Abstract

   Hemorrhagic fever with renal syndrome (HFRS) occurs throughout Eurasia
   with considerable morbidity and mortality. Currently, the absence of
   specific treatments or effective antiviral drugs for hantavirus
   infection makes developing safe and effective vaccines a high priority.
   Here, we report the development of three novel nucleic acid vaccine
   candidates, mRNA, naked DNA, and DNA encapsulated in lipid
   nanoparticles, encoding the glycoproteins of the Hantaan virus (HTNV).
   To comprehensively evaluate the potential of candidate HTNV nucleic
   acid vaccines in preventing HFRS, we focus on evaluating their
   immunogenicity and efficacy in mice and comparing them with an
   inactivated vaccine as the benchmark. Our findings reveal that all
   candidate vaccines activated instant and sustained immune responses,
   offering comparable in vivo protective efficacy to the inactivated
   vaccines. Notably, compared to the inactivated vaccine, mRNA vaccine
   induced stronger virus-specific T-helper 1 cell immune response, while
   DNA-LNP elicited higher levels of neutralizing antibodies in mice.
   These results mark a significant step in developing nucleic acid
   vaccines for HTNV, suggesting that sequential immunization with DNA and
   mRNA vaccines could further amplify the advantages of nucleic acid
   vaccines.

   Subject terms: RNA vaccines, DNA vaccines

Introduction

   Old World hantaviruses are predominantly found in Europe and Asia,
   where they usually cause hemorrhagic fever with renal syndrome (HFRS),
   resulting in symptoms including fever, bleeding tendency, and renal
   impairment^[64]1–[65]4. Rodents are the primary natural hosts of
   hantaviruses. Human infection is primarily through contact infection or
   aerosolized excreta inhalation^[66]2,[67]5. HFRS leads to considerable
   morbidity and mortality and thus represents a significant public health
   concern. Infections caused by the Hantaan virus (HTNV) strain of HFRS
   are especially severe, with reported mortality rates ranging from 5% to
   15%^[68]6,[69]7.

   Currently, there are no specific treatment methods or effective
   antiviral drugs for hantavirus infection^[70]8. Therefore, vaccination
   is the most viable strategy for disease prevention. In response to
   HFRS, both China and South Korea have authorized the use of inactivated
   vaccines developed either from mouse brain or cell cultures^[71]9.
   These vaccines demonstrate strong immunogenicity and are proven
   effective in reducing the incidence of HFRS^[72]10. However, existing
   studies indicate that inactivated vaccines require multiple doses to
   achieve sufficient protection, and their safety in the elderly and
   children still requires further validation^[73]11,[74]12. The parallel
   development of multiple vaccine technologies offers more choices, which
   is important for enhancing the ability to tackle new pathogens and
   their variants. Besides, it is believed various vaccine types can meet
   the specific needs of diverse demographic groups, and respond rapidly
   to global health emergencies^[75]13. Hantavirus infection has been
   noted to fluctuate periodically, and recent observations suggest an
   expansion in the range of its natural hosts^[76]14,[77]15.
   Consequently, HFRS is garnering increasing attention and underscores
   the importance of developing new, safe, and effective vaccines against
   Hantavirus.

