mRNA vaccines targeting malaria protein pieces protect mice from infection

We constructed immunogens displaying subunits from PfCSP strain 3D7 (Accession No. AL844502), the strain used in RTS,S/AS01, as shown in Fig. 1a. First, we fused (NANP)₆ as a 24-residue peptide to the N-terminus of the nanoparticle protomers as a control for the NANP epitope of RTS,S in the context of our nanoparticles; we matched the number of repeat motifs found in the junctional region (NANPNVDP)₃ as a higher number of repeat motifs interfered with nanoparticle formation. We displayed three additional peptides on nanoparticles to encompass the junctional epitope with the sequences beginning with Region I, the cleavage site important for parasite invasion²⁸. We constructed a minimal junction peptide termed J2, a full-length junctional peptide termed J3, and a peptide termed J3R3 that added three NANP repeats onto the junctional peptide to combine junctional and NANP repeat epitopes. To focus the C-terminal immune response to the conserved epitope region, we used Rosetta to design a modified αTSR termed MD1 that minimized binding of known strain-specific antibodies without impacting antibody binding to the conserved epitope (Supplemental Fig. 1). We displayed both wild-type (WT) αTSR and MD1 αTSR on nanoparticles. We did not include the C-terminal linker, because no protective antibodies have been identified for this region^52,53. We were unable to express the full N-terminal domain on nanoparticles. Instead, novel linear epitopes were identified through binding of novel antibodies to overlapping peptide sequences⁴⁵, and these peptides, termed P1, P2, P7, P8, and P15, were selected for nanoparticle display as a representative set of N-terminal domain immunogenic epitopes. Nanoparticle formation was confirmed by size exclusion chromatography (SEC), SEC with multi-angle light scattering (SEC-MALS), and negative stain electron microscopy (nsEM) (Supplemental Figs. 2–10). Prospective epitopes for mRNA development were displayed on membrane-bound platforms for FACS characterization before production.

We evaluated antibody binding responses using sera collected one week after the second vaccination (Week 5). For mice vaccinated with nanoparticles displaying peptides J2, J3, J3R3, and (NANP)₆, we tested binding to a junctional peptide corresponding to J2. We found that 60mers elicited a stronger response than 24mers aside from J3 which merely trended higher (Fig. 3b). We also found that J3R3-60mer elicited the highest response to junctional epitopes compared to all immunogens targeting PfCSP repeats (Fig. 3b). For the C-terminal immunogens, we measured the serological response to the C-terminal αTSR domain captured on ELISA plates by either anti-alpha site antibody 236 or anti-beta site antibody 1512 thus binning the sera into either anti-alpha site response or anti-beta site response. We found that the WT αTSR immunogen elicited antibodies that bound to both alpha and beta C-terminal regions, while the MD1 αTSR immunogen elicited antibodies that bound to only the beta C-terminal region (Fig. 3c). To evaluate the effect of combined immunization with J3R3 and C-terminal immunogens, we compared the serum binding responses from mice immunized with single immunogens to those immunized with the combination. We found that the epitope-specific responses to the combination immunization were indistinguishable from immunization with the individual components (Supplemental Fig. 12a). Finally, we evaluated responses to P1, P2, P8, P9, and P15 immunogens except for P15 60mer, which did not form nanoparticles. We found P1 and P2 60mer but not 24mer sera showed detectable binding to full length N-terminal domain; all P8 and P9 sera showed detectable binding; and P15 24mer did not show detectable binding (Fig. 3d). For all protein candidates, we observed that repeat epitopes and C-terminal immunogens elicited stronger serological responses to their respective epitopes than N-terminal immunogens elicited for their respective epitopes (Fig. 3).

For all junctional and NANP repeat immunogens, we found that the 60mers tended to provide larger reduction in liver burden than the 24mers, consistent with the serological responses shown in Fig. 3. Immunization with (NANP)₆ peptide 60mer provided 60% reduction in liver burden, and short junctional peptide (J2) 60mer provided 50% reduction in liver burden (Fig. 4a). The full-length junctional epitope (J3) 60mer demonstrated a larger reduction in liver burden at 80%, showing that this epitope contributes more than the traditionally studied NANP repeats. A 60mer combining J3 with three repeats of NANP (J3R3), mimicking the native composition of CSP, resulted in 90% reduction in liver burden (Fig. 4b). In a separate experiment, a head-to-head comparison of the best-performing immunogen J3R3-60mer/SMNP and RTS,S/AS01 was carried out in an independent laboratory. J3R3-60mer/SMNP showed comparable reduction in liver burden (median ± interquartile range, 71 ± 13%) to that for RTS,S/AS01 (82 ± 10%) and monomer CSP (83 ± 7%), despite having a substantially shorter epitope length (Fig. 4c).

