An mRNA-based rabies vaccine induces strong protective immune responses in mice and dogs

BHK-21 cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM) containing 10% FBS at 37 °C with 5% CO₂. RABV strain CVS-24 (GenBank: ADR03123.1) was provided by the Institute of Animal Health, Guangdong Academy of Agricultural Sciences (Guangzhou, China). RABV strain BD06 (GenBank: ACB38373.1) was provided by the Institute of Military Veterinary Medicine, Academy of Military Medical Sciences (Changchun, China). BD06 strain could cause 80% mortality in unvaccinated dogs after challenge [35] and is responsible for most rabies cases in humans and dogs in China [36].

mRNA vaccines were produced based on the Liverna Therapeutics platform (China patent ZL201911042634.2). The mRNA molecules included a 5’ cap structure, a 5’ UTR, an ORF, a 3’ UTR and a poly(A) tail. The ORF in this study encodes the glycoprotein (RABV-G) of the CTN-1 strain (GenBank: ACR39382.1), which has been used for production of human rabies vaccine in China and is Chinese domestic isolates [37, 38]. Three different mRNA constructs were developed: RABV G-A, RABV G-B, and RABV G-C. The RABV G-A sequence consisted of a 5’ UTR, the ORF of RABV-G, a 3’ UTR and a 64A + 36C + histone stem loop. RABV G-B was an optimized RABV G-A sequence with a different ORF. RABV G-C was identical to RABV G-B with the exception of its poly(A) tail. Unless specifically noted, RABV G-C was the sequence of our mRNA vaccine, named LVRNA001.

mRNAs were produced by in vitro transcription (IVT). DNA templates were linearized from plasmids containing the open reading frames flanked by 5’/ 3’UTR and Poly-A sequences. IVT reactions were performed using DNA templates, an optimized T7 RNA polymerase (Novoprotein Scientific Inc., China), NTP and Cap GAG m7G(5′)ppp(5′ (2′-OMeA)pG (Jiangsu Synthgene Biotechnology Co, China). The reaction was terminated by addition of DNase I (Novoprotein Scientific Inc., China). mRNAs were purified using Oligo-dT affinity column (Sepax Technologies, Inc., China) and Tangential Flow Filtration (TFF, Repligen Corporation, America). Microfluidic capillary electrophoresis (Fragment Analyzer systems 5200, Agilent) was used to assess RNA integrity, and the characterization including concentration, pH, residual DNA, proteins, and dsRNA impurities of the solution were also performed. To prepare the mRNA vaccines, purified mRNAs were encapsulated in LNPs according to a modified procedure wherein cholesterol, 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), a polyethylene glycol-lipid and a cationic lipid was rapidly mixed with an aqueous solution containing mRNA. Then, the analytical characterization of product was carried out, including the determination of particle size and polydispersity, encapsulation, pH, endotoxin, and bioburden. The final products were lyophilized into powder in 2 ml glass vials. Sterile water (1 ml) was added to each vial to produce a 25 μg/ml (mRNA) solution before use. Appropriate volumes were taken from the vial for animal injection according to the experimental plan.

Inactivated vaccines were purchased from an animal hospital (Rabvac®, produced by Boehringer Ingelheim Vetmedica, Inc. Lot number 4150466A, labelled potency is 1 dose ≥ 2.0 IU) or a domestic vaccine manufacturer (rabies vaccine made from aGV strain for human use, freeze-dried, labelled potency is 1 dose ≥ 2.5 IU).

Protein expression was measured using flow cytometry analysis as previously described [39]. Briefly, HEK293 cells were transfected with either RABV-G mRNA or Luc mRNA (negative control) for 24 h. Cells were then collected and stained with a monoclonal mouse anti-rabies antibody (HyTest Ltd, Turku, Finland) and an FITC-labeled goat anti-mouse IgG (Life Technologies GmbH, Darmstadt, Germany). Flow cytometric analysis of FITC-positive cells confirmed protein expression.

The morphology of the nanoparticles was analyzed using TEM. Briefly, a drop of aqueous nanoparticle sample was dropped onto a carbon-coated copper grid. Subsequently, the grid was air-dried completely at room temperature. TEM micrograph images were captured at an accelerating voltage of 80 kV.

