Antigens that group together after mRNA vaccine delivery improve antibody responses

Alphafold 2.3.1 [33] was used to predict structures for multimerized mSb constructs with foldon (3-mer), IMX313 (7-mer), and ferritin (24-mer). Each prediction run used one NVIDIA RTX A6000 GPU, 32 CPUs, and 220 GB RAM. The "-m multimer" parameter was used. For foldon and IMX313, protein reference files were provided with 3 and 7 copies of the construct sequence, respectively. For ferritin, to make the prediction computationally tractable, 24 copies of the ferritin sequence were predicted in a multimer run and, separately, 1 copy of the mSb-Ferritin construct was predicted. The highest confidence model for each prediction run was selected. For mSb-Ferritin, the alpha-carbons of the ferritin portion of the monomer were aligned to the structure of each of the multimer 24-mer ferritins using ChimeraX (version 1.5), generating the final multimerized structure.

DNA sequences for all plasmids used in this study are listed in Supplementary Data File S1. Coding sequences were codon-optimized using GenSmart (accessed 26 September 2022) (GenScript [Piscataway, NJ, USA]) and then manually edited to remove uridine residues in the wobble position of two-family codons. RNA structures were predicted by RNAFold [34] and coding sequences were modified as needed to maintain an optimal RNA structure in the 5′ UTR. All golden gate assembly parts were domesticated to remove Type IIs restriction sites and then incorporated into pNL and/or pLZ1/4. To create the pRNA series of plasmids, pRNA-Destination plasmids were combined with appropriate pLZ plasmids and assembled in a PaqCI-based golden gate assembly reaction performed using the manufacturer's protocols. In brief, reactions contained 50 ng of destination vector and equimolar amounts of pLZ, 10 U PaqCI, 400 U of T4 DNA ligase, and 5 pmoles of PaqCI activator in T4 DNA ligase buffer and were cycled 30 times for 37 °C for 5 min, then 16 °C for 5 min. All pDNA constructs were sequence-verified. All enzymes were purchased from New England Biolabs (Ipswich, MA, USA). All golden gate assembly parts were ordered from GenScript, Integrated DNA Technologies (Coralville, IA, USA), or Twist Bioscience (San Francisco, CA, USA).

NEB stable E. coli (New England Biolabs, Cat. #C3040H) were used for propagation of DNA. E. coli was transformed following manufacturer protocols. E. coli was grown in LB broth at 30 °C, shaking at 250 rpm. For solid medium, LB broth was supplemented with Bacto agar (1.5% [w/v]) (Becton Dickinson [Franklin Lakes, NJ, USA], Cat. #214010) and further supplemented with 100 μg/mL carbenicillin or 50 μg/mL kanamycin, as appropriate. For blue/white screening, media was further supplemented with 1 mM IPTG and 0.1 mg/mL of X-Gal (prepared as a 20 mg/mL stock solution in DMF). All pDNA was quantified by Qubit assay and stored at −20 °C.