In contrast, the C-terminal domain of CSP, whether wild-type or modified, provided a small reduction in liver burden, reaching a maximum of 36% (Fig. 4d, e). Combined immunization with C-term WT and J3R3 immunogens, approximating the possible combination of repeat and C-terminal epitopes on native CSP, showed reduction no higher than J3R3 alone (Supplemental Fig. 12b). Immunogens presenting peptides from the N-terminal domain, including peptides P1, P2, P8, P9, and P15, exhibited varying degrees of reduction in liver burden ranging from 27% to 60% (Fig. 4f).

Immunization with "naked" nanoparticles lacking malaria epitopes (CSP-negative nanoparticles) conferred nonsignificant levels of reduction in liver burden in this assay (Fig. 4g), suggesting that the base nanoparticle itself made minimal contributions to liver burden reduction for nanoparticles displaying CSP epitopes. However, despite the lack of significance, vaccination with CSP-negative nanoparticles trended toward lower flux levels compared to the adjuvant-only group, implying a potential impact on sporozoite fitness and liver invasion (Fig. 4g). This effect is unlikely to be driven by a humoral IgG response, as no circulating CSP-specific IgG titers were detected in these animals, as expected from the absence of CSP in the immunogen (Supplemental Fig. 13). All immunogens delivered as protein nanoparticles, despite varying levels in liver burden reduction, failed to provide sterilizing immunity, as all challenged mice, aside from those in groups receiving sterilizing mAbs, developed parasitemia. Therefore, we explored alternate methods of immunogen delivery.

We next immunized with protein monomer CSP, J3R3-24mer mRNA-LNP, J3R3-GPI mRNA-LNP, J3R3-60mer mRNA-LNPs, and J3R3-TM mRNA-LNP using the same schedule but with lower parasite challenge with 300 sporozoites. Small bleeds were taken from the mice following challenge, and red blood cells were monitored for parasitemia by flow cytometry (Fig. 5d, Supplemental Fig. 14). At 42 hours post infection, a majority of the immunogens exhibited levels of liver invasion inhibition like that of the mAb317 sterilizing control (Supplemental Fig. 15). At day three, mice in the adjuvant-only group begin to show parasitemia followed shortly by mice in other immunogen groups. At day six, all mice in the adjuvant-only group showed increased parasitemia, along with mice in all groups except those in the noninfected negative control group and the mAb 317-sterilized group. At day eight, all mice that showed parasitemia previously had increased parasite burden, while all mice that had shown no infection remained uninfected. At day nine, all mice in the adjuvant-only group exhibited sufficient malaria symptoms to require humane euthanasia as determined in the animal use protocol (Fig. 5e). Survival was monitored, and mice that exhibited no symptoms by day fourteen were considered to have to have received sterilizing protection from their respective immunogen (Fig. 5e). Mice passively immunized with 300 µg of mAb 317 remained parasite free and survived for the duration of the study (Fig. 5d, e). All mRNA-LNP immunogen groups except J3R3-TM showed significant drops in survival, despite a majority of the immunogens providing 97% or greater of inhibition of liver invasion (Fig. 5e).

