CD44 cross-linking promotes Plasmodium falciparum invasion

Abstract

The ability of Plasmodium falciparum to invade and replicate asexually within human red blood cells (RBCs) is central to its pathogenicity. Invasion involves several host-parasite interactions, yet the required host factors remain underexplored, largely due to the intractability of mature RBCs. The transmembrane protein CD44 was identified as a host factor for P. falciparum invasion through a novel forward genetic screen. Here, we identify an anti-CD44 monoclonal antibody, BRIC 222, that significantly promotes P. falciparum invasion through CD44 cross-linking. CD44 cross-linking induced changes in the phosphorylation of RBC cytoskeletal proteins, consistent with a proposed role for CD44 as a co-receptor during invasion. CD44 cross-linking also altered the RBC membrane, increasing the accessibility of several surface proteins, including the essential invasion receptor Basigin. The parasite ligand Erythrocyte Binding Antigen-175 (EBA-175), which interacts with CD44, enhanced P. falciparum invasion and induced RBC membrane changes similarly to BRIC 222. Moreover, both BRIC 222 and EBA-175 increased binding of the PfRH5/PCRCR invasion complex to Basigin, an interaction known to be essential for invasion. We propose that CD44 cross-linking, potentially by EBA-175, serves to coordinate and enhance ligand-receptor interactions and promote signaling to the host cell cytoskeleton, making RBCs more permissive to P. falciparum invasion.

Fig. 1. Cross-linking of RBC CD44 promotes Plasmodium falciparum invasion.

a Approximate target epitope sites for each anti-CD44 antibody (BRIC 222, BRIC 235, KZ1, IM7), highlighted and color coded. b Invasion efficiency of Plasmodium falciparum strain 3D7 in presence of anti-CD44 monoclonal antibodies relative to an isotype control (BRIC 170). The parasitemia was measured ~21 h post-infection, and normalized to isotype control. Each dot represents one biological replicate (N = 3 biological replicates (individual donors) for all conditions except for IM7 dilutions where N = 4; n = 3 technical replicates for each condition). Error bars indicate SEM. Statistical analysis: Kruskal-Wallis with FDR correction; * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001. c Invasion efficiency of P. falciparum strain 3D7 in the presence of the indicated antibodies. The parasitemia was measured ~21 h post-infection, and normalized to an infection with no antibody (No Ab). Each dot represents one biological replicate (N = 3 biological replicates; n = 3 technical replicates). Error bars indicate SEM. Statistical analysis: One-way ANOVA with FDR correction; ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001. d–f Invasion efficiency of P. falciparum strains W2mef (d), D10 (e), and 7G8 (f) in the presence of the indicated antibodies, relative to no antibody (No Ab). Antibody concentration at 10⁰ is 25 µg/ml for all except the BRIC 222 F(ab) used in W2mef assay (34 µg/ml). The parasitemia was measured ~21 h post-infection. Each dot represents one biological replicate (N = 2 biological replicates; n = 3 technical replicates). Error bars indicate SEM. Statistical analysis: One-way ANOVA with FDR correction; * p = 0.0108 (e), 0.0219 (f), ** p = 0.0067. g Invasion efficiency of P. falciparum strain 3D7 in presence of 0.75 µM RII EBA-175, 25 µg/ml BRIC 222, or vehicle control, relative to no additive. Each dot represents one biological replicate (N = 4 biological replicates; n = 3 technical replicates). Error bars represent SEM. Statistical analysis: One-way ANOVA with FDR correction; * p = 0.0126, *** p = 0.0003, ns: non-significant. Source data are provided as a source data file.

Fig. 2. RBC agglutination is not directly correlated with enhanced P. falciparum invasion.

a Gating strategy for flow cytometry analysis of RBC clusters events from a representative experiment. RBC agglutination/clustering of three cells or more is denoted as “Clusters” in orange. b RBCs were treated with 25 µg/ml anti-CD44 monoclonal antibodies or 2 µM RII EBA-175 and assessed for agglutination/clustering by flow cytometry, measured by forward scatter (size). Frequency of cluster events from three donors with average ± SD is shown. Statistical analysis: Kruskal-Wallis with FDR correction; * p = 0.0468, ** p ≤ 0.01. c Representative images of RBCs treated with 25 µg/ml anti-CD44 monoclonal antibodies (BRIC 222, IM7, BRIC 235, KZ1) or 2 µM RII EBA-175, taken with 20X objective. d RBCs were treated with 25 µg/ml of the indicated monoclonal antibodies or their respective F(ab)s. Average frequency of cluster events from 5 donors (No Ab, BRIC 222 F(ab)) or 6 donors (BRIC 222, KZ1, KZ1 F(ab)) ± SD is shown. Statistical analysis: Kruskal-Wallis with FDR correction; *** p = 0.0006. e The relative invasion efficiency within populations of “Clusters” or “Single cells” was measured by flow cytometry and normalized to the mean of the no antibody (No Ab) condition. Each dot represents an individual biological replicate (N = 4 biological replicates; n = 3 technical replicates). Error bars represent SEM. Statistical analysis: paired t-test, one-tailed; ** p = 0.0069, * p = 0.0112.