pDNA template was linearized in preparation for IVT by restriction digest using BspQI (New England Biolabs, Cat. #R0712L) in 1× NEBuffer 3.1 (New England Biolabs, Cat. #B7203) with incubation at 50 °C for ≥1 h. Linearized pDNA was then column-cleaned using the Monarch PCR & DNA cleanup kit (New England Biolabs, Cat. #T1030) following the manufacturer's protocols and eluting with 50 °C elution buffer. IVT was performed by combining 1000 ng of pDNA with RNA pol reaction buffer (New England Biolabs, Cat. #B9012) (1× final), magnesium acetate (Sigma-Aldrich, [St. Louis, MI, USA], Cat. #63052-100ML) (12.5 mM final), Triton X-100 (Sigma-Aldrich, Cat. #93443) (0.002% final), ATP (New England Biolabs, Cat. #N0450S) (5 mM final), GTP (New England Biolabs, Cat. #N0450S) (5 mM final), CTP (New England Biolabs, Cat. #N0450S) (5 mM final), N1-methyl-pseudouridine-TP (TriLink Biotechnologies [San Diego, CA, USA], Cat. #N-1081-1) (5 mM final), CleanCap Reagent AG (TriLink Biotechnologies, Cat. #N-7113-5) (4 mM final), RNase inhibitor (New England Biolabs, Cat. #M0314L) (10 U/μL final), DTT (Thermo Fisher Scientific [Waltham, MA, USA], Cat. #707265ML) (5 mM final), pyrophosphatase (New England Biolabs, Cat. #M0361L) (0.0025 U/μL final), and T7 polymerase (New England Biolabs, Cat. #M0251) (5 U/μL final) in a total reaction volume of 20 μL (brought to 20 μL with UltraPure dH₂O [Invitrogen (Waltham, MA, USA), Cat. #10977023]) and incubating at 37 °C for 4–6 h. After incubation, 28 μL of UltraPure dH₂O and 2 μL of DNAse I (New England Biolabs, Cat. #M0303S) was added; then, samples were incubated at 37 °C for 15 min. Resulting mRNA was then column-cleaned using the Monarch RNA cleanup kit (New England Biolabs, Cat. #T2040) following manufacturer's protocols with 2 additional wash steps and 5 min of RT incubation before centrifugation steps after the addition of washing or elution buffers. After cleanup, mRNA was quantified using a Qubit BR RNA assay kit (Thermo Fisher Scientific, Cat. #Q10211) and integrity was verified by gel electrophoresis. mRNA was stored at −80 °C.

For encapsulation, mRNA was diluted to 1 mg/mL with UltraPure dH₂O (Invitrogen, Cat. #10977023) in a total volume of 22 μL. mRNA was then encapsulated using the NanoAssemblr Spark Hepato9 siRNA kit (Cytiva [Marlborough, MA, USA], Cat. #NWS0009) following manufacturer's protocols and using 16 μL of Spark Nanoparticle mix. Resulting mRNA LNP was evaluated for concentration and encapsulation efficiency using the Quant-IT RiboGreen RNA assay kit (Thermo Fisher Scientific, Cat. #R11490) based on manufacturer's protocols with or without the addition of Triton-X 100 (Sigma-Aldrich, Cat. #93443) (2% final) and measured on a VICTOR2 plate reader. All mRNA LNPs were determined to have encapsulation efficiencies of ≥93.6%. Concentration of encapsulated mRNA was used for all further assays. mRNA LNP of each set were generated within 1 day of each other and stored at 4 °C for ≤7 days before use.

HEK293T/17 cells were cultured in cDMEM (complete DMEM; DMEM [Thermo Fisher Scientific, Cat. #11995073] supplemented with 10% heat-inactivated FBS (Thermo Fisher Scientific, Cat. #A3840302 or Cat. #12484028), 1X GlutaMAX [Thermo Fisher Scientific, Cat. #35050061], 1X Penicillin-Streptomycin [Thermo Fisher Scientific, Cat. #15140122], 1X MycoZap [Lonza (Basel, Switzerland), Cat. #VZA-2031]). HEK293T/17 cells were dissociated using TrypLE Express (Thermo Fisher Scientific, Cat. #12604013) for routine subculturing. K562 cells were cultured in cRPMI (complete RPMI; RPMI 1640 [Thermo Fisher Scientific, Cat. #11875119] supplemented with 10% heat-inactivated FBS, 1X GlutaMAX, 1X Penicillin-Streptomycin, 1X MycoZap, 10 mM HEPES [Thermo Fisher Scientific, Cat. #15630080], 1 mM sodium pyruvate [Thermo Fisher Scientific, Cat. #11360070], and 55 μM 2-mercaptoethanol [Thermo Fisher Scientific, Cat. #21985023]). All cells were grown in an incubator at 37 °C, 5% CO₂. Cells were cryopreserved in 90% FBS/10% DMSO in vapor-phase liquid nitrogen. HEK293T/17 and K562 cells were purchased from ATCC.

