Repertoire, function, and structure of serological antibodies induced by the R21/Matrix-M malaria vaccine

Abstract

The World Health Organization (WHO) recently recommended the programmatic use of the R21/Matrix-M vaccine for Plasmodium falciparum malaria prevention in children living in malaria-endemic areas. To determine its effects on humoral immunity, we conducted a proteomic analysis of polyclonal IgG antibodies directed against the NANP tetrapeptide of the circumsporozoite protein (CSP), which comprises the vaccine's core immunogen. In 10 malaria-naïve adult volunteers, R21/Matrix-M induced polarized IgG anti-NANP repertoires, heavily skewed for IGHV3-30/3-33 genes bearing minimal somatic mutation, which remained static in composition following a controlled human malaria infection challenge. Notably, these vaccine-generated antibodies cross-reacted with another protective CSP epitope, the N-terminal junction region, despite its absence from the R21 construct. NANP-specific IGHV3-30/3-33 mAbs mined from polyclonal IgG repertoires blocked sporozoite invasion in vitro and prevented parasitemia in vivo. Overall, R21/Matrix-M elicits polarized, minimally mutated, polyclonal IgG responses that can target multiple protective CSP epitopes, offering molecular insight into the serological basis for its demonstrated efficacy against P. falciparum malaria.

Figure 1.

Minimally mutated IGHV3-30/3-33 genes dominate the IgG anti-NANP ₆ plasma repertoire following three doses of R21/Matrix-M vaccine. (A) Schematic of the VAC053 and VAC065 clinical trials. Plasma and cells (PBMCs) were collected at the times indicated. Volunteers (n = 10) were vaccinated at three time points (days 0, 28, and 56). Ig-Seq analyses were performed at day 63 (VAC053) or day 84 (VAC065). (B) Relative abundance of NANP₆-reactive IgG lineages in a representative volunteer. Each bar represents a single IgG lineage identified by its respective CDR-H3. Red indicates the unique peptides for each lineage that were detected by LC-MS/MS. (C) Relative plasma antibody IGHV gene segment usage for each volunteer. IGHV gene identity by color is shown in D. (D) Average IGHV gene abundance in NANP₆-reactive IgG. Bar represents mean. (E) IGHV gene frequency, by lineage, in total peripheral B cells averaged across all volunteers. (F) IGHV gene frequency, by lineage, in NANP₆-nonreactive IgG averaged across all volunteers. (G) IGHJ gene usage in the same volunteer as B. (H) IGHV gene SHM levels for NANP₆-reactive and nonreactive plasma IgG antibodies and BCR sequences grouped by lineage. ****P < 0.0001 (two-tailed Kruskal–Wallis test). (I and J) Amino acid usage in CDR-H2 (I) position 50 and (J) position 52 of IGHV3-33 plasma antibodies separated by NANP₆ reactivity. I = isoleucine; L = leucine; V = valine; W = tryptophan. For each Ig-Seq analysis of polyclonal anti-NANP₆ IgG, the affinity purification was performed once, and the purified material was analyzed in triplicate by mass spectrometry.

Figure S1.

Profile of NANP ₆ -reactive IgG repertoires across all 10 volunteers after vaccination. (A) Number of IgG lineages per volunteer. (B) Frequency of CDR-H3 lengths for all lineages. (C) Checkerboard graphic illustrating no shared CDR-H3s (no overlap) across volunteers. (D) Dissimilarity of VH repertoires according to tSNE analysis. (E) Frequency of IGHV gene use by volunteer.

Figure 2.