   Nucleic acid vaccines, including DNA and mRNA vaccines, can enter human
   cells and produce endogenous antigens like viruses. By presenting
   antigens to the immune system via dual pathways, nucleic acid vaccines
   activate both cellular and humoral immune responses for enhanced
   protection^[78]16. And nucleic acid vaccines have become a swift and
   flexible platform, particularly valuable in emergency response
   scenarios^[79]17,[80]18. mRNA vaccine represents a significant
   advancement in nucleic acid vaccine technology^[81]19. With the
   clinical application of mRNA vaccines in the prevention of COVID-19,
   mRNA vaccine has demonstrated significant potential in combating
   various virulent pathogens^[82]20–[83]23. Yet, the efficacy and
   suitability of mRNA vaccine in addressing HTNV is still an area of
   uncertainty. Notably, the advancement of mRNA vaccine technology has
   benefited from pseudouridine modification and the delivery system of
   lipid nanoparticles (LNP). Recent studies indicate that LNP is crucial
   in maintaining vaccines at the injection site, allowing for prolonged
   antigen presentation and immunostimulation of innate immune
   cells^[84]24. While LNP as effective antigen-delivery systems have been
   thoroughly exploited in the field of mRNA vaccines, further research is
   needed to assess their potential for DNA vaccination against various
   infectious agents. Leveraging the proven potential of nucleic acid
   vaccine technology in combating viral infections, the pursuit of
   developing a safe and effective vaccine specifically for Hantavirus is
   paramount, which would be a valuable addition against this serious
   viral threat.

   Hantavirus is classified within the group of single-stranded
   negative-sense RNA viruses^[85]7. Similar to other members of the
   Bunyaviridae family, its viral envelope glycoprotein (GP) plays a
   critical role in the virus’s attachment to and entry into host cells
   and serves as the primary target for the immune system to recognize and
   neutralize the virus^[86]25. The envelope GP of Hantavirus is currently
   reported with multiple neutralizing epitopes which induce the
   production of neutralizing antibodies to resist viral
   attacks^[87]26,[88]27. Consequently, the HTNV GP is a preferred
   candidate for HFRS vaccine development. In the present research, we
   report the development of three novel candidate nucleic acid vaccines
   targeting the GP and evaluate the immunogenicity of the candidate
   vaccines from humoral immune response, cellular immune response, and in
   vivo protection.

Result

Construction and identification of candidate vaccines

   M segment of HTNV 76–118 strain was selected as the target antigen
   encoding sequence for three nucleic acid vaccines. We successfully
   constructed the DNA candidate vaccine pVAX1-GP[HTNV] and the vector
   pGEM-UTR-GP[HTNV] for mRNA in vitro transcription. All relevant vectors
   were verified by Sanger sequencing and enzyme digestion electrophoresis
   (Fig. [89]1a). Here, we designed an mRNA platform encoding the
   full-length HTNV GP protein (Fig. [90]1b). The N-terminus of this
   protein carries an HTNV GP signal peptide for endoplasmic reticulum
   translocation and secretion. The desired mRNA molecules were
   synthesized through in vitro transcription of the pGEM-UTR-GP[HTNV]
   vector, followed by further capping modification. Microfluidic
   capillary electrophoresis of RNA revealed a single peak at 4000 nt,
   indicating its high purity and integrity (Fig. [91]1c). And the dsRNA
   content in the mRNA-GP[HTNV] vaccine candidate was 0.07%. We developed
   a vaccine platform based on DNA encapsulated in LNP (DNA-LNP) for in
   vivo delivery (Fig. [92]1d). DNA (pVAX1-GP[HTNV])-LNP, manufactured
   using a microfluidic flow-focusing strategy, exhibited an average
   particle size of 98.83 nanometers (Fig. [93]1e) with an encapsulation
   efficiency exceeding 70% (Table [94]S4). Transmission electron
   microscopy (TEM) analysis revealed that DNA-LNP particles exhibited a
   uniform solid spherical morphology (Fig. [95]1f). mRNA-GP[HTNV] is
   encapsulated into in vivo-jetRNA, forming mRNA-lipo[jet]. The average
   particle size of the mRNA-lipo[jet] was 144.57 nanometers (Fig.
   S[96]2a), with an encapsulation efficiency of 76% (Table [97]S4). TEM
   shows that these particles are uniformly shaped like solid spheres
   (Fig. S[98]2b). During the expression of HTNV GP, its N-terminal Gn
   undergoes autophagy^[99]28,[100]29. Hence, only the expression of Gc
   could be detected by western blot analysis in transfected cells. Major
   protein bands containing Gc were detected, with a molecular weight of
   approximately 61 kDa (Fig. [101]1g and Fig. S[102]3). We confirmed
   transcription of DNA and transfection of mRNA through qPCR, and the
   transcription level between Gn and Gc exhibits no substantial
   differences (Fig. [103]1h). HTNV GP was localized to the cytoplasm in
   transfected cells, as observed by immunofluorescence microscopy (Fig.
   [104]1i).

