Analysis of the diverse antigenic landscape of the malaria protein RH5 identifies a potent vaccine-induced human public antibody clonotype

Abstract

The highly conserved and essential Plasmodium falciparum reticulocyte-binding protein homolog 5 (PfRH5) has emerged as the leading target for vaccines against the disease-causing blood stage of malaria. However, the features of the human vaccine-induced antibody response that confer highly potent inhibition of malaria parasite invasion into red blood cells are not well defined. Here, we characterize 236 human IgG monoclonal antibodies, derived from 15 donors, induced by the most advanced PfRH5 vaccine. We define the antigenic landscape of this molecule and establish that epitope specificity, antibody association rate, and intra-PfRH5 antibody interactions are key determinants of functional anti-parasitic potency. In addition, we identify a germline IgG gene combination that results in an exceptionally potent class of antibody and demonstrate its prophylactic potential to protect against P. falciparum parasite challenge in vivo. This comprehensive dataset provides a framework to guide rational design of next-generation vaccines and prophylactic antibodies to protect against blood-stage malaria.

Keywords: PfRH5; Plasmodium falciparum; RH5; antibody; blood stage; mAb; malaria; monoclonal antibody; public clonotype; vaccine.

Figure 1:. The functional epitope landscape of PfRH5.

(A) Community network plot illustrating the competitive relationship between 206 vaccine-induced anti-PfRH5 human mAbs. Supercommunities and communities are defined by number code and color. Individual mAbs are represented as nodes. Solid lines between nodes indicate bidirectional competition. Dashed lines between nodes indicate unidirectional competition. Square nodes indicate mAbs that were excluded as either a ligand or an analyte. Community 6 (representing the IDL binders; N=7) was analysed separately and is shown as an inset. (B) Community network plot overlaid with blocking category for PfRH5 binding to basigin or PfCyRPA as defined by BLI, or (C) GIA % as tested using a high concentration (0.8–2 mg/mL) of each mAb. (D) Violin plots showing the GIA potency of each epitope community as measured by the effective concentration (EC) needed to reach 30%, 50% or 80% GIA. Data are log transformed and lines indicate the median and quartiles. Dashed line indicates the previously reported best-in-class sentinel mAb, R5.016, with GIA EC₅₀ against 3D7 clone P. falciparum = 20.7 μg/mL (Data S1C). Weak or non-active mAbs, for which EC values could not be determined, were assigned values of 10 mg/mL for the purpose of analysis.

Figure 2:. Binding kinetics of anti-PfRH5 mAbs.

(A) Iso-affinity plot showing kinetic rate constants for binding of mAbs to RH5.1 (full-length PfRH5 protein) as determined by HT-SPR. Diagonal lines represent equal affinity (K_D) = K_off / K_on. Red vertical line indicates lowest limit of K_off measurement (6 × 10⁻⁵ s⁻¹). mAbs colored by epitope community (N=213 in total). (B) The RH5.1-binding parameters of K_on, (C) K_off and (D) K_D were correlated with GIA EC₅₀ for all antibodies in the growth inhibitory antibody epitope (super)communities 1, 2 and 3 (N=159). (E) The RH5.1-binding parameter of K_on or (H) K_D was correlated with GIA EC₅₀ for all antibodies in the growth inhibitory antibody epitope supercommunity 1 (N=83), (F,I) community 2 (N=44) and (G,J) supercommunity 3 (N=32). Anti-PfRH5 clones R5.129 and R5.036 are circled in panels E and F, respectively. Spearman’s rank correlation coefficient (ρ) and two-tailed P value are shown.

Figure 3:. Sequence analysis of anti-PfRH5 mAbs.