For mRNA LNP transfections of K562 cells, cells were seeded at a density of 600,000 cells/mL in 6-well tissue culture-treated plates (2.5 mL/well) in cRPMI supplemented with 1 μg/mL ApoE4 (Peprotech [Cranbury, NJ, USA], Cat. #350-04). mRNA LNP were adjusted to 8.33 ng/μL with D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250) and then cells were transfected by addition of 1.25 μg of mRNA LNP per well dropwise; then, plates were swirled and rocked to disperse complexes. For mock-transfected negative controls, D-PBS (MgCl⁻, CaCl⁻) was used in place of mRNA LNP. For conditions with protein transport inhibition, protein transport inhibitor cocktail (Invitrogen, Cat. #00-4980-83) was diluted in media and added to transfectants at ~1× final concentration at 16 h following transfection.

For pDNA transfections, HEK293T/17 cells were seeded at a density of 140,000 cells/cm² in 6-well tissue culture-treated plates (540,000 cells/mL, 2.5 mL/well) (Corning [Corning, Somerville, MA, USA], Cat. #353046). Transfections were performed at ~16.5 h following seeding. Based on the manufacturer's recommendations, pDNA (volume normalized with buffer EB [Qiagen (Venlo, The Netherlands), Cat. #19086]) was combined with TransIT-LT reagent (Mirus Bio [Madison, WI, USA], Cat. #MIR2305) at a ratio of 3 μL of TransIT-LT per 1 μg of pDNA per 108 μL of total volume in OptiMEM (Thermo Fisher Scientific, Cat. #3195062) and incubated at RT for ~20 min to allow complex formation. For mock-transfected negative controls, buffer EB alone was used in place of buffer EB/pDNA. Following complex formation, 2.5 μg of pDNA was added dropwise per well, and plates were swirled and/or rocked to disperse complexes.

For assessment of mSb fluorescence alone in transfectants, at ~24 h following transfection, cells were transferred to microfuge tubes, centrifuged at 300× g (4 °C), supernatant was removed, and cells were resuspended in D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250) and counted on an EVE automated cell counter. A volume equivalent to 1,000,000–1,500,000 cells was then centrifuged at 300× g (4 °C), supernatant was removed, and cells were resuspended in 500 μL FACS buffer (5% FBS in D-PBS). Cells were centrifuged at 300× g (4 °C), supernatant was removed, and then cells were resuspended in 400 μL of FACS buffer with 0.1 μg/mL DAPI (Sigma-Aldrich, Cat. #D9542).

For assessment of mSb fluorescence and surface mSb expression in transfectants, at ~24 h following transfection, cells were dissociated using 1 mL/plate of enzyme-free cell dissociation buffer (Thermo Fisher Scientific, Cat. #13151014) (to preserve surface proteins). Cells were transferred to microfuge tubes, centrifuged at 300× g (4 °C), supernatant was removed, and cells were resuspended in D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250) and counted on an EVE automated cell counter. A volume equivalent to ~500,000 cells was then centrifuged at 300× g (4 °C), supernatant was removed, and then cells were stained with 100 μL of 1/1000 anti-mSb monoclonal mouse IgG antibody (OriGene [Rockville, MD, USA], Cat. #TA180039) (diluted 1/1000 in FACS buffer [5% FBS in D-PBS]) for 30 min at 4 °C. Cells were then washed in 500 μL FACS buffer twice. Cells were then centrifuged at 300× g (4 °C), supernatant was removed, and cells were then stained with 100 μL of 1/200 diluted AlexaFluor647-conjugated goat anti-mouse IgG polyclonal antibody (Biolegend [San Diego, CA, USA], Cat. #405322) (diluted 1/200 in FACS buffer) for 30 min at 4 °C. Cells were then washed in 500 μL FACS buffer twice. Cells were then centrifuged at 300× g (4 °C), supernatant was removed, and cells were resuspended in 400 μL of FACS buffer with 0.1 μg/mL DAPI (Sigma-Aldrich, Cat. #D9542).

Following sample preparation, cells were strained through a 0.35 μm FACS tube cap (Corning, Cat. #352235), placed on ice, and analyzed on a Becton Dickinson LSRFortessa flow cytometer.

For assessment of transfectant media in Western blot, at ~24 h following transfection, media from K562 cells was collected from wells by transferring to a microfuge tube. Samples were centrifuged at 300× g for 5 min to pellet cells and supernatant was transferred to a new tube twice. cOmplete EDTA-free protease inhibitor (Roche [Basel, Switzerland], Cat. 11873580001) was then added to 1× final. Media samples were then stored at 4 °C for multimer experiments or −20 °C for all other experiments until Western blot analysis.

