetting, with a particular emphasis on low A antigen expression levels.
Methods
Parasite cultures
The parasite lines used in this study were the two laboratory parasite clones FCR3S1.2 and PaloAltoVarO (PAvarO) and four culture-adapted clinical isolates, UCM83, UKM74, UKM104 and UKS111 collected in Kampala, Uganda [
24,
25]. All parasite lines were cultivated according to standard procedures [
26,
27]. Briefly, all cultures were grown in malaria culture medium (MCM) (RPMI 1640 (Gibco) supplemented with 2 mM
l-glutamine (Hyclone), 2.5 µg/mL gentamicin (Gibco), bicarbonate at final concentration of 32 mM (Sigma) and 10% A + serum) at 4% haematocrit under constant microaerophilic condition (5% O
2, 5% CO
2, 90% N
2) and in suspension on an orbital shaker. Prior to assaying, parasites were propagated in O+ erythrocytes with 10% AB+ serum for the laboratory clones and 15% AB+ serum for the patient isolates.
Sample collection
All RBC and serum samples used in the study were obtained from healthy Swedish blood donors. For RBC samples, intravenous blood was collected in Vacutainer® citrate tubes (BD) and serum samples were collected into Vacutainer® SST™II Advance gel tubes (BD). Aliquots of sera were stored at − 20 °C and heat inactivated at 56 °C for 30 min prior to use.
ABO antigen semi-quantification
Subdivision of A1-negative and A1-positive BgA RBCs was performed using anti-A1 lectin (Ortho Clinical Diagnostics, catalogue number 711830) at the site of collection. This serological classification was not available for the BgAB samples, since this was not part of routine clinical practice. The semi-quantification of A and B antigens by flow cytometry was performed as previously described, with 10,000 RBCs acquired per sample [
28]. In brief, RBCs were blocked in 2% bovine serum albumin (fraction V, HyClone) in PBS (BSA-PBS) for 5 min at room temperature. Thereafter primary antibodies (anti-BgA (sc-52367) and anti-BgB (sc-52371, Santa Cruz Biotechnology) were added at 10 µg/mL for 10 min at room temperature. After two washes with PBS, FITC conjugated secondary anti-mouse IgM (μ-chain specific) (Sigma, F9259) was used at 1:10 dilution in 2% BSA-PBS for 10 min at room temperature. Lastly, RBCs were washed with PBS and resuspended in PBS for acquisition.
Rosetting assays
For rosetting assays involving RBCs of different Bgs, late stage parasite cultures were diluted 1:25 in donor RBCs in triplicates in 96-well plates for two cell cycles, with a desired final parasitaemia of 5–7.5%. For serum rosetting assays, parasites were grown in BgO RBCs in 10% of each serum for one cell cycle in triplicates. All samples were kept at 1% haematocrit. The rosetting assays were either investigating the innate rosetting characteristics or the rosette-disrupting propensity in different ABO Bgs. To evaluate rosette disruption by heparin and PfEMP1-DBL1α-antibodies (anti-var60 and anti-varO) [
29], parasite cultures were incubated either with heparin (from Porcine intestine, Sigma, H3393) or antibodies at concentrations indicated for specific experiments at 37 °C for 45 min in MCM and measured by flow cytometry. For determination of rosetting rate (RR) and rosette size (RS) by microscopy, parasite cultures were stained with acridine orange [
26]. The RR of 150–200 late stage iRBCs was counted for each sample and the number of bound RBCs in 100 rosettes were counted to determine RS.
Flow cytometric analysis of rosetting rate and rosette size
The flow cytometric assays were performed using FACSVerse with universal loader (BD Bioscience). The cuvette cross section of this machine is 430 μm × 180 μm and data was acquired at medium flow rate (60 µl/min) with sheath core stream fluid velocity of 5.4 m/s. Samples were kept in suspension between the reads by 5 s shaking at 1400 rpm every 10 min. To discriminate between rosetting and non-rosetting late stage iRBCs, cells were co-stained with 10 μg/mL Hoechst 33,342 (Invitrogen, H3570) and 5 μg/mL Dihydroethidium (DHE, Invitrogen, D1168) in RPMI 1640 (HyClone) for 45 min at 37 °C. As described previously [
30,
31]. The percentage of multiplets (events represented by 2 cells or more) after gating on this late-stage parasite population has previously been shown to correlate with RR determined by microscopy [
31]. This gating strategy was adopted, with further optimizations, in order to enable absolute quantification of RR, as well as relative comparison of RS. Briefly, the initial FSC-SSC gating was excluded in order to include larger multiplets. Key to successful absolute quantification of RR and relative measures of RS was to allow larger multiplets to be identified, which simply accommodates inclusion of larger rosetting events. For this, voltages were adjusted so the cell population would be located closer to the lower left quadrant in the SSC-A–FSC-A plot. This resulted in tube target values of approximately 0.09 for FSC and 0.75 for SSC, which is on average 30% lower than for the previously described assay.
