Modified mRNA-Based Vaccines Elicit Robust Immune Responses and Protect Guinea Pigs From Ebola Virus Disease

We developed 2 novel lipid nanoparticle-encapsulated, nonreplicating RNA-based Ebola virus vaccines encoding glycoprotein with 1-methylpseudoUTP nucleoside modifications and different signal peptides. Vaccination of guinea pigs elicited high Ebola-specific IgG and neutralizing antibodies and conferred complete protection against Ebola virus disease.

Abstract

Most current Ebola virus (EBOV) vaccine candidates are based on viral vectors, some of which cause side effects or require complex manufacturing. Modified mRNA vaccines are easily produced, safe, and are highly immunogenic. We developed 2 mRNA vaccines based on the EBOV envelope glycoprotein, which differed by the nature of signal peptide for improved glycoprotein post-translational translocation. The mRNAs were formulated with lipid nanoparticles to facilitate delivery. Vaccination of guinea pigs induced EBOV-specific IgG and neutralizing antibody responses and 100% survival after EBOV infection. The efficacy of our mRNA vaccine combined with preclinical safety data supports testing in clinical studies.

There is no commercially available vaccine against EBOV. Promising developmental vaccines were rushed into clinical trials through collaborative efforts involving the World Health Organization in response to the unprecedented 2014–2016 epidemic of Ebola virus disease in West Africa, which led to over 11000 fatalities. A recent phase 2 clinical trial of the lead candidate, a vesicular stomatitis virus (VSV) expressing the EBOV glycoprotein (GP) in place of its G protein, demonstrated a durable antibody response [1]. However, adverse effects were previously exhibited [2] and facing the potentiality of unpredictable and uncontrolled outbreaks, the demand for safe, rapidly scalable, and efficient vaccines remains. Vaccine manufacturing lag times of up to 6 months cannot be afforded during attempts to contain this highly infectious virus.

Recently, nucleic acid-based vaccines against EBOV have come into focus. An RNA replicon based on Venezuelan equine encephalitis virus encoding EBOV GP, delivered in 2 doses using modified dendrimer nanoparticles, protected laboratory mice against lethal infection [3]. The large size of this self-replicating mRNA, however, introduces additional challenges for large-scale production and purification. DNA vaccines expressing the GP of EBOV or closely related Marburg virus entered phase 1 clinical trials, but not all participants responded despite multiple boosters for sustained immune responses [4].

Interest in developing mRNA-based vaccines against infectious diseases was sparked after pioneering experiments demonstrated in vivo protein expression of exogenously introduced naked mRNA in mouse muscle [5]. mRNA vaccines are highly advantageous over viral-based replicons or plasmid DNA vaccines, enabling rapid discovery, development, and ease of manufacture, mitigating risk of insertional mutagenesis, and facilitating strong and balanced antibody and cell-mediated immune responses. Recently, we [6] and others [7] have shown how lipid nanoparticle (LNP)-formulated, modified mRNA vaccines can overcome some of the issues that can impede mRNA vaccines development; namely the possible indiscriminate recognition by innate immune sensors and sensitivity to degradation.

In this study, using the technology described in Richner et al [6], we developed 2 novel mRNA vaccines encoding the EBOV GP in a membrane-bound form. A human kappa immunoglobulin (Igκ) signal peptide or the wild-type signal peptide sequence of GP were also incorporated into the vaccine transcripts for efficient translocation through the cell's secretory network and display on the cell's surface. We show here that 2 doses of either vaccine were highly efficacious in an animal model and demonstrate all the attributes of a promising candidate to advance to clinical trials.

METHODS

mRNA Synthesis and Formulation

Two mRNA vaccines were synthesized in vitro by T7 polymerase-mediated transcription where the UTP were substituted with 1-methylpseudo UTP, using a linearized DNA template encoding EBOV GP (variant Yambuku, isolate Mayinga, GenBank accession number: AAN37507), which incorporates 5′ and 3′ untranslated regions (UTRs) [6]. The 2 vaccines differed in the signal

peptide sequence: Vaccine A uses the wild-type sequence of GP (MGVTGILQLPRDRFKRTSFFLWVIILFQRTFS), whereas Vaccine B uses the signal sequence from human Igκ subgroup V (METPAQLLFLLLLWLPDTTG). The mRNAs were purified and resuspended in a citrate buffer at the desired concentration. A donor methyl group S-adenosylmethionine (SAM) was added to methylated capped RNA (cap-0), resulting in a cap-1 structure to increase mRNA translation efficiency [8]. LNP formulations were prepared as previously described [6].

Animal Experiment

Three groups of adult female guinea pigs, strain Hartley (n = 5), were intramuscularly vaccinated on days 0 and 21 with 0.1 mL of Vaccine A (0.2 mg/mL), Vaccine B (0.2 mg/mL), or phosphate buffered saline (PBS). On day 42, the animals were intraperitoneally infected with 1000 plaque-forming units (PFU) of guinea pig-adapted EBOV variant Yambuku, isolate Mayinga. Serum was collected from the anterior vena cava on days 0, 21, and 42 postimmunization and at 3-day intervals over 12 days postinfection, and at day 28 postinfection, the time of euthanasia. The animal experiment was conducted in compliance with guidelines approved by the University of Texas Medical Branch Institutional Animal Care and Use Committee, Galveston.