For assessment of transfectant cell lysate in Western blot, cells were collected at ~24 h following transfection. For HEK293T/17 cells, cells were dissociated using enzyme-free cell dissociation buffer (Thermo Fisher Scientific, Cat. #13151014) (to preserve surface proteins). Cells were then transferred to microfuge tubes, centrifuged at 300× g, supernatant was removed, and cells were resuspended in D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250) and counted on an EVE automated cell counter. For K562 cells, cells were transferred to microfuge tubes, centrifuged at 300× g, supernatant was removed, and cells were resuspended in D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250) and counted on an EVE automated cell counter. A volume equivalent to ~500,000 cells was then centrifuged at 300× g, supernatant was removed, and cells pellets were stored at −20 °C. Cell pellets were then thawed and resuspended in 100 μL of RIPA buffer (Thermo Fisher Scientific, Cat. #89900) with cOmplete EDTA-free protease inhibitor (Roche, Cat. #11873580001) added to 1–1.25× final. Samples were then incubated on ice for 15 min to lyse, then centrifuged at 14,000× g for 15 min at 4 °C. Supernatant was then transferred to a new tube. Lysed cell pellets and prepared cell lysates were stored at 20 °C. As needed, lysed cell pellets were re-centrifuged and supernatant was re-transferred to generate more prepared cell lysate sample.

Transfectant media or prepared cell lysate samples were then combined with a 1:3 volume ratio with Laemmli buffer (4×) (BioRad [Hercules, CA, USA], Cat. #1610747) supplemented with (reducing) or without (non-reducing) 2-mercaptoethanol (1.42 M final) (BioRad, Cat. #1610710), or with (reducing) DTT (50 mM final). Samples were heated (85 °C for sample sets with multimers, 95 °C for 5 min for sets without multimers) and then 5 μL were loaded per lane onto a 4–15% mini-PROTEAN TGX polyacrylamide gel (BioRad, Cat. #4561085 or 4561086) and ran at 50 V for 15–30 min; then, 150 V for 35–60 min in 1× Tris/Glycine/SDS running buffer (BioRad, Cat. #1610732). Proteins were then transferred to nitrocellulose blots in transfer buffer (1× Tris/Glycine [BioRad, Cat. #1610734], with 20% methanol) at 0.30 A for 1.33–1.5 h on ice. After transfer, blots were cut as needed, washed 1× in 10 mL TBS-T (TBS with 0.1% Tween-20), then blocked with 10 mL of EveryBlot blocking buffer (BioRad, Cat. #12010020) for 10 min at RT. Blots were then washed 2× in 10 mL TBS-T and then probed with 10 mL of mouse anti-StrepTagII monoclonal antibody (diluted to 1/5000) (LSBio, Cat. #LS-C413454), mouse anti-mSb monoclonal antibody (diluted 1/2000) (OriGene, Cat. #TA180039), or rabbit anti-vinculin monoclonal antibody (diluted 1/5000) (Abcam [Cambridge, UK], Cat. #ab129002) (all dilutions were in EveryBlot blocking buffer) overnight at 4 °C. Blots were then washed 4 times in 10 mL TBS-T and probed with 10 mL of HRP-conjugated goat anti-mouse IgG polyclonal antibody (diluted 1/10,000 in EveryBlot blocking buffer) (Jackson Immunology [West Grove, PA, USA], Cat. #115035003) or HRP-conjugated goat anti-rabbit IgG polyclonal antibody (diluted 1/10,000 in EveryBlot blocking buffer) (Jackson Immunology, Cat. #111035003), as appropriate, for 1 h at RT. Blots were then washed 4 times in 10 mL TBS-T, developed using Clarity Western ECL substrate (BioRad, Cat. #1705060) and then imaged on a BioRad ChemiDoc imaging system. SeeBluePlus2 pre-stained protein standard (Invitrogen, Cat. #LC5925) or HiMark pre-stained protein standard (Invitrogen, Cat. #LC5699) was run alongside samples to evaluate molecular weight.