The gating was performed in the following order: Double-positive events for Hoechst and DHE (trophozoites and schizonts), FSC-A vs FSC-H for multiplets to determine RR based on the uninfected RBCs from the same sample and from there the mean SSC-A of multiplets for RS (Fig.
2A–C). Data from 1000–5000 late stage iRBCs were collected for the assays, depending on numbers and parasitaemias of samples.
Antibody recognition of PfEMP1
Antibody recognition of PfEMP1 on iRBCs cultured in the four major ABO Bgs were performed using strain-specific PfEMP1-DBL1α-antibodies according to previously described methods [
17]. Briefly, late-stage FCR3S1.2 and PAvarO cultures were incubated in 2% BSA-PBS for 45 min, followed by 45 min of incubation with strain-specific antibodies at 10 μg/mL or non-immune goat IgG 10 μg/mL. Thereafter, samples were washed with 2% BSA-PBS and incubated for 45 min with Alexa-488 coupled rabbit anti-goat IgG antibodies (1:100) (Invitrogen, A11078) and Hoechst 33342 (10 μg/mL) (Invitrogen, H3570) in 2% BSA-PBS. The samples were then washed three more times in 2% BSA-PBS and resuspended in 2% BSA-PBS with subsequent flow cytometric cell acquisition. All steps were done at 37 °C. Data was acquired with a flow cytometer as described above. The production and purification of anti-var60 (from strain FCR3S1.2), anti-varO (PAvarO) and non-immune polyclonal ChromPure goat IgG (Jackson Immuno Research, 005-000-003) used in this work was previously described in [
29].
To retrieve detailed ABO blood group allele frequencies, the Erythrogene database (
https://www.erythrogene.com/) was used, containing next-generation sequencing data for all 36 blood group alleles in the 1000 Genomes Project [
32]. 235 alleles were retrieved upon searching for ABO alleles, for which the ABO phenotype was classified into O, A1, A2, Aweak, Ax/Aweak, B, Bweak, B(A), B3 and unknown (previously uncharacterized ABO alleles). The results were analysed and displayed per continent (Africa, America, East Asia, Europe, South Asia).
Data analysis and statistics
The cell acquisition was done using FACSVerse (BD Bioscience) flow cytometer and data was analysed using FlowJo (v10). All statistical analyses were performed in R version 4.0.3. Unpaired two-tailed t test was used for all pairwise comparisons of RR, RS and the relative mean fluorescence intensity (MFI) between different ABO Bgs, with p-values adjusted for multiple comparisons using the Holm method. Spearman rank’s correlation test was used for correlation testing.
Ethics statement
All RBC and plasma samples used were collected from the Karolinska University Hospital Blood Bank and approved by the Regional Ethical Review Board in Stockholm, Sweden (Dnr 2009/668-31/3). The collection of clinical isolates in Uganda was approved by Karolinska Institute’s Regional Ethical Review Board (permission 03/095) and the Uganda National Council for Science and Technology (permission MV717). Written informed consent was obtained from the parents or guardians of the patients.
Discussion
That ABO Bg influences susceptibility to severe
P. falciparum malaria and rosetting has been implicated mainly by the previously observed BgO protective effect [
17,
33]. Opposite to this, the heterogenous BgA has been suggested as a potential risk factor for SM and although data is limited, there appears to be a more heterogenous distribution of BgA alleles in Africa, which coincides with the highest transmission of
P. falciparum. The aim of this paper was to gain insights to this by evaluating the role of Bgs in rosetting, a well-known and SM associated virulence feature of the parasite, with an emphasis on the sub-division of BgA [
34‐
36].