Fig. 1. Vaccine design and characterization of the expressed antigens.

   [105]Fig. 1
   [106]Open in a new tab

   a Plasmids (pVAX1-GP[HTNV], and pGEM-UTR-GP[HTNV]) identified by 1%
   agarose gel electrophoresis. b Structure of mRNA vaccine expressing the
   HTNV GP. UTR, untranslated region; SP, signal peptide. c Liquid
   capillary electropherograms of in vitro-transcribed mRNA. Peaks
   represent mRNA of HTNV GP (4000 nt) and an indicative marker (20 nt). d
   Schematic diagram of DNA-LNP, encoding the full-length GP protein of
   HTNV. e Representative intensity-size graph of HTNV DNA-LNP measured by
   dynamic light-scattering method. f TEM image of HTNV DNA-LNP. g GP
   protein expression in HEK293T cells. Cells were transfected with HTNV
   GP-encoding DNA or mRNA (2 μg/mL), and immunoblotting was performed at
   24 h after transfection. The GP protein was detected with a rabbit
   monoclonal antibody against the flag. h qRT-PCR analyses for HTNV Gn
   and Gc expression following transfection of HTNV GP-encoding DNA or
   mRNA (2 μg/mL) into HEK293T cells, β-actin was used as control. i
   Immunofluorescence assay of GP protein expression in HEK293T cells 24 h
   post-transfection. GP (red) and DNA (blue). Scale bar, 30 μm.

Candidate vaccines induced humoral immune protection in mice

   We conducted experiments on a group of immunocompetent female BALB/c
   mice to assess the immunogenicity and effectiveness of candidate
   vaccines. Following the strategy illustrated in Fig. [107]2a, the mice
   received two doses of either inactivated vaccine, naked DNA, DNA-LNP,
   or mRNA-lipo[jet], administered via intramuscular injection. PBS and
   empty LNP served as control groups. The instant immune responses to
   each vaccine group were evaluated accordingly. In addition, no local
   inflammatory or other adverse reactions were observed at the injection
   sites during the observation period post-immunization. HTNV-specific
   IgG antibodies in the sera of the immunized mice were quantified using
   ELISA. Following both prime and boost immunizations, four candidate
   vaccine groups successfully induced the generation of antigen-specific
   IgG antibodies. Serum samples from mice immunized with PBS and empty
   LNP did not exhibit specific IgG detection. Notably, the group
   receiving the DNA-LNP vaccines demonstrated significantly higher
   specific IgG titers compared to the inactivated vaccines after the
   initial immunization (Fig. [108]2b).

Fig. 2. Schematic diagram of immunization and evaluation of humoral immune
responses.

   [109]Fig. 2
   [110]Open in a new tab

   a Schematic diagram of immunization, sample collection, and challenge
   schedule (the prone mouse element was created by Figdraw). b
   HTNV-specific IgG antibody titers were determined by ELISA (n = 8). The
   red dashed lines indicate the limit of detection (LOD). The LOD of the
   HTNV-specific antibody ELISA is 100. c 50% virus-neutralization titers
   determined using infectious HTNV (n = 6) and the LOD of the HTNV
   neutralization assay is 10. Data are represented as the mean ± SEM.
   Statistical differences were analyzed using one-way ANOVA with multiple
   comparison test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