We then tested lower doses (3 µg and 1 µg) of J3R3-60mer mRNA and J3R3-TM mRNA using the standard challenge model (Supplemental Fig.). 14G2-equivalent titers were measured from week six sera for each vaccine group. Increasing dosages of both 60mer mRNA and mRNA-TM resulted in corresponding increases in 14G2 equivalence, with maximum average titers of 2509 µg/mL and 2363 µg/mL, respectively (Supplemental Fig.). Significant differences in liver burden reduction were observed between the 1 µg, 3 µg, and 10 µg groups, as evidenced by liver burden measurements and day six parasitemia levels (Supplemental Fig.). Dose-dependent reduction in liver burden in the J3R3-60mer mRNA groups was most apparent in the survival analysis (Supplemental Fig.). Overall, the trends observed in liver burden, parasitemia, and survival analyses suggested dose-dependent reductions in liver burden from mRNA-VSV-TM that peaked at the maximum tested dose of 10 µg. Dose-dependent liver burden reduction for J3R3-60mer was less clear, though the data suggested that doses higher than 1 µg confer increased effect. 16 16a 16b, c 16d

Subsequently, mRNA-delivered J3R3-60mer and J3R3-TM were directly compared to RTS,S/AS01_E in an independent laboratory, using the same immunization schedule and a modified challenge protocol utilizing 2000 sporozoites per mouse. CSP-specific IgG titers were quantified by interpolation using the murine NANP-specific mAb 2A10 in an ELISA format. Both J3R3-60mer mRNA-LNP and J3R3-TM mRNA-LNP elicited higher NANP-specific IgG titers, with values of 101 µg/mL ± 53 and 272 µg/mL ± 53.4, respectively, compared to RTS,S, which averaged 82 µg/mL ± 61, but the difference was significant only for J3R3-TM mRNA-LNP (p = 0.0012) (Fig. 5f). Immunization with J3R3-TM mRNA-LNP, J3R3-60mer mRNA-LNP, or RTS,S resulted in inhibition of liver invasion, with similar median inhibition values of 92%, 87%, and 79%, respectively (Fig. 5g). Given the lower inhibition of liver invasion of RTS,S, it is unlikely that it would provide sterilizing protection and significant survival to these animals like we show with J3R3-TM mRNA-LNP in Fig. 5e. This is in line with what has been shown elsewhere with two doses of RTS,S at 5 µg per dose⁵⁹.

All genes were synthesized at Genscript, Inc. Nanoparticles were cloned into pHLsec between the leader and a double stop coding using the AgeI and KpnI cloning sites. Antibody heavy chains were cloned into pCW-CHIg-hG1 between the leader and IgG1 human constant domain using EcoRI and NheI cloning sites. Double stop codons were inserted after the CH₁ region for Fab production. Kappa chains were cloned into pCW-CLIg-hk between the leader and human kappa constant region using the EcoRI and BsiWI. Lambda chains were cloned into pCW-CLIg-hL2 between the leader sequence and the human lambda constant region using the EcoRI and AvrII cloning sites.

We made immunogens displayed on modified self-assembling nanoparticles ferritin and lumazine synthase. We modified H. pylori ferritin (PDB:3bve) to remove a problematic naturally occurring glycosylation site at position 18 and engineered four additional glycosylation sites introduced at positions 68, 78, 95, and 102. Peptides were linked directly without a GGS linker to utilize proline in orienting the peptide away from the particle. We modified A. aeolicus lumazine synthase (PDB:1hqk) to mutate away an unpaired cysteine at position 37 and a buried glycosylation site at position 102. Four cysteines were introduced to design two extra disulfide bridges. Three mutations were introduced to knock out the active sites of the nanoparticle. And finally, one extra glycosylation site was introduced at position 71 and two glycosylation sites were added to the C-terminus.

DNA was transfected with a plasmid into FreeStyle 293 F cells (Invitrogen, Cat no. R79007) using 293Fectin (ThermoFisher), and proteins were expressed at 37 ˚C for four days. NPs were purified using snowdrop lectin-conjugated agarose beads (Vector laboratories), followed by gel-filtration using a Superose 6 size-exclusion chromatography column (GE Healthcare). NP-assembly was assessed by SEC + multiangle light scattering (SEC-MALS) using a Superose 6 10/300 column (GE Healthcare) at a flow rate of 0.5 ml/min followed by DAWN HELEOS II and Optilab T-rEX detectors (Wyatt Technology), correcting for the glycan molecular mass by applying the built-in protein-conjugate analysis (ASTRA).

Full-length monomeric protein PfCSP was produced in Lactococcus lactis as described previously⁵⁷.