(A) Chord plot representing pairings of immunoglobulin (IG)HV and (IG)KV/(IG)LV gene families and (B) genes used by N=206 anti-PfRH5 human mAbs. The width of the cord is proportional to the number of mAbs which utilize that pairing. (C) Gene pairs in the anti-PfRH5 panel with N≥4 representative mAbs plotted in groups along with their GIA EC₅₀ value. Each mAb is colored by its epitope community. Boxes show the mean with minimum to maximum. Dashed line shows the R5.016 bench mark EC₅₀ of 20.7 μg/mL. (D) GIA assay titration curves of mAbs utilizing the (IG)HV3–7/(IG)LV1–36 gene combination and sentinel mAb R5.016 for comparison. Data were combined from repeat assays: N=14 for R5.034; N=4 for R5.237; N=3 for R5.237; N=3 for R5.270; and N=6 for R5.016. Data were log transformed and a four parameter nonlinear regression was fitted. Individual points are from the all replicate titrations. The shaded regions show the 95% confidence limits of the curves. Dashed line shown at 50% GIA and dotted lines at EC₅₀ readouts for the R5.034 and R5.016 curves. (E) For each mAb utilizing the (IG)HV3–7/(IG)LV1–36 public clonotype gene combination (color-coded as in D) a panel of four germline revertant antibodies was designed. Each mAb was tested at a concentration of 0.5 mg/mL in the GIA assay against 3D7 clone P. falciparum parasites. Points shown the mean of triplicate test wells and connecting lines are shown for clarity. WT = wild-type mAb sequence; (IG)HV3–7_GL has all mutations up to the beginning of the CDRH3 sequence mutated to germline combined with WT light chain; (IG)LV1–36_GL has all mutations reverted to germline, including the CDRL3 and J-region, combined with WT heavy chain; (IG)HV3–7_GL/(IG)LV1–36_GL is mAb with both of these heavy and light chain sequences; (IG)HV3–7_KO is the (IG)HV3–7 WT sequence for each respective mAb but with the CDRH3 mutated to a random sequence of 13 amino acids, combined with WT light chain.

Figure 4:. The antigenic landscape of PfRH5.

(A) Crystal structure of PfRH5 using RH5ΔNL protein (grey) bound to R5.251 Fab fragment (violet). (B) Close-up view of the PfRH5 (grey) and R5.251 (violet) binding interface. Complementarity determining regions (CDRs) are highlighted in magenta, and labelled with their identifier, according to IMGT annotation. (C) Structure of human basigin (CD147, yellow) (PDB: 4UOQ) aligned to the structure of PfRH5 (grey) in complex with the heavy (violet) and light (pink) chains of R5.251 Fab. (D) Structure of PfRH5 (grey) and R5.251 Fab (violet, community 3a) aligned to the structures of Fabs R5.004 (blue, community 1a, PDB: 6RCU); R5.016 (red, community 2, PDB: 6RCU); QA1 (pale blue, community 1a, PDB: 4U1G) and 9AD4 (pale red, community 2, PDB: 4U0R); and to basigin (yellow, PDB: 7CKR). (E) Structure of PfRH5 (grey, PDB: 4WAT) colored by the interface residues with Fabs R5.251 (violet, community 3a); R5.004 (blue, community 1a, PDB: 6RCU); R5.011 (green, community 4b, PDB: 6RCV); R5.015 (orange, community 5b, PDB: 7PHU); R5.016 (red, community 2, PDB: 6RCU); QA1 (pale blue, community 1a, PDB: 4U1G) and 9AD4 (pale red, community 2, PDB: 4U0R). Interfacing residues used by two or more different communities are highlighted in magenta. Basigin (PDB: 4U0R) and CyRPA (PDB: 6MPV) are shown as silhouettes. The leftmost and rightmost images are flipped 180° relative to one another. The centre images is a top-down view, centred on the apex of PfRH5; PfCyRPA has been omitted from this view.

Figure 5:. Assessment of intra-PfRH5 antibody synergy.