   Neutralizing antibody levels serve as a pivotal indicator of HTNV
   vaccine efficacy. We conducted HTNV-neutralizing antibody tests on sera
   from mice immunized using a fixed virus (100 plaque forming units
   (PFUs)) serum dilution method. The results showed that after two doses
   of vaccine immunization, the PBS group and the empty LNP group induced
   average neutralizing titers of 1:11 and 1:13, respectively. Meanwhile,
   the inactivated vaccine, naked DNA, DNA-LNP, and mRNA-lipo[jet] groups
   induced mean neutralizing titers of 1:79, 1:75, 1:118 and 1:65,
   respectively. Significantly, the HTNV-neutralizing antibody titers in
   sera from mice vaccinated with the four candidate vaccine groups were
   significantly higher than the control group. Furthermore, the DNA-LNP
   group induced significantly higher neutralizing antibodies than the
   inactivated vaccine group. Compared to naked DNA, the DNA-LNP group
   demonstrated a significant increase in neutralizing antibody titers,
   indicating that LNP-mediated delivery enhances the immune response for
   the same DNA dosage (Fig. [111]2c). Overall, the sera from mice
   immunized with the four candidate vaccines effectively neutralized the
   infectious HTNV and initiated antiviral protective immunity, featured
   as the DNA-LNP yielding the highest serum neutralizing titers.

Candidate vaccines elicited T cell responses in mice

   T cells play a crucial role in immune protection against hemorrhagic
   fever viruses. Therefore, we investigated the T cell responses induced
   by two doses of each candidate vaccine through intramuscular injection
   and whether these immunizations elicited HTNV-specific T cell immune
   responses in mice. 24 days after booster vaccination, we characterized
   the antigen-specific responses of splenocytes in immunized mice. Upon
   stimulation with a mixture of ten HTNV GP 15-mer peptides, as
   determined by ELISpot assay, four candidate vaccine groups showed a
   significant increase in HTNV GP-specific T lymphocytes secreting IFN-γ
   compared to the control group, indicating the successful induction of
   antiviral T-helper 1 (Th1) cell-mediated immune responses. Furthermore,
   naked DNA, DNA-LNP, and mRNA-lipo[jet] vaccines induced significantly
   higher levels of IFN-γ-secreting T lymphocytes than the inactivated
   vaccine. DNA-LNP resulted in significantly high IFN-γ levels in the
   spleen in comparison with the naked DNA vaccine, suggesting that DNA
   vaccine-induced T cell immune responses were further enhanced under LNP
   delivery (Fig. [112]3a and Fig. S[113]4a). Furthermore, ELISpot
   experiments revealed a significant increase in the secretion of
   interleukin-4 (IL-4) in the splenocytes of mice immunized with
   inactivated vaccine, DNA-LNP, and mRNA-lipo[jet] vaccines, as compared
   to mice receiving the control vaccines. And mRNA-lipo[jet] vaccine
   induced a significantly higher number of IL-4-secreting T lymphocytes
   compared to the inactivated vaccine (Fig. [114]3b). Our results
   indicate that naked DNA, DNA-LNP, and mRNA-lipo[jet] vaccines
   successfully induced Th1-biased, HTNV-specific cellular immune
   responses. Yet the inactivated vaccine did not exhibit a clear Th1/Th2
   bias after two doses of immunization (Fig. [115]3c). We performed
   single peptide verification on ten GP 15-mer peptides by ELISpot, and
   the results revealed that stimulation with the 15-mer peptides
   IALGPYRVQVVYERS, SYCMTGVLIEGKCFV, and KKVMATIDSFQSFNT notably enhanced
   the IFN-γ secreting T lymphocyte response across all immunization
   groups (Fig. [116]3d). Additionally, stimulation with KKVMATIDSFQSFNT
   led to a remarkable increase in the IL-4 secreting T lymphocyte
   response in all groups. In summary, all ten 15-mer peptides could
   induce splenocyte responses specific to GP, qualifying as GP-specific
   dominant peptides for evaluating GP-based HTNV vaccine efficacy.

Fig. 3. Analysis of T cells after immunization.