BALB/c mice (~ 6 weeks old, 20–25 g) were purchased from the Institute of Animal Health, Guangdong Academy of Agricultural Sciences (Guangzhou, China). Dogs (~ 3 months old beagles, 5–6 kg) were obtained from the Institute of Military Veterinary Medicine, Academy of Military Medical Sciences (Changchun, China). All experiments were approved by the Research Ethics Committee of the College of Animal Health, Guangdong Academy of Agricultural Sciences, and the Research Ethics Committee of the College of the Institute of Military Veterinary Medicine, Academy of Military Medical Sciences (IACUC of AMMS-11–2021-19). Mouse immunizations were carried out according to the methods described previously [40], LVRNA001 (0.2–5 μg) or 0.1 dose (1 dose ≥ 2.0 IU) of Rabvac® were intramuscularly (i.m.) injected into the thigh of hindlimb once (0d) or twice (0d/7d); 14 days post immunization (dpi), blood samples were collected for antibody tests and mice were injected with viruses (strain CVS-24, 50-fold LD₅₀ pre-determined by the Institute of Animal Health, Guangdong Academy of Agricultural Sciences for mouse in this experiment) intracerebrally, followed by observation of animal survival and last round of blood sample collection on day 21 post challenge for neutralization antibody test.

For dog challenge studies, LVRNA001 (5 or 25 μg, 0d/7d or 0d/7d/21d) or an inactivated vaccine made for human use (1 dose, ≥ 2.5 IU, 0d/3d/7d/14d/28d) were i.m. injected into the lateral thigh of hind limb. Blood samples were collected on days 0, 7, 9, 11, 13, 35 post first injection for antibody tests. Viruses (strain BD06) were given to the test dogs on day 35 by i.m. injection to the biceps femoris of the hind limb at 50-fold LD₅₀ (pre-determined by the Institute of Military Veterinary Medicine, Academy of Military Medical Sciences for dog in this experiment). Observation of animal survival continued through 3 months after challenge, when blood samples were collected from survived dogs for neutralization antibody test.

Dogs were intramuscularly injected with 50-fold LD₅₀ of virulent RABV-BD06 strain in the biceps femoris of the hind limb. Six hours after challenge infection, dogs were i.m. injected with LVRNA001 or inactivated vaccine. Immunization procedures and experimental steps were the same as pre-exposure protocols shown above.

Spleen lymphocytes from mice in each immunization group were collected and resuspended in RPMI 1640 medium containing 10% fetal bovine serum (FBS). Then, 4 × 10⁶ cells were seeded into a 24-well flat-bottom tissue culture plate and incubated with 5 μg of a synthesized peptide library of RABV-G for 72 h at 37°C. The cell supernatants were collected, the interferon (IFN)-γ and interleukin (IL)-4 levels were measured using ELISA kits (Neo Bioscience, China) according to the manufacturer’s instructions.

TNF-α-producing CD3⁺/CD4⁺ or CD3⁺/CD8⁺ T cells from mouse spleen lymphocytes (collected 7 days after booster immunization) were analyzed using flow cytometry. Spleen lymphocytes (4 × 10⁶ cells/ml) were seeded into a 24-well flat-bottom tissue culture plate and incubated with 5 μg of a synthesized peptide library of RABV-G for 72 h at 37°C. Cells were then stained with the following antibodies at 4°C for 20 min: CD8-FITC (1:200), CD3-PerCP-Cy5.5 (1:200), CD4-BV605 (1:200) (all from BD Biosciences, Heidelberg, Germany), and TNF-α-PE (1:100) (BioLegend, San Diego, CA, USA). The fluorescence signals from staining were measured by flow cytometry, and data analysis was conducted using FlowJo software (Tree Star, Inc., Ashland, USA).

Mouse serum samples were collected and scanned for RABV-G-specific immunoglobulin using a commercially available RABV antibody detection kit (Synbiotics Corp, France) following the manufacturer’s instructions. A positive antibody titer was recognized as OD₄₅₀ > 0.2.

Viral neutralizing antibody (VNA) titers against RABV were determined by fluorescent antibody virus neutralization (FAVN) tests according to standards set forth by the World Organization for Animal Health. Serum samples were serially diluted into a 96-well plate along with WHO reference serum diluted to 0.5 IU/ml. A 50 μL suspension containing the 50% tissue culture infectious dose (TCID₅₀) of challenge virus standard strain 11 (CVS-11, obtained from Chinese National Institutes for Food and Drug Control) was added to each well, and the plates were incubated at 37°C for 1 h. BHK-21 cells were cultured with DMEM containing 10% FBS and then added to the wells, and the plates were incubated at 37°C in a humidified incubator with 5% CO₂ for 48 h. Afterward, the cells were washed and fixed in 80% cold acetone for 30 min. Last, the cells were covered with a RABV-specific monoclonal antibody (KPL, Gaithersburg, MD, USA) followed by FITC-conjugated goat anti-mouse IgG (KPL, Gaithersburg, MD, USA). Fluorescent signals in each well were detected with a fluorescence microscope.