Female C57BL/6J-A0201 (C57BL/6J-Mcph1^{Tg(HLA-A2.1)1Enge}/J [Jackson strain #003475]) mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA) and allowed to acclimate for ≥7 days following arrival. Mice were 7–8 weeks of age at the beginning of experiments. Each group consisted of 4 mice. All experiments involving animals were performed within the BC Cancer Animal Resource Centre (ARC) under specific pathogen-free conditions.

mRNA LNP vaccines were adjusted to 50 ng/μL using D-PBS (MgCl⁻, CaCl⁻) (Thermo Fisher Scientific, Cat. #14190250). Immediately prior to injection, mRNA LNP vaccines were warmed to RT. All vaccinations were performed as a prime-boost regimen, with each mouse receiving 20 μL of mRNA LNP vaccine (1 μg/mouse, intramuscularly) for prime and boost doses. Boost vaccinations were performed 21 days following prime vaccinations and experiment endpoint was 14 days following boost vaccinations.

For endpoint plasma/serum, immediately following euthanasia at experimental endpoint by isoflurane followed by CO₂ inhalation, whole blood was collected via cardiac puncture with a 25G needle. Whole blood was placed in an SST microtainer tube (Becton Dickinson, Cat. #365967), clotted for 30–60 min at RT, and then placed on ice/at 4 °C. Clotted whole blood was then centrifuged at 2000× g for 10 min at 4 °C and the supernatant was then transferred to a new tube and stored at −80 °C.

Immulon 2HB plates (Thermo Fisher Scientific, Cat. #14-245-61) were coated by addition of 100 μL/well of 2 μg/mL of recombinant mStrawberry (OriGene, Cat. #TP790044) in ELISA carbonate coating buffer (Invitrogen, Cat #CB01100) and incubation overnight at 4 °C. Plates were washed 3 times with 200 μL/well of 1× wash buffer (Invitrogen, Cat. #WB01, then blocked by addition of 200 μL/well of 1× ELISA buffer (Invitrogen, Cat. #DS98200) and incubation at RT for 3 h. Plates were then washed 3 times with 200 μL/well of 1× wash buffer. Next, 100 μL/well of individual serum samples from vaccinated mice or negative control C57BL/6J-A0201 serum (Jackson Laboratory, custom product) serially diluted in 1× ELISA buffer (Invitrogen, Cat. #DS98200) were added and plates were incubated overnight at 4 °C. Plates were then washed 3 times with 200 μL/well of 1× wash buffer and 100 μL/well of 0.2 μg/mL HRP-conjugated donkey anti-mouse IgG polyclonal antibody (Invitrogen, Cat. #A16017) diluted in 1× ELISA buffer was then added and plates were incubated for 2 h at RT. Plates were then washed 3 times with 200 μL/well of 1× wash buffer and 100 μL/well of TMB stabilized chromogen (Thermo Fisher Scientific, Cat. #SB01) was then added. Reactions were allowed to develop for ~3 min and were then stopped by addition of 100 μL/well of ELISA stop solution (Invitrogen, Cat. #SS03). OD₄₅₀ values were read on a Molecular Devices SpectraMax iD3 microplate reader, then adjusted to account for path length (dividing by the correction factor of 0.596, corresponding to a volume of 200 μL/well). All buffers were diluted to 1× using UltraPure dH₂O (Invitrogen, Cat. #10977023). ANOVA analysis was performed in R (v4.5.2) in RStudio (v2025.05.2, Build 517). Raw ELISA data is provided in Supplementary Data File S2.