   [117]Fig. 3
   [118]Open in a new tab

   a HTNV GP-specific IFN-γ spot-forming units (SFCs) determined by
   ELISpot (n = 8). b HTNV GP-specific IL4 SFCs determined by ELISpot
   (n = 6). c Scatter plot showing the correlation of HTNV GP-specific
   IL-4 and IFN-γ secreting cells by ELISpot (n = 6). d IFN-γ and IL-4
   SFCs stimulated by single 15-mer peptide of HTNV GP by ELISpot assay
   after the second immunization and presented as a heatmap of Z scores
   (n = 8). e Gating strategy to identify cytokine secretion (take the
   IL-2 release from CD4^+ T cells for example). f IFN-γ, IL-2, IL-4 or
   Granzyme B release from CD4^+ and CD8^+ T cells by flow cytometry
   (n ≥ 5). g Gating strategy to identify effector memory T (Tem) cells
   (CD44^+ CD62L^-) in the spleen (take the CD4^+ Tem cells in splenocytes
   for example). h HTNV GP-specific CD4^+ (n ≥ 6) and CD8^+ (n ≥ 4) Tem
   cells in splenocytes were detected by flow cytometry. Data are
   represented as the mean ± SEM. Statistical differences were analyzed
   using one-way ANOVA with multiple comparison test. *P < 0.05,
   **P < 0.01, ***P < 0.001, ****P < 0.0001.

   Further confirmation of Th1-type cell-mediated immune response was
   obtained by flow cytometry, which detected CD4^+ and CD8^+ T
   lymphocytes secreting the cytokines IFN-γ or IL-2 in immunized mice
   following mixed stimulation (Fig. [119]3f). The results showed that
   splenocytes from mice immunized with the four candidate vaccine groups
   exhibited a significant increase in the secretion of both IFN-γ and
   IL-2 by CD4^+ and CD8^+ T cells compared to the control mice. The
   cellular response patterns for naked DNA and DNA-LNP candidate vaccines
   were similar, while compared to mice immunized with naked DNA, mice
   immunized with DNA-LNP produced stronger secretion of IL-2 by CD8^+ T
   cells (Fig. [120]3f). Furthermore, compared to the control group, mice
   immunized with inactivated vaccine and mRNA-lipo[jet] showed a
   significant increase in the IL-4 secretion of CD4^+ and CD8^+ T cells
   (Fig. [121]3f). Additionally, CD8^+ T cells secreting IL-4 were
   significantly increased in mice immunized with DNA-LNP when compared to
   the control group. Compared to mice receiving the control group, mice
   from four candidate vaccine groups showed a significant increase in
   granzyme B secretion by CD4^+ and CD8^+ T cells, indicating increased
   cytotoxic activity. Additionally, CD8^+ T cells secreting granzyme B
   were significantly increased in the mRNA-lipo[jet] vaccine group
   compared to the inactivated vaccine group (Fig. [122]3f). The flow
   cytometry results showed that as compared to the control group,
   splenocytes from immunized mice in all experimental groups exhibited a
   significant increase in virus-specific CD4^+ effector memory T (Tem)
   cells following stimulation with the HTNV GP peptide pool (Fig.
   [123]3h). Moreover, a significant increase in CD8^+ Tem cells was
   observed in the inactivated vaccine, naked DNA and mRNA-lipo[jet]
   vaccine immunized mice compared with PBS immunized mice.