Mouse brain tissue was collected and fixed with 4% paraformaldehyde. Tissue samples were blocked with bovine serum albumin (BSA) and incubated with a RABV-specific primary antibody for 1 h. Samples were then washed with an FITC-conjugated secondary antibody and imaged with a fluorescence microscope.

Data are presented as the mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 6.0 software. Comparisons between groups were performed with one-way ANOVA followed by Tukey’s test. A p value less than 0.05 was considered statistically significant.

RABV-G is the major viral antigen that induces neutralizing antibody production in infected species [41, 42]. To both improve its expression and stabilize its desired conformation, we designed a panel of RABV-G mRNA antigens: RABV-G A, RABV-G B, and RABV-G C (Fig. 1A). We then compared the antigen expression induced by these optimized mRNA constructs. Flow cytometry analysis revealed that the expression of RABV-G induced by RABV-G C (95.0%) was the highest among the three mRNA constructs (Fig. 1B), suggesting that both sequence optimization and poly(A) tailing are necessary for cellular G protein expression. As lipid shells are necessary for mRNA vaccine delivery [43], the RABV-G C sequence was then encapsulated into lipid nanoparticles (LNPs) to become our mRNA vaccine candidate LVRNA001. Briefly, the LVRNA001 construct includes a 5′ cap structure, a 5’ UTR, an open reading frame (ORF), a 3′ UTR and a poly(A) tail. The ORF in this study encodes the glycoprotein (RABV-G) of the CTN-1 strain (GenBank: ACR39382.1). Spherical nanoparticles with a uniform size distribution were observed under an TEM (Fig. 1C).

To evaluate the immunogenicity of LVRNA001, BALB/c mice were vaccinated via intramuscular (i.m.) injection with LVRNA001 at different dosages: 5 μg, 1 μg, and 0.2 μg. ELISA detected anti-RABV-G IgG levels and demonstrated that the IgG levels induced by LVRNA001 were dose-dependent (Fig. 2A). Specifically, the FAVN test revealed that mice immunized with 5 μg or 1 μg of LVRNA001 achieved neutralizing antibody levels above the threshold concentration of antibody (0.5 IU/ml), considered as protective in humans, dogs and cats [44], whereas mice immunized with 0.2 μg of LVRNA001 or PBS did not achieve this threshold antibody concentration (Fig. 2B). Then, we monitored the IgG antibody titers in the serum of mice at different time points (0.5 months, 1 month, 3 months and 6 months) after vaccination with 5 μg or 1 μg of LVRNA001. We found that both the 5 μg and 1 μg doses of LVRNA001 sustained high levels (> 0.2 ELISA cutoff) of IgG antibody titers at all time points, with levels peaking at 3 months post-immunization (Fig. 2C). Our results indicate that LVRNA001 administered at a dose of 5 μg or 1 μg can effectively induce antibody production in mice and sustain high antibody levels for at least 6 months post-immunization.

In mice vaccinated with 5 μg or 1 μg of LVRNA001, the Th1 and Th2 cellular immune responses were detected and characterized (Fig. 3) in comparison with an inactivated vaccine Rabvac®. One (0 d) or two inoculations (0d/7d) with LVRNA001 at either dosage produced significantly higher IFN-γ levels than with PBS or Rabvac® (Fig. 3A). No differences in IL-4 levels were observed between the LVRNA001-immunized groups and the PBS or inactivated vaccine-immunized groups (Fig. 3B). The expression of TNF-α-producing CD4⁺ and CD8⁺ T cells in mice receiving LVRNA001 inoculations was elevated compared to that in mice receiving inactivated vaccine (Fig. 3C–D). These findings imply that LVRNA001 effectively promotes the Th1 cellular immune response in mice.