Plasmid DNA (pDNA) constructs for generating mRNA for encapsulation and transfection/vaccination were designed incorporating several features for enhanced antigen expression, immunogenicity, and ease of use. These constructs have a T7 promoter followed by an AG initiator for use with TriLink CleanCap AG during in vitro transcription (IVT) reactions to generate 5′ Cap-1 (m7GpppN1mp) capped mRNA (which, as opposed to Cap-0 [m7GpppN1p], has an additional 2′-O-methylation at the nucleotide 1 position). Such Cap-1 capping has been shown to be essential in mimicking self mRNA of higher eukaryotes, which has 2′-O-methylation at the nucleotide 1 position and thereby reduces type I interferon activation and translational inhibition typically induced by uncapped or Cap-0 mRNAs [35,36,37]. The construct is also designed such that the IVT transcript has a 5′ α-globin UTR and a eukaryotic Kozak sequence motif for efficient protein translation [38,39] and 2 × 3′ β-globin untranslated regions (UTRs) for greater mRNA translation/stability [40]. mRNA CDS were codon-optimized using GenSmart codon optimization to enhance translation efficiency. The poly-A tail is 120 bp to promote translation/stability [40], bisected to reduce pDNA recombination [41] and directly incorporated into the plasmid sequence (rather than added following IVT via enzymatic polyadenylation, which results in inconsistent polyA lengths [40]). The construct was also designed for linearization prior to IVT via a type IIS restriction enzyme (SapI or BspQI), eliminating any 3′ overhang at the end of the polyA tail following linearization, which have been reported to inhibit mRNA translation/stability [40]. Lastly, within the pDNA construct, an upstream CMV promotor enables expression of the mRNA CDS and thereby allows for in vitro assessment of designs by pDNA transfection in mammalian cells, such as HEK293T/17 cells. In addition, the generation of mRNA was also performed using methodology for enhanced antigen expression and immunogenicity—during in vitro transcription, N¹-methylpseudouridine was incorporated to reduce TLR activation and associated decreases in cellular translation [42].

For the mRNA CDS sequence, four different designs were created: for intracellular antigen expression, for expression and secretion of antigen as a monomer, for expression and secretion of antigen as a multimer, and for expression and surface-display of antigen (Figure 1). Given that mRNA LNP-encoded antigens are expressed in the cytoplasm, the mRNA CDS design for intracellular expression contains the antigen of interest with no additional domains. To stimulate a more robust antibody response, the vaccine-encoded antigen should be accessible to the extracellular environment. Therefore, a murine-derived Igκ signal peptide was added to the N-terminus of the antigen of interest for the mRNA CDS for secretion of monomeric antigen, which targets the encoded protein to the ER for subsequent secretion in mammalian cells [43,44,45]. Multimerization domains (foldon-based, IMX313-based, or ferritin-based) were added in the C-terminal position in combination with Igκ signal peptide to generate an mRNA CDS for secretion of multimeric antigen. Surface-display is a commonly used alternative approach to extracellularly expose antigen. Furthermore, surface-displayed RBD-encoding saRNA LNP vaccines have shown greater antibody titers compared to secreted monomeric RBD [46]. Therefore, designs for surface-display were also created for direct comparison to secreted multimeric antigen designs. For the mRNA CDS for surface-displayed designs, the Igκ signal peptide is included in combination with a C-terminal transmembrane domain, either B7 (including the cytoplasmic tail) or PDGFR-based. In combination with the ER-directing signal peptide Igκ, both the B7 TM domain and PDGFR TM domains have been shown to lead to surface-display of the protein [47,48,49]. Lastly, a StrepTagII sequence was included at the C-terminal end of constructs. This enables the immobilization of antigen for various downstream applications, such as ELISA (using streptavidin-coated plates) or purification. StrepTagII sequences were chosen as tags, rather than other commonly used tags, such as HA or Myc sequences, which are microbe-derived and therefore potentially highly immunogenic. The StrepTagII sequences were not included in the mRNA CDS designs for secretion of multimeric antigen, given concerns that it may interfere with multimerization. For simplicity, these various mRNA CDS are named according to their features from N-terminus to C-terminus (e.g., Igκ-mSb-TM(B7)-Strep) and are referred to as "intracellular", "secreted monomeric", "secreted multimeric", or "surface-displayed" designs.