Viremia

Vero-E6 cells were inoculated with 10-fold serially diluted serum in MEM (ThermoFisher Scientific, Waltham, MA) containing 0.05 mg/mL gentamicin (ThermoFisher Scientific, Waltham, MA). After 1 hour absorption at 37°C, the inoculum was replaced with carboxymethyl cellulose overlay, plates were incubated for 6 days, monolayers were fixed with formalin, and EBOV plaques were immunostained [9].

EBOV Plaque Reduction Assays

Plaque reduction neutralization assays were performed as described [9].

Enzyme Linked Immunosorbent Assays

Enzyme linked immunosorbent assays (ELISAs) were performed as described [9] with modifications. Plates were coated with EBOV GP (7.25 ng/well, IBT Bioservices, Rockville, MD) and blocked with 3% milk powder/PBS. Sera diluted 4-fold in blocking solution starting at 1:10, was added to the plate. A peroxidase-labeled goat anti-guinea pig IgG (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) detected bound antibodies.

Statistics

Two-way ANOVA was used to determine the P values between the vaccine groups (GraphPad software, Inc.). Differences were considered significant when P < .05.

RESULTS

We designed 2 EBOV GP mRNA-based vaccines (Figure 1A). The in vitro transcribed mRNAs for both vaccines encoded GP of EBOV, variant Yambuku isolate Mayinga. Vaccine A encoded full-length EBOV GP, while in Vaccine B the authentic signal peptide of GP was swapped with that from human Igκ. The vaccine mRNA transcripts were encapsulated in a LNP formulation for intramuscular delivery in vivo.

Figure 1.

EBOV mRNA vaccine constructs and antibody responses in vaccinated guinea pigs. A, The mRNA vaccines encode nucleoside modified EBOV glycoprotein (GP), flanked by 5′ and 3′ untranslated regions (UTRs) and a polyA tail. Both vaccines possess a cap-1 structure but differ by the presence of a code for a signal peptide (wt-SP or Igκ-SP). Antibody responses in vaccinated guinea pigs: (B) total GP-specific IgG and (C) EBOV neutralizing titers. Vaccine A, orange circle symbols. Vaccine B, blue square symbols. Mock, black triangle symbols. Bars denote group means. Vaccine A versus Vaccine B: *P = .0175, ****P = < .0001.

Three groups of guinea pigs were immunized with vaccine formulations A, B or PBS (the mock-treated group) on days 0 and 21. EBOV GP-specific serum IgG analyzed by ELISA were detected in both vaccinated groups at day 21 by which time the responses to Vaccine B were significantly higher than for Vaccine A (Figure 1B). At day 42, three weeks after the second dose, the IgG responses further increased in both groups, with group B maintaining higher titers. EBOV-neutralizing antibodies were detected in response to Vaccine B (1:160), but not Vaccine A on day 21 (Figure 1C). After the second dose, EBOV-neutralizing antibodies were also detected for Vaccine A (1:140) whilst that for Vaccine B increased to 1:1800.

To test protective efficacy of the vaccine, the animals were infected with lethal guinea pig-adapted EBOV on day 42. The mock-immunized animals developed severe disease as evidenced by the increase in clinical score, body temperature, weight loss, and death or euthanasia of moribund animals on days 7–10 (Figure 2). In contrast, all animals immunized with either Vaccine A or B demonstrated no clinical signs of illness, normal body temperature, no weight loss, and 100% protection against death. Analysis of serum samples demonstrated a viremia in the control group starting at 3 days postinfection, which increased to high levels until the time of euthanasia. Very low levels of viremia were detected on day 3 and day 6 in some vaccinated animals, which was no longer detectable by day 28.

Figure 2.

Survival (A), disease scores (B), body temperature (C), weight loss (D), and viremia (E) in guinea pigs vaccinated with Vaccine A (orange circles), Vaccine B (blue squares), or mock-vaccinated (black triangles) on days 0 and 21 and challenged with EBOV on day 42. Mean values ± SD. E, Dotted line represents the limit of detection.