Following vaccination with R21/Matrix M, NANP ₆ -reactive plasma antibody repertoires remain highly correlated before and after challenge with live sporozites. A subset of thrice-vaccinated VAC065 volunteers (n = 5) was analyzed at day 84 (pre-challenge; C₋₁) and 5 wk later at day 112 (post-challenge; C₊₃₅). (A and B) Relative abundance of NANP₆ reactive IgG lineages before and after challenge for one representative protected individual (A) and one non-protected individual (B). Each bar represents one lineage. Inset indicates relative abundance for each lineage before and after challenge (Pearson r² = 0.98 [A] and r² = 0.88 [B]). (C and D) Estimated plasma titer (C) and (D) change in titer of NANP₆-reactive IgG for volunteers before and after challenge (n = 17). Yellow markers indicate the group average. (E) Abundance of persistent antibodies before challenge (n = 5). (F) Abundance of emergent antibodies after challenge (n = 5). (G) Bray–Curtis dissimilarity index between pre- and post-challenge repertoires (n = 5). Lower values indicate higher degrees of similarity. Green indicates protection from infection; red indicates no protection. For each Ig-Seq analysis of polyclonal anti-NANP₆ IgG, the affinity purification was performed once, and the purified material was analyzed in triplicate by mass spectrometry.

Figure S2.

Static composition and high correlation between NANP ₆ pre-challenge and post-challenge IgG repertoires among VAC065 volunteers. A green outline indicates the volunteer was protected from infection; red indicates they were not protected.

Figure 3.

R21/Matrix-M elicits polyclonal IgG plasma antibodies cross-reactive with the N-terminal JR of CSP despite its absence in the vaccine construct. NANP₆ affinity chromatography depletes the plasma of JR-specific antibodies. (A) The schematic (A) illustrates the truncated CSP used in R21 (and RTS,S) vaccine constructs and the corresponding lack of the N-terminal JR. (B) Heat maps illustrate the signal loss (OD) in four VAC065 volunteers when measured by indirect ELISA using NANP₆ peptide or JR peptide (JR: KQPADGNPDPNANPNVDPN) after affinity chromatography and the depletion of NANP₆-reactive IgG from day 84 plasma at pre-challenge (C₋₁) and post-challenge (C₊₃₅) time points. (C) The bar graph illustrates Ig-Seq analysis of pre-challenge IgG in one volunteer (MA-1010).

Figure S3.

MA6 epitope specificity. MA6 Fab fragments displaying yeast were incubated with 10 µM biotinylated NANP₆ or JR peptide (Jct Pept: KQPADGNPDPNANPNVDPN) and subsequently labeled with streptavidin-conjugated Alexa-Fluor 647 to detect binding via flow cytometry.

Figure 4.

Monoclonal IgG plasma antibodies elicited by R21/Matrix-M vaccine provide sterile protection in mice. (A) Binding of mAbs to either WT 3D7 Pf (WT Pf) or transgenic Pb expressing PfCSP (Tg Pb) measured by indirect IFA. (B) Relationship between either IFA or ISI and NANP-binding EC₅₀. (C and D) C57BL/6J mice (n = 6 per group) were injected i.p. with 400 µg of mAb on day −1. The following day, mice were challenged with 5,000 Pb-PfCSP Pb sporozoites administered by s.c. injection at the base of the tail. From 5 days after sporozoite injection, mice were monitored by thin blood film for the development of blood-stage malaria. Parasitemia was calculated daily, and linear regression was used to determine the time mice reached a threshold of 1% parasitemia. Sterile protection was defined as any animal that remained parasite-free until day 9 after sporozoite administration. C and D are replicate experiments. (E) Summary of sterile protection for each mAb (grey bars represent the percent survival for each group over the two experiments).

Figure 5.

Elucidation of IGHV3-33 antibody binding modes to the NANP polypeptide. Crystal co-structure of MA6 bound to NPNA₃ peptide. (A) Overview of peptide in groove. The peptide lies within a groove formed primarily by the heavy chain, with CDR-H3 residues forming one side of the groove and CDR-H1 and -H2 residues forming the other side. (B) Specific contacts. Compared to related structures, the MA6 Trp52 sidechain is flipped such that the pyrrole ring of the indole, rather than the benzene ring, packs against Pro6 of the peptide. (C) Superposition with 1210 F_ab. MA6 and 1210 bind the peptide in an almost identical conformation (RMSD of 0.23 Å for the first 8 Cα atoms).