Paired HC and LC Fab variable region sequences were gene-synthesized and inserted into human Fab HC constant region expressing vector pFabCW and human lambda or kappa expressing vectors pCW-CLig-hL2 or pCW-CLig-hk, respectively. Fabs were expressed in 500 mL FreeStyle™ 293 F cell cultures or 30 mL ExpiCHO™ cell cultures (Thermo Fisher Scientific, Cat# A29133). For 293 F cell transfection, 300 μg of HC and 150 μg of LC plasmids were mixed with 225 μg polyethylenimine (PEI; 1:3 DNA:PEI ratio) in 5 mL of Opti-MEM™ reduced serum medium (Thermo Fisher Scientific, Cat# 31985070) for 30 min, then added to 293 F cells. Supernatant was collected after 5–6 days. ExpiCHO™ cell cultures were transfected according to manufacturer instructions, using 12.5 μg HC and 31.2 μg LC plasmids. Supernatant was collected 8 days post-transfection. Harvested supernatants were filtered through 0.45 or 0.25 μm membrane filters and batch bound to CaptureSelect CH1-XL Affinity resin (Thermo Fisher Scientific, Cat# 1943462005). Resin was washed with PBS, and captured Fabs were eluted with 50 mM NaOAc pH 4.0, buffer exchanged into 1× PBS, and concentrated using a 30k MWCO concentrator.

For expression of IgG the HC variable region was cloned into the pCW-CHIg-hG1 vector. Transfection was carried out as described above, but batch binding occurred overnight at 4 °C to Protein A resin (Thermo Fisher Scientific, Cat# 20334) while on a rocker. Unbound supernatant was allowed to flow through, and the resin was washed with PBS until protein A280 reading of the flowthrough measured by a nanodrop reached background levels. Protein A bound IgG was eluted with 0.1 M Glycine pH 2.7. Eluted mAbs were buffer exchanged into 1× PBS and concentrated using a 50 K MWCO concentrator (Millipore).

Analysis for possible glycosylation sites was done using GlycanTreeModeler as described previously⁷⁶. Rosetta fixed backbone, relax, and scoring were used to prescreen core mutations to find suitable mutations.

mRNA-LNPs for J3R3-60mer, J3R3-24mer, J3R3-TM, and J3R3-GPI were synthesized in vitro and passed quality control tests as previously described in ref. ⁷⁷. The secretion signal sequence for human IL-2 (MYRMQLLSCIALSLALVTNS) was used for J3R3-TM and J3R3-GPI mRNA constructs, and the secretion signal sequence from the pHLsec vector (MGILPSPGMPALLSLVSLLSVLLMGCVAETG)⁷⁸ was used for J3R3-60mer and J3R3-24mer mRNA constructs.

ELISA buffers were made as follows: washing buffer (PBS + 0.2% tween 20), blocking buffer (5% skim milk in PBS with 1% FBS and 0.2% tween 20), dilution buffer (PBS with 1% FBS and 0.2% tween 20), and stop solution (340 mL of dH₂O with 10 mL of sulfuric acid). 96-well Corning costar assay plates were coated with 25 μl of antigen or 236 and 1512 Fab at 2 μg/ml in PBS overnight at 4 °C. After each incubation, plates were washed three times. Blocking buffer was added, incubating at room temperature for one hour. Sera were diluted 3-fold from a starting dilution of 1:500 and incubated for one hour. Antibodies were diluted 4-fold from a starting concentration of 10 μg/mL and incubated for one hour. 25 μl of peroxidase secondary antibody (Peroxidase AffiniPure Goat Anti-Mouse IgG H + L, Jackson ImmunoResearch or Peroxidase Donkey Anti-Human IgG, Jackson ImmunoResearch) was added for one hour incubation. A 1:5 dilution of TMB was added to each plate at 25 μL per well. Plates were left to stand for approximately four minutes before 25 μL per well of stop solution was added. A plate reader read at 450 nm and 570 nm, with the delta between the two readings measured and analyzed. Area under the curve (AUC) was calculated using Prism 9.1.0. Equivalence titers were determined by interpolation using standard curves of murine antibodies 14G2 and 2A10.

Membrane-tethered constructs were produced in Freestyle 293-F cells transfected with the indicated DNAs according to manufacturer's instructions (ThermoFisher Scientific). Two days after transfection, cells were stained with the indicated antibodies 10 nM in FACS buffer (PBS + 0.1% BSA), washed twice, labelled with Alexa Fluor 647-conjugated anti-human IgG (Jackson Immunoresearch) and analyzed on a NovoCyte 3000 (ACEA Biosciences). Data were analyzed by gating on single cells using FSC, SSC, and width in FlowJo 10.7.1 (Beckton Dickinson). The median fluorescence intensity of the Alexa-647 signal was calculated and plotted in Prism 9 for macOS (Graphpad).