Fig. 3. CD44 cross-linking induces phosphorylation of RBC cytoskeletal proteins.

RBCs were stimulated with BRIC 222 or BRIC 222 F(ab) at 25 µg/ml, processed into ghosts, and used for 2D-DIGE. a Total Protein and Phosphoprotein gels, along with the overlays of each condition are shown. Spots with differential phosphorylation, with ≥1.5 or ≤0.67 fold-change in phosphorylation, between the BRIC 222 F(ab)- and BRIC 222-stimulated conditions are circled. White circles indicate higher phosphorylation in BRIC 222-stimulated samples; blue circles indicate higher phosphorylation in BRIC 222 F(ab)-stimulated samples. b Top phosphorylated candidates, numbered in (a) with their respective phospho-volume. Adjusted phospho-ratio indicates phosphorylation ratio adjusted based on protein expression. The candidate proteins were identified by mass spectrometry, with the associated gene indicated in the parentheses.

Fig. 4. CD44 cross-linking alters the accessibility of Basigin and other RBC membrane proteins.

RBCs were stimulated with BRIC 222 or BRIC 222 F(ab) at 25 µg/ml and then assessed for surface protein changes. a Representative gating for RBC surface Basigin (BSG) detected with monoclonal anti-BSG antibody TRA-1-85, indicating BSG^high population. b–e Frequencies of BSG^high (b), GYPA^high (c), GYPC^high (d), and CD47^high (e) as well as the respective MFI of five distinct RBC donors are plotted. For each graph, a color dot represents an individual donor. Statistical analysis: Friedman test with FDR correction; * p ≤ 0.05, ** p ≤ 0.01, ns non-significant.

Fig. 5. CD44 cross-linking promotes P. falciparum PCRCR binding.

a Left: Representative images from immunofluorescence assay for Basigin (BSG) and CD44 on RBCs from an individual donor. Pearson correlation coefficient for co-localization (r) is indicated. Right: Pearson correlation coefficients for co-localization of BSG and CD44 on individual cells from three distinct donors are plotted; each dot represents a single RBC; error bars indicate SEM. b RBCs were stimulated with the indicated anti-CD44 antibodies (25 µg/ml) and then surface BSG was detected by flow cytometry with anti-BSG polyclonal antibody. Each dot represents an individual donor, and frequency of BSG^high (left) and MFI (right) are plotted. The average BSG^high or MFI ± SD for five donors is indicated. Statistical analysis: Friedman test with FDR correction; * p ≤ 0.05. c RBCs were stimulated with the indicated anti-CD44 antibodies (250 µg/ml) and then detected through surface staining of anti-BSG monoclonal antibody TRA-1-85. Frequency of BSG^high population (left) and MFI (right) are plotted. Each dot represents one donor. N = 6 donors for all conditions except for BRIC 235 (N = 4). Mean ± SD is indicated. Statistical analysis: Kruskal-Wallis with FDR correction; * p ≤ 0.05, *** p ≤ 0.001. d PCRCR binding to RBCs was assessed following stimulation with 25 µg/ml BRIC 222. PCRCR binding was detected using anti-PfRH5 (R5.011). Flow cytometry gating strategy for a representative donor is shown. Percentage of PCRCR binding is boxed and indicated as A488 + . e PCRCR binding results from six donors, comparing the unstimulated and BRIC 222-stimulated conditions, are plotted. Each color dot represents an individual donor. Statistical analysis: Wilcoxon matched-pairs signed rank test, one-tailed; * p = 0.0156.

Fig. 6. CD44 cross-linking by P. falciparum EBA-175 increases PCRCR binding to RBCs.

a–d RBCs were stimulated with 2 µM RII EBA-175. (a) Basigin (BSG) was detected with the anti-BSG monoclonal antibody TRA-1-85. Frequency of BSG^high population (left) and MFI (right) of five donors are shown. Frequencies of GYPA^high (b), GYPC^high (c), CD47^high (d) and the respective MFI of five donors are plotted. Each color dot represents an individual donor. Statistical analysis: Wilcoxon matched-pairs signed rank test, one-tailed; * p ≤ 0.05. e PCRCR binding of RBCs was assessed following 1.5 µM RII EBA-175 stimulation. Representative gating for one donor is shown. Percentage of PCRCR binding is boxed and indicated as A488 + . f PCRCR binding of five donors, comparing unstimulated and EBA-175-stimulated conditions, is plotted. Each color dot represents an individual donor. Wilcoxon matched-pairs signed rank test, one-tailed; * p = 0.0312. g PfRH5 binding on RBCs was assessed following stimulation with 2 µM RII EBA-175. PfRH5 was detected using an anti-PfRH5 antibody conjugated with A488 fluorophore (5A9-488). Representative gating for one donor is shown, with percentage of PfRH5 binding boxed and indicated as A488 + . h Percentage of PfRH5 binding compared to unstimulated condition, across six donors, is plotted. Each color dot represents an individual donor. Statistical analysis: Wilcoxon matched-pairs signed rank test, one-tailed; * p = 0.0312.