Candidate vaccines conferred protective efficacy in vivo

   To further evaluate in vivo protective efficacy, we assessed the
   immune-mediated in vivo protective effect using the HTNV 76–118 strain
   challenge model. On the day of euthanasia, organs (heart, liver,
   kidney, lung, and spleen) were collected to determine viral loads,
   revealing HTNV replication in all collected organs. Notably, the levels
   of viral RNA were significantly reduced in the hearts and livers of all
   mice immunized with the four candidate vaccines compared to those
   receiving the control group (Fig. [124]4a). In addition, the spleen
   viral RNA levels of the DNA-LNP and mRNA-lipo[jet] vaccine groups were
   significantly lower than those of the control group (Fig. [125]4a).
   Furthermore, immunohistochemical analysis revealed abundant expression
   of HTNV NP protein in the heart and liver tissues of mice receiving PBS
   and LNP, whereas minimal positive cells were detected in the hearts and
   livers of mice immunized with the candidate vaccines (Fig. [126]4b).
   These results suggest that two doses of vaccination can prevent HTNV
   replication in vivo and indicate that three nucleic acid vaccines
   exhibited in vivo protective efficacy comparable to inactivated
   vaccines. Based on the viral copy number analysis of the spleen, the
   groups of DNA-LNP and mRNA-lipo[jet] vaccine induce a better protective
   immune response.

Fig. 4. Candidate vaccines protect mice from HTNV challenge.

   [127]Fig. 4
   [128]Open in a new tab

   a Viral RNA copies in the main organs (heart, liver, kidney, lung, and
   spleen) of HTNV-infected mice were determined by qRT-PCR (n = 6). b
   Immunostaining results for HTNV NP protein in the heart and liver
   tissues (n = 3). Scale bar, 100 μm. Data are represented as the
   mean ± SEM. Statistical differences were analyzed using one-way ANOVA
   with multiple comparison test. *P < 0.05, **P < 0.01, ***P < 0.001.

Candidate vaccines induced robust immunoreactive transcriptomes

   To investigate the potential mechanisms of inactivated vaccine, naked
   DNA, DNA-LNP, and mRNA-lipo[jet] immunization strategies, we
   characterized the transcriptome by performing RNA-seq on mouse
   splenocyte samples after two doses of different vaccine regimens
   following HTNV peptide stimulation. The pathway enrichment analysis,
   utilizing both Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene
   Ontology (GO) frameworks, revealed significant insights into the
   immunological responses across all immunization groups. The KEGG-based
   analysis underscored the enrichment of pathways involved in immune
   activation, lymphocyte function, Th sub-population differentiation, and
   the regulation of inflammatory responses, etc. For instance, the Th1
   and Th2 cell differentiation pathway increased in the inactivated and
   mRNA-lipo[jet] vaccines groups compared to the PBS group. The
   inactivated vaccine, naked DNA, DNA-LNP, and mRNA-lipo[jet] exhibited
   an increase in the B cell receptor signaling pathway compared to the
   control group. Additionally, the DNA-LNP group showed an increase in
   the B cell receptor signaling pathway compared to the naked DNA group
   (Fig. [129]5a). Concurrently, the GO-based analysis pointed to a
   pronounced enrichment of pathways associated with innate and adaptive
   immunity, T cell-mediated cellular immunity together with cytokines
   secretion, and B cell-mediated humoral immunity. It showed that all
   four candidate vaccine groups exhibited an increase in T cell
   activation, cytokine-mediated signaling, cellular response to INF-γ,
   and B cell proliferation pathway. Furthermore, the inactivated vaccine,
   DNA-LNP, and mRNA-lipo[jet] groups showed an increase in the cellular
   response to IL-4 pathway (Fig. [130]5b). Volcano plots highlighted the
   relative increases in the expression of selected immune-related genes
   between each two groups of all four candidate vaccines and control
   (Fig. [131]5c). Immune component analysis of mouse splenocytes was
   performed using ssGSEA and mMCPcounter algorithms. Clustering analysis
   based on cell types revealed a distribution pattern as shown in Fig.
   [132]5d and Fig. S[133]5. Both ssGSEA and mMCPcounter analysis
   indicated a sequence of PBS, naked DNA, inactivated vaccine, DNA-LNP,
   and mRNA-lipo[jet], roughly indicating a trend of increasing immune
   cell activation. In summary, all four candidate vaccine regimens
   effectively activated the immune response in vivo and induced robust
   immunoreactive transcriptomes.