ELISAs revealed that at 14 days post-immunization (dpi), the IgG levels in the mice that received LVRNA001 were above the cutoff line and slightly higher than those in mice immunized with inactivated vaccine (Fig. 4A). The overall neutralizing antibody levels at 14 dpi reflected the same pattern (Fig. 4B). Mice were then intracerebrally challenged with RABV, and the survival rates (14 days post-challenge) were evaluated (Fig. 4C). We found that administering two doses (0d/7d) of LVRNA001 at either 5 μg or 1 μg resulted in 100% of the mice surviving the challenge, whereas a single injection (0d) of LVRNA001 at either 5 μg or 1 μg resulted in only 77.8% and 55.6% survival, respectively. In addition, the survival rates were also evaluated in the mice that received inactivated vaccine. The results showed that 22.2% of these mice injected once (0d) survived, while 66.7% those injected twice (0d/7d) survived. All mice in the PBS group died within 5 days. At 21 days post-challenge, the neutralizing antibody levels in the surviving mice were measured. Both LVRNA001 and the inactivated vaccine induced high antibody levels (≥ 0.5 IU/ml), indicating potential humoral responses and protective efficacy (Fig. 4D). Additionally, indirect immunofluorescence did not show the presence of RABV in the brain tissue of mice receiving two inoculations (0d/7d) of 5 μg of LVRNA001, while the virus was detected in PBS-treated mouse brain tissue (Fig. 4E). These data indicate that LVRNA001-induced antibodies can efficiently prevent viral replication in mice.

Determining the dosing interval for a two-dose vaccine regimen is critical. We therefore compared the serological responses and survival rates of the mice inoculated with a second dose of LVRNA001 at 3 days, 7 days, and 14 days after their first immunization (Fig. 5). Regardless of whether the dosages were larger (2.5 μg, Fig. 5A) or smaller (0.625 μg, Fig. 5B), the neutralizing antibody levels at 28 dpi and 84 dpi were significantly increased in the 14-day interval group versus the 3-day interval and 7-day interval groups. However, no significant difference between any of the groups was observed at 14 dpi. The survival rates of the mice at 14 days post-challenge were also observed to be higher in mice immunized a second time after 14 days than that of the mice immunized again after 3 or 7 days (Fig. 5C).

Dogs were intramuscularly immunized with LVRNA001 (5 μg or 25 μg dosages) for either two doses (0d/7d) or three doses (0d/7d/21d). Antibody response following pre-exposure immunization against RABV was measured. For all dosage and interval combinations, the induced neutralizing antibody levels exceeded 0.5 IU/ml at 9, 11, 13, and 35 dpi (Fig. 6A). Dogs vaccinated with 25 μg of LVRNA001 (0d/7d/21d) induced the production of more neutralizing antibodies than dogs immunized with inactivated vaccine (0d/3d/7d/14d/28d) at 35 dpi; no neutralization activity was detected in the PBS group throughout the experiment (Fig. 6A). These dogs were then infected with 50-fold LD₅₀ of the virus by i.m. injection at 35 dpi, and the survival rates (3 months post-challenge) were compared. Dogs that received 2 doses or 3 doses of LVRNA001 (5 μg or 25 μg, 0d/7d or 0d/7d/21d) demonstrated 100% survival (Fig. 6B), as did the dogs that received 5 doses of inactivated vaccine (0d/3d/7d/14d/28d). In contrast, all dogs in the PBS group died by end of the 3-month observation period (Fig. 6B). Neutralizing antibody levels were measured in the surviving dogs at 3 months post-challenge, the results showed that all four LVRNA001-immunized groups and the inactivated vaccine group sustained high levels of antibody production (Fig. 6C).

To further assess effect of LVRNA001 on protective efficiency in dogs, post-exposure immunization against RABV was conducted. Dogs were first infected (i.m.) with 50-fold LD₅₀ of the virus, six hours later, inoculated twice (0d/7d) or three times (0d/7d/21d) with LVRNA001 (25 μg or 5 μg dosages). Three months post-infection, the survival rates of the dogs immunized with 25 µg of LVRNA001 (0d/7d or 0d/7d/21d) were 100%, and the survival rates of the 5 μg dosage groups were 83.33%. In contrast, only 33.33% of dogs vaccinated with inactivated vaccine (0d/3d/7d/14d/28d) survived, and all dogs given PBS died (Fig. 7A). Analysis of the neutralizing antibody levels 3 months post-challenge found that all immunized dogs sustained high neutralizing antibody levels (> 0.5 IU/ml) (Fig. 7B).