(A) Growth-inhibitory/neutralizing antibodies (nAbs) were tested at a final concentration equivalent to their EC₅₀ value and non-nAbs were tested at a final concentration of 0.3 mg/mL. The predicted Bliss additivity % GIA was subtracted from the measured % GIA of the test antibody combination, and the difference is plotted as a percentage in the heatmap. Thresholds were used to categorize combinations as synergistic (≥10%; red), additive (grey), or antagonistic (≥–10%; blue). Test mAbs are annotated with their epitope community assignment. (B) As for panel A except pairs of non-nAbs or mAbs with minimal GIA were combined. Antibodies were tested at a final concentration of 0.2 mg/mL each. (C) GIA assay dilution curve of R5.028 (community 3b) in combination with R5.246 (community 4b). Antibodies were combined in an equal ratio and run in a 5-fold dilution curve starting from ^~1 mg/mL. Data were log transformed and a four-parameter non-linear regression was plotted. The predicted Bliss additivity % GIA of the mixture is shown as a dashed curve. A dotted line is shown at 50% GIA for reference. (D) GIA assay dilution curves of the community 1a nAb R5.077 run in a 5-fold dilution starting from ^~0.5 mg/mL under various test conditions. Data were log transformed and a four-parameter non-linear regression was plotted. For each curve, a non-nAb (R5.028 from community 3b or R5.246 from community 4b), or a combination of both non-nAbs was added at a fixed concentration of 0.2 mg/mL each. Predicted Bliss additivity GIA curves for each combination are shown as a dashed line. The black dotted line indicates 80% GIA. The red dotted line indicates the level of GIA measured alone for the fixed concentration combination of the two non-nAbs (R5.028+R5.246). (E) Same assay set-up as in D except using R5.034 (community 2) in place of R5.077.

Figure 6:. Structure of R5.034 and efficacy against P. falciparum challenge.

(A) Crystal structure of PfRH5 using RH5ΔNL protein (grey) in complex with R5.034 Fv region (maroon). The image is of a top-down view, tilted 10° along the x-axis to view the binding interface on PfRH5 as predicted by PDBePISA, which is coloured in pink. Common PfRH5 polymorphisms are colored white and annotated. CDR loops in the R5.034 structure are labelled with their IMGT identifier. Residues in the IGHV and IGLV regions which are mutated from germline are highlighted in magenta. (B) Structure of PfRH5 (grey) and R5.034 modelled Fab (maroon) aligned with the community 2 Fabs of 9AD4 (pale red, PDB: 4U0R) and R5.016 (red, PDB: 6RCU) and the basigin ectodomain with transmembrane helix (yellow, PDB: 7CKR). An AlphaFold predictive model^, of R5.034 Fab was aligned with the experimentally observed R5.034 Fv structure to hypothesize the spatial arrangement of the three antibodies. (C) Steady-state affinity, as assessed using SPR, of the R5.034 and R5.034LS mAb binding to human FcRn at pH 6.0. Sensorgrams are shown of a 9-step dilution curve beginning at the indicated concentration of mAb. Calculation of steady-state affinity (K_D) at pH 6.0 is shown in Figure S6B. (D) Titration of the R5.034 and R5.034LS mAbs in the assay of GIA against 3D7 clone P. falciparum parasites. Dots show the mean and error bars the range of N=3 triplicate test wells per test mAb concentration. Non-linear regression curve is shown. (E) FRG huHep mice were each exposed to the equivalent of five bites using mosquitos infected with N54 strain P. falciparum. On day 5 post-infection, mice were administered 675 μg control mAb or 625 μg R5.034, as well as human RBC, via the intravenous route. Administration of more human RBC was repeated on days 6, 9 and 11. Parasitemia in the blood was monitored on days 7, 9, 11 and 13 by quantitative RT-PCR. Data from individual mice are shown combined from two independent experiments (Control N=8; R5.034-treated N=13). Dashed line is the lower limit of detection of the qRT-PCR assay at 5 parasites per mL blood.