We firstly performed structural predictions of the different secreted multimeric design types using AlphaFold2 [33] to determine whether the structures were amendable to expected multimer formation (Figure 2). When inspecting the structures of a single subunit of each multimerization domain-containing design alone, each design appeared to have the multimerization domain portion accessible, suggesting multimerization should be possible. Predictions using the multimer parameter supported the ability of the structures to form multimers of 3-mers, 7-mers, and 24-mers for mSb-Foldon, mSb-IMX313, and mSb-Ferritin, respectively, as expected. Additionally, these structures also show the varying geometry of each multimerization domain: foldon generates a "bouquet"-like geometry, IMX313 generates a circular ring geometry, and ferritin generates a spherical geometry. Finally, these data suggest approximate protein particle sizes of 9.5 nm W/L × 6.6 nm H for mSb-Foldon, 11.5 nm W/L × 7.5 nm H for mSb-IMX313, and 22 nm W/L and H for mSb-Ferritin. Therefore, these predicted multimer sizes support APC-independent lymph node drainage (10–100 nm [11,12,13]) as a potential mechanism towards increasing immunogenicity for IMX313 and ferritin multimers.

Secreted monomeric and secreted multimeric designs were then evaluated in vitro for expression/secretion and expression/secretion/multimerization, respectively (Figure 3). For these experiments, mSb-Strep (intracellular) served as a negative control for evaluating secretion. K562 cells were transfected with mSb-Strep (intracellular), Igκ-mSb-Strep (secreted monomeric) mRNA LNP, or mock-transfected (negative control for transfection/mSb expression) with or without addition of protein transport inhibitor cocktail (PTIC); then, mSb fluorescence was evaluated by flow cytometry (Figure 3A). PTIC contains brefeldin A and monensin which are well-established inhibitors of the export pathway in mammalian cells [50,51]. The addition of PTIC had no effect on mock-transfected negative control cells or mSb-Strep mRNA LNP-transfected cells. Conversely, the addition of PTIC in Igκ-mSb-Strep mRNA LNP-transfected cells resulted in increased mSb fluorescence. Given that inhibition of secretion led to intracellular accumulation of mSb, these data indicate the inclusion of the N-terminal Igκ signal peptide successfully leads to secretion in our designs. All mRNA LNP successfully led to mSb expression—transfection resulted in mSb fluorescent positivity of 99.0% for mSb-Strep, 85.7% for Igκ-mSb-Strep, 98.5% for Igκ-mSb-Foldon, 99.0% for Igκ-mSb-IMX313, and 67.8% for Igκ-mSb-Ferritin, compared to 0.1% positivity in mock-transfected negative controls (Figure 3B).

mSb expression, secretion, and multimerization were evaluated by Western blot analysis of cell lysate and cell culture media samples (Figure 3C). Cell lysate samples run under reducing conditions showed the expected banding of all mSb designs (Igκ = 2.4 kDa, mSb = 26.6 kDa, StrepTagII = 1.1 kDa, foldon = 3.1 kDa, IMX313 = 6.2 kDa, ferritin = 20.0 kDa). Cell culture media samples run under reducing conditions similarly demonstrated expected banding of all mSb designs. Of note, cell culture media of Igκ-mSb-Strep appeared as two similarly sized bands, possibly reflecting a species with post-translational modifications associated with the secretion pathway. Cell culture media of mSb-Strep had only a faintly present band, likely representing release of protein from cells which died/lysed in the period between transfection and collection. Thus, these data further support that the N-terminal Igκ signal peptide successfully leads to antigen secretion, as observed by flow cytometry, and demonstrates successful secretion of Igκ-mSb-Foldon, Igκ-mSb-IMX313, and Igκ-mSb-Ferritin. Given that foldon, IMX313, and ferritin-based multimerization results in 3-mer, 7-mer, and 24-mer products, respectively, a corresponding 3×, 7×, and 24× shift in molecular weight is expected. Comparing reducing and non-reducing conditions, a shift in band size was observed for Igκ-mSb-IMX313, corresponding to the formation of a 7-mer product (32 kDa × 7 = 224 kDa). Additionally, the presence of only the 7-mer band in the cell media sample of Igκ-mSb-IMX313 suggests complete multimerization of these monomers prior to secretion and that the multimer is relatively stable in the extracellular environment. The presence of multiple other bands in the cell lysate sample of Igκ-mSb-IMX313 possibly reflects a stepwise nature of multimer formation. No bands corresponding to multimers were observed for Igκ-mSb-Foldon or Igκ-mSb-Ferritin. Under non-reducing or reducing conditions, multimeric and monomeric forms, respectively, have been observed for foldon-based [52], IMX313-based [22,53], and ferritin-based multimers [52]. Therefore, under conditions permissive to multimer detection, we find no evidence of multimerization for IgK-mSb-Foldon or IgK-mSb-Ferritin.