DISCUSSION

In this study, we examine the efficacy of 2 novel LNP-encapsulated, nonreplicating EBOV mRNA vaccines encoding full-length GP with 1-methylpseudo UTP nucleoside modifications and different signal peptides. Both vaccines successfully protected guinea pigs against lethal EBOV infection and signs of clinical illness. Titers of EBOV-neutralizing antibodies exceeded levels induced by the VSV-vectored vaccine in guinea pigs and nonhuman primates, while GP-specific IgG titers were, however, lower in comparison [10], though side-by-side comparisons using the same study schedule and ELISA assay conditions are required to draw firm conclusions. The human Igκ signal peptide in Vaccine B GP mRNA may have contributed to improved translocation through the intracellular secretory network for subsequent display on the cell surface. However, we cannot rule out the different codons in the 2 constructs affording greater translational efficiency of Vaccine B over Vaccine A. More potent GP-specific IgG and EBOV-neutralizing antibodies were generated after one dose of Vaccine B compared to Vaccine A, suggesting that a single dose of the Vaccine B may likely be sufficient for protection. We show that mRNA vaccines may represent a viable addition to current EBOV vaccine contenders, given the need for safer, rapidly designed, and less-reactogenic platforms. Thus far, mRNA vaccines targeting influenza A virus [11], rabies virus [12], HIV-1 [13], and Zika virus [6] have been tested in infectious disease animal models and, most recently, in humans [14].

Enhancing the potency of mRNA vaccines can involve an optimal delivery vehicle. LNPs, used to deliver our mRNA template, are proving to be a viable option already utilized to improve the uptake and stability of therapeutic siRNA in human trials [15] and other mRNA vaccines [6, 7, 14]. Antibody titers can be substantially greater than that achieved with vaccination with purified inactivated virus or plasmid DNA-based vaccine candidates [7]. Modifications to the 5′ cap structure and the incorporation of 5′ and 3′ UTRs are all tools used to improve intracellular stability and translation [8]. RNA has the ability to stimulate pattern recognition receptors, such as Toll-like receptors or retinoic acid-inducible gene I-like receptors, creating an adjuvant effect to nonspecifically activate the immune response, which in turn prevents efficient translation of the mRNA vaccine. We incorporate nucleoside modifications, which minimize the indiscriminate activation of these innate immune sensors and provide stability for greater expression of the encoded GP protein.

The majority of EBOV vaccines advancing through clinical trials are based on virus vectors. During the initial clinical trials of the VSV vector-based vaccine, oligoarthritis and other adverse effects were observed in some high-dose recipients [2]. GP is the known causative agent of inflammation, vascular leakage, and cytotoxicity and may alter the tropism of the VSV backbone. However, dosage, replicative properties of the viral delivery system, and the route of vaccination may be the critical determinants. Tapering the dosage of live vaccines may not solve vaccination reactogenicity on a per case basis and could compromise the protective response [2]. Nevertheless, nonreplicative vaccines offer a safer alternative to virus-vectored vaccines and may address concerns from the VSV-vectored EBOV vaccine trials. The nonreplicative nature of the mRNA modality may be well tolerated in humans while still eliciting a robust response, unhampered by vector-derived or pre-existing immunity and devoid of a virus vector with potentiated alteration in virulence.

Currently, virus-vectored vaccines in clinical trials and our mRNA vaccines encode the full-length GP. The heavily glycosylated mucin-like domain and the transmembrane domain of GP are associated with cytotoxicity and may be not necessary for generating a neutralizing antibody response. mRNA vaccines can be tailored to exclude cytotoxic domains and focus on displaying protective epitopes, while maintaining a conformational integrity and stability, which yields efficacious antibody and T-cell responses. The removal of large structural domains from GP is unlikely to be tolerated by live attenuated chimeric vaccine viruses, such as the VSV-vectored EBOV vaccine used in the recent clinical trials [1], which lack their own sole G envelope protein and require a functional EBOV GP to enable entry. In contrast, such modifications would not impede nonreplicating nucleic acid vaccine platforms, which use alternate vehicles for delivery of genetic information.

Other nonreplicating EBOV vaccine candidates include DNA-based, adjuvanted GP nanoparticles, virus-like particles, GP subunit, and replication-deficient viral vectors. An EBOV mRNA vaccine is advantageous over the plasmid DNA-based vaccines as there is no risk of host chromosomal integration. Unlike DNA vaccines, mRNA vaccines do not require entry into the nucleus for transcription. Translation occurs directly in the cytoplasm and therefore eliminates the rate-limiting step of nucleus translocation, which may be defined by the cell type and delivery vehicle, and which is the likely cause for large doses of several milligrams, followed by electroporation, typically required for DNA vaccines. Indeed, in recent phase 1 clinical trials, not all participants responded to DNA vaccination despite receiving up to 4 doses [4]. Nonreplicative mRNAs vaccine modalities are gaining momentum in the infectious disease field. They are advantageous in their ability to be rapidly tailored to sequences of outbreak viruses and new emergent strains. There is potential to deliver a cocktail of mRNAs, which encompass the growing repertoire of EBOV strains that differ by several amino acids. The potential to use combinations of mRNAs encoding GP antigens from the different filovirus species may allow for a broadly cross-protective pan-filovirus vaccine. The promising results from this efficacy study warrants further testing of EBOV-targeting mRNA vaccines in nonhuman primates for potential advancement into clinical development.

Notes

Acknowledgments. We thank Jessica Graber and ARC, UTMB, for assistance with the vaccination and challenge study.

Financial support. This work was supported by a research grant from Moderna Therapeutics (A. B.) and National Institutes of Health (grant number 1R01AI102887-01A1) (A. B.).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.