Fig. 5. Transcriptome analysis for different vaccination strategies.

   [134]Fig. 5
   [135]Open in a new tab

   a Enriched KEGG immune pathways between each two groups of all four
   candidate vaccination strategies and control. b Enriched GO immune
   pathways between each two groups of all four candidate vaccination
   strategies and control. c Volcano plots of Log2 fold change of the
   differential expression between each two groups of all four candidate
   vaccines and control. (log2 fold change cut-off value at ±1, P < 0.05).
   d Immune component analysis using ssGSEA algorithm.

Candidate vaccines elicited long-term immune response in mice

   To evaluate the long-term efficacy of the immune response, we adhered
   to the protocol outlined in Fig. [136]6a. Six months after the initial
   immunization, each group was administered the same dose of booster. The
   antibody response dynamics demonstrated a significant surge in specific
   antibody titers by day 14 following each vaccination, reaching their
   peak levels. Furthermore, all candidate vaccines sustained high
   antibody titers up to five months after two vaccination doses,
   indicating prolonged immunogenicity (Fig. [137]6b). After the final
   immunization, four candidate vaccine groups successfully induced the
   production of antigen-specific IgG antibodies (Fig. [138]6c). Following
   the administration of a long-term booster, we conducted
   HTNV-neutralizing antibody assays on sera collected from immunized
   mice. The results revealed that after receiving the vaccine booster,
   the PBS group and the empty LNP group induced average neutralizing
   titers of 1:9 and 1:7, respectively. The inactivated vaccine, naked
   DNA, DNA-LNP, and mRNA-lipo[jet] groups generated average neutralizing
   titers of 1:87, 1:104, 1:113 and 1:89, respectively (Fig. [139]6d). In
   comparison to the immediate immune response, the average neutralizing
   antibody titers in the long-term immune response were higher in groups
   receiving the inactivated vaccine, naked DNA, and mRNA-lipo[jet] boost
   vaccines. This underscores the importance of long-term booster shots in
   strengthening and prolonging immune protection. All candidate vaccine
   groups effectively induced HTNV-neutralizing antibodies, with antibody
   levels significantly higher than those in mice immunized with PBS or
   empty LNP controls (Fig. [140]6d). Besides, no significant differences
   were observed among the experimental groups, indicating that the
   humoral response of nucleic acid vaccines is not inferior to that of
   clinically used inactivated vaccines.

Fig. 6. Schematic diagram of long-term immunization and evaluation of immune
responses.

   [141]Fig. 6
   [142]Open in a new tab

   a Schematic diagram of immunization, sample collection, and challenge
   schedule (the prone mouse element was created by Figdraw). b Dynamic
   HTNV-specific antibody response following vaccination (n ≥ 7) (arrows
   below the x-axes indicate the day of vaccine injection). c
   HTNV-specific IgG antibody titers were determined by ELISA (n = 7). The
   red dashed lines indicate the limit of detection (LOD). The LOD of the
   HTNV-specific antibody ELISA is 100. d 50% virus-neutralization titers
   determined using infectious HTNV (n = 6) and the LOD of the HTNV
   neutralization assay is 10. e Long-term HTNV GP-specific IFN-γ SFCs
   determined by ELISpot (n = 6). f Long-term HTNV GP-specific IL4 SFCs
   determined by ELISpot (n = 6). g Scatter plot showing the correlation
   of long-term HTNV GP-specific IL-4 and IFN-γ secreting cells by ELISpot
   (n = 6). h IFN-γ and IL-4 SFCs stimulated by single 15-mer peptide of
   HTNV GP by ELISpot assay after the long-term immunization and presented
   as a heatmap of Z scores (n = 3). i Long-term HTNV GP-specific CD4^+
   (n ≥ 3) and CD8^+ (n ≥ 3) Tem cells in splenocytes were detected by
   flow cytometry. Data are represented as the mean ± SEM. Statistical
   differences were analyzed using one-way ANOVA with multiple