Purified samples were diluted to 0.1 mg/ml in HBS and mixed with uranyl formate for negative-stain EM using a Talos TEM. Carbon-coated copper grids (400 mesh) were glow-discharged and 10 µL of each sample was adsorbed for 2 min. Excess sample was wicked away and grids were negatively stained with 2% uranyl formate for 2 min. Excess stain was wicked away, and the grids were allowed to dry. Samples were analyzed at 80 kV with aD ThermoFisher Talos L120C transmission electron microscope, and images were acquired with a CETA 16 M CMOS camera.

Our standard two-dose schedule utilized female C57BL/6 J mice that were 7–8 weeks old at the time of the first immunization. Mice were immunized at week 0 and week 5 with 0.5 nanomoles of immunogen and 5 μg of SMNP adjuvant, SQ scruff of the neck. Mice were bled one week before the first immunization and one week after the second immunization for 100 μl of blood.

The only study using female C57BL/6 J mice that differed from the standard was the study utilizing RTS,S/AS01. Here, mice were immunized with 15 μg of our immunogens and 5 μg of SMNP adjuvant, SQ base of the tail. One group received RTS,S/AS01 (5 µg RTS,S adjuvanted with 5 µg of AS01_E) via intramuscular route.

We observed no difference between administering malaria immunogens and controls SQ scruff of neck and SQ base of tail.

10 µg, 3 µg, or 1 µg of formulated mRNA/LNP was gently mixed with PBS to bring up to a total volume of 50 µl. mRNA/LNP i.m. injections were done in the left leg.

For standard challenge, at either two weeks or five weeks after last immunization, all mice were anesthetized and challenged with 5000 P. Berghei transgenic sporozoites expressing P. falciparum CSP and luciferase through IV tail injection. Forty-two hours after challenge, mice were anesthetized and injected IP with 100 μL of D-luciferin (30 mg/mL), having been anesthetized by exposure to isoflurane. Bioluminescence in the liver was measured using an AMI HTX (Spectral Instruments). Percent inhibition of liver invasion was calculated with the following formula: Percent Inhibition=100-100*(experimental flux/adjuvant flux).

Parasitemia protocols were adapted from previously described methods⁷⁹. Briefly, animals were anesthetized, and small bleeds were taken from mice after challenge. 1 µL of whole blood was added to PBS containing 1 μg/ml Hoechst 33258 (Thermo Fisher) and 2 µg/mL of anti-mouse CD45.2 conjugated to Alexa-647 (BioLegend) and analyzed on a ZE5 Cell Analyzer (BioRad). Live/dead stains were not used in these experiments; therefore, gated populations may contain dead cells. Data were analyzed by gating on single cells using FSC, SSC, and width in FlowJo 10.7.1 (Beckton Dickinson), followed by gating on Alexa 647 negative cells. The percentage of the GFP signal was calculated and plotted in Prism 10.1.0 for macOS (Graphpad).

At the conclusion of each experiment, mice were euthanized through CO₂ exposure in accordance with approved guidelines. The research protocol was approved and performed in accordance with Scripps Research IACUC Protocol #22-0007.

Data were analyzed and plotted using Prism 10.1.0 (Graphpad). Medians and interquartile ranges were shown when appropriate. All measurements were taken using discrete samples. Statistical analysis was performed using Prism software using two-tailed Mann–Whitney assuming non-normal distribution, Kruskal-Wallis test assuming non-normal distribution, or log-rank Mantel-Cox test for survival curves. P-values below 0.05 (*), 0.005 (**), and 0.0005 (***) were considered significant and indicated by asterisks.

Figure 1 schematics adapted from "Membrane Proteins", by BioRender.com (2023). Retrieved from biorender-templates↗.

All the experimental animal work was performed in strict compliance to the guidelines of Scripps Institutional Animal Care and Use Committee, which approved this study (under animal use protocol authorization 15-0003-4 and 22-0007-1).