Key to this interrogation was the development of a more robust methodology for high-throughput characterization of the rosetting phenotype as the traditional analysis by microscopy is laborious, quantitatively weaker and prone to subjective interpretations. This work led to the possibility of accurately measuring absolute RRs by flow cytometry, which abolished the need for microscopy completely in that respect. For RS however, which is a more heterogenous and less on/off phenotypic read out, the dynamic range of the data prevented absolute measures. Still, the correlation between RS determined by flow cytometry and microscopy was deemed sufficient.
The use of this methodology allowed for a thorough investigation of Bg influence on the rosetting phenotype for a wide panel of parasite strains and clinical isolates using a large number of RBC and serum donor samples. Indeed, an associations between RR and RS and the major blood groups was identified as expected, associations that were further strengthened upon addition of rosette disrupting agents. BgA and BgAB clearly stood out in mediating both elevated RR and RS and protecting against disruptive agents, although with large inter-parasite variations. The latter was expected as the rosetting phenotype is not solely dependent on Bgs but is phenotypically diverse and also entails the use of other receptors on the host RBCs [
37]. In fact, the plasticity of the phenotype also goes beyond the use of host cell receptors and also includes a wider repertoire of parasite antigens beyond the PfEMP1 family of proteins. Thus, it is plausible that the parasites used in this study that did not reveal any Bg dependency of their rosetting make use of additional ligands and/or non-Bg RBC receptors. Still, for the Bg dependent parasite lines with known PfEMP1 expression and availability of specific DBL1α antibodies, a robust protection of parasite antigens in BgA rosettes was seen. It has been previously observed that rosettes formed with BgA RBCs have reduced epitope accessibility for antibodies against PfEMP1, but here these findings were extended to all non-O Bgs [
17]. PfEMP1 serves as a target of naturally acquired immunity and thus, a reduced epitope accessibility in non-O Bgs might be an important factor underlying the increased risk of contracting severe malaria [
17,
38].
Several subtypes of the BgA have been defined, with two major subgroups being A
1 and A
2. A
1 RBCs possess approximately five times more A-antigen than A
2 RBCs as well as a more of the repetitive A epitope glycolipid motifs [
18]. Structural studies of the PfEMP1 RBC binding DBL1α
1 domain of PAvarO parasites showed a higher binding of A1 RBCs compared to A2 and Ax RBCs [
20]. These findings got further support from Goel et al. who demonstrated that FCR3S1.2 forms bigger rosettes with A1 RBCs but no difference in RR [
13]. Herein, a dependence of RS on the amount of A-antigen present on RBCs was observed with significantly larger rosettes formed in A1-positive blood. Furthermore, A1-positivity made rosettes more resistant to the disruptive agent heparin and antibodies against PfEMP1. Thus, the data presented here support BgA in general and subtypes of BgA in particular as risk factors for SM. In the view of these findings, one would anticipate a stronger selection of BgO, and within BgA a selection for low antigen A, in malaria endemic regions. Although GWAS studies have confirmed the protective effects of BgO for SM, the relative allele distribution of Bgs between populations in malaria endemic and non-endemic continents are not that striking. This is possibly a result of the phenotypically plastic rosetting phenomenon, where a single family of receptors is likely not enough of a force for stronger evolutionary selection.
Another contributing factor behind the plastic rosetting phenomenon was evaluated, namely the role of human serum. Differences between ABO-Bgs stretch beyond the RBC surface antigens. Variations in the levels of plasma components such as VWF and ICAM-1 have been reported [
21,
22,
39]. Further, serum factors have been indicated to form intermolecular bridges between parasite antigens on iRBCs and receptors on surrounding RBCs. However, the current study did not find any significant changes in RR nor RS when exposing iRBCs to sera from different Bg donors. These findings suggest that Bg mediated rosetting fully relies on host cell resident ABO antigens. It is however important to note that this might be influenced by limitations of the in vitro cultivation system in mimicking the more complex in vivo setting, where serum ABO Bgs might still play a more substantial role.
The potential impact of SM on the distribution of different BgA alleles in different continents has been previously suggested but the significance of weak phenotypes has not been properly explored [
13,
20]. Past epidemiological studies have been pooling the heterogeneous BgA subgroups into the same category, potentially resulting in an underestimation of risk for severe disease. The future studies ideally should subdivide BgA in order to comprehensively investigate the effect of high and low expressing BgA phenotypes on the risk of contracting SM. This is probably also important beyond the role of Bgs for malaria pathogenesis, as BgA has been associated to adverse outcomes also for other diseases, such as in thromboembolic diseases and more recently for COVID-19 [
40,
41].
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.