Exploring in vitro expression and immune potency in mice using mRNA encoding the Plasmodium falciparum malaria antigen, CelTOS

Abstract

The secreted malarial protein, Cell-Traversal protein for Ookinetes and Sporozoites (CelTOS), is highly conserved among Plasmodium species, and plays a role in the invasion of mosquito midgut cells and hepatocytes in the vertebrate host. CelTOS was identified as a potential protective antigen based on a proteomic analysis, which showed that CelTOS stimulated significant effector T cells producing IFN-γ in peripheral blood mononuclear cells (PBMCs) from radiation attenuated sporozoite-immunized, malaria-naïve human subjects. In a rodent malaria model, recombinant full-length CelTOS protein/adjuvant combinations induced sterile protection, and in several studies, functional antibodies were produced that had hepatocyte invasion inhibition and transmission-blocking activities. Despite some encouraging results, vaccine approaches using CelTOS will require improvement before it can be considered as an effective vaccine candidate. Here, we report on the use of mRNA vaccine technology to induce humoral and cell-mediated immune responses using this antigen. Several pfceltos encoding mRNA transcripts were assessed for the impact on protein translation levels in vitro. Protein coding sequences included those to evaluate the effects of signal sequence, N-glycosylation on translation, and of nucleoside substitutions. Using in vitro transfection experiments as a pre-screen, we assessed the quality of the expressed CelTOS target relative to the homogeneity, cellular localization, and durability of expression levels. Optimized mRNA transcripts, which demonstrated highest protein expression levels in vitro were selected for encapsulation in lipid nanoparticles (LNP) and used to immunize mice to assess for both humoral and cellular cytokine responses. Our findings indicate that mRNA transcripts encoding pfceltos while potent for inducing antigen-specific cellular cytokine responses in mice, were less able to mount PfCelTOS-specific antibody responses using a two-dose regimen. An additional booster dose was needed to overcome low seroconversion rates in mice. With respect to antibody fine specificities, N-glycosylation site mutated immunogens yielded lower immune responses, particularly to the N-terminus of the molecule. While it remains unclear the impact on CelTOS antigen as immunogen, this study highlights the need to optimize antigen design for vaccine development.

Keywords: CelTOS; N-glycosylation; Plasmodium falciparum; lipid nanoparticles; mRNA; malaria; malaria vaccine; nucleoside modifications.

Figure 1

In vitro translation of pfceltos mRNAs with different signal sequences. (A) CHO cells were transfected with pfceltos mRNAs, having identical non-coding and coding sequences except for the signal sequences, and harvested at 8-, 24-, and 48-hours post-transfection. Western blot analysis was performed to assess for in vitro translation and translocation of PfCelTOS. The right panel outlines each lane and provides a detailed description of the test samples. Semi-quantitative analysis of the PfCelTOS protein levels at the different time points was performed by using ImageJ for the (B) supernatant and (C) cell lysate relative to the 10ng recombinant PfCelTOS protein (Lane 1, for each panel).

Figure 2

Effect of N-glycosylation on in vitro and in vivo responses. (A) Western blot analysis of CHO cells transfected with mRNAs pfceltos ARCA, pfceltos Cap1 and pfceltos ARCA GM. pfceltos mRNA transcripts were codon harmonized (CH) for optimal expression in mice and encoded with the native falciparum celtos signal sequence (Wt-SS). Cell culture supernatants and cell lysates were harvested at 24 hours post-transfection to assess for the effect of N-glycosylation on protein translation. (B) BALB/cJ mice were immunized intramuscularly (IM) two times at a three-week interval with 10µg of pfceltos mRNA encapsulated into LNP1 (10µg LNP1), 10µg pfceltos mRNA with glycosylation site modified (GM) in LNP1 (10µg GM LNP1) or LNP1 alone (n=5 per group). PfCelTOS-specific IgG antibody concentrations (µg/mL) were quantified in sera at pre-immunization, three weeks after the primary dose and two weeks after the final dose by ELISA. Antibody concentrations are reported as the geometric mean and 95% confidence intervals. (C, D) Splenocytes were harvested and IFN-γ and IL-4 cytokines were detected by ELISpot assay. The mean number of spot-forming cells (SFC) per splenocytes were reported with standard errors of the mean (SEM). Statistical analysis was performed using Mann-Whitney test (**p<0.01).

Figure 3

Effect of mRNA dose and nucleoside modification on immune responses. BALB/cJ mice were immunized intramuscularly (IM), two times at three-week intervals, with a low dose (10µg) or a high dose (30µg) of pfceltos mRNA that was N-glycosylation site modified (GM), containing human IgE signal sequence (IgE-SS) (GM IgE-SS), and with nucleoside ψ-pseudouridine and 5’-methylcytosine substitutions (PU5MC), (GM PU5MC IgE-SS), in LNP (LNP1 or LNP3), (n=5 per group). (A) Antigen-specific IgG antibody concentrations against PfCelTOS were quantified in sera two weeks after the final dose by ELISA. Antibody concentrations are represented as the mean and standard deviation (SD). Statistical analysis was performed using an unpaired t-test (*p<0.05). (B, C) IFN-γ and IL-4 cytokines were detected by ELISpot. The mean number of spot-forming cells (SFC) per splenocytes were reported with standard errors of the mean (SEM). Statistical analysis was performed using Mann-Whitney test, (*p<0.05, **p<0.01).

Figure 4

A three-dose regime overcomes an “all-or-none” pattern for humoral immune responses. Mice were immunized intramuscularly (IM) thrice at a three-week interval with 10µg of University of Pennsylvania (UPenn) N-glycosylation site modified (GM) and encapsulated in LNP1 (UPenn GM LNP1), or without N-glycosylation modification (UPenn LNP1) that were one-methylpseudouridine (m1Ψ)-5′-triphosphate modified, and cellulose affinity purified or with TriLink pfceltos mRNA with N-glycosylation site modified (GM) in LNP1 (TriLink GM LNP1) or LNP3 (TriLink GM LNP3) that were ψ-pseudouridine and 5-methylcytosine modified (n = 15 per group; n = 10 in LNP alone groups). (A) Kinetics of PfCelTOS antibody concentrations measured by ELISA. Antibody concentrations are reported as the geometric mean and 95% confidence intervals. (B) IFN-γ cytokine responses were detected by ELISpot, (n=5 per group). The mean number of spot-forming cells (SFC) per splenocytes were reported with standard errors of the mean (SEM). Statistical analysis was performed using Mann-Whitney test, (*p<0.05, **p<0.01).

Figure 5

Fine specificity of antibody responses. Sera from mice that were immunized IM thrice at a three-week interval with 10µg mRNA pfceltos UPenn N-glycosylation site modified (GM) in LNP1 (UPenn GM LNP1), or without N-glycosylation modification (UPenn LNP1) that were one-methylpseudouridine (m1Ψ)-5′-triphosphate, and cellulose affinity purified or with TriLink pfceltos mRNA that were N-glycosylation site modified (GM) in LNP1 (TriLink GM LNP1) and were ψ-pseudouridine and 5-methylcytosine modified were characterized for antibody specificities by ELISA against (A) PfCelTOS N-term, (B) CelTOS Peptide 1-2, (C) CelTOS Peptide 2-3, (D) PfCelTOS C-term, (E) CelTOS Peptide 4 and (F) Full length PfCelTOS protein (n = 15 per group). Antibody titers are reported as the geometric mean and 95% confidence interval of the dilution required to achieve an OD₄₀₅ = 1. Statistical analysis was performed using a Mann-Whitney test (*p<0.05, **p<0.01, ***p<0.001).