Next, surface-displayed designs were evaluated in vitro for expression and surface-display (Figure 4). For these experiments, mSb-Strep (intracellular) served as a negative control for evaluating surface-display. HEK293T/17 cells were transfected with mSb-Strep, Igκ-mSb-TM(PDGFR)-Strep, or Igκ-mSb-TM(B7)-Strep pDNA or mock-transfected (negative control for transfection/mSb expression) and then mSb fluorescence and mSb surface staining was evaluated by flow cytometry (Figure 4A). Transfection resulted in mSb fluorescent positivity of 69.6% for mSb-Strep, 81.1% for Igκ-mSb-TM(PDGFR)-Strep, and 70.7% for Igκ-mSb-TM(B7)-Strep, compared to 0.1% positivity in mock-transfected negative controls, indicating all designs successfully lead to mSb expression. Alternatively, surface staining was detected at a positivity of 0.3% for mSb-Strep, 98.2% for Igκ-mSb-TM(PDGFR)-Strep, and 98.1% for Igκ-mSb-TM(B7)-Strep, compared to 0.3% positivity in mock-transfected negative controls. These data showed that the inclusion of the N-terminal Igκ signal peptide with a PDGFR or B7 TM domain results in successful surface-display.

As a secondary validation of expression, Western blot analysis was performed on cell lysates of mSb-Strep, Igκ-mSb-TM(PDGFR)-Strep, or Igκ-mSb-TM(B7)-Strep transfectants (Figure 4B). A dominant band corresponding to expected sizes for all mSb designs was observed (Igκ = 2.4 kDa, mSb = 26.6 kDa, TM(PDGFR) = 5.4 kDa, TM(B7) = 7.7 kDa, StrepTagII = 1.1 kDa). Igκ-mSb-TM(B7)-Strep was selected for further evaluation given the ~7.6-fold longer half-life of B7-based surface-display compared to PDGFR-based surface-display [54].

Next, to determine the relative immunogenicity of the various mSb-encoding designs, an in vivo prime-boost vaccination study was performed in C57BL/6-A0201 mice, comparing mRNA LNP vaccines of the following designs: Igκ-mSb-Strep, Igκ-mSb-TM(B7)-Strep, Igκ-mSb-Foldon, Igκ-mSb-IMX313, or Igκ-mSb-Ferritin (Figure 5A) (n = 4 mice/group). Of note, the mice used were transgenic for HLA-A*02:01⁺ so as to allow future assessment of MHC class I epitopes presented by this allele, which is beyond the scope of the present study. At experimental endpoint (14 days following boost vaccination), blood samples were collected, and serum was isolated. To assess immunogenicity following vaccination, an indirect ELISA using recombinant mSb-coated plates was performed for the serum of each mouse (Figure 5B). All vaccines elicited responses compared to negative control serum samples derived from the same mouse strain (p ≤ 0.05; ANOVA), except Igκ-mSb-Ferritin. No significant difference was observed between Igκ-mSb-Strep and Igκ-mSb-TM(B7)-Strep (p > 0.05), indicating secretion and surface-display are equally immunogenic in this context. Similarly, Igκ-mSb-Foldon or Igκ-mSb-Ferritin showed no difference (p > 0.05) in antigen-specific IgG relative to the Igκ-mSb-Strep. Assuming this lack of increased antigen-specific IgG was due to lack of multimerization, these findings are in agreement with our in vitro observations. Alternatively, Igκ-mSb-IMX313 generated ~5-fold greater antigen-specific IgG compared to its monomeric counterpart (p ≤ 0.05). These data indicate that high rates of antigen self-multimerization can be achieved from mRNA LNP, translating to improved antibody responses, and suggest that IMX313 may be well-suited as a multimerization domain in the context of mRNA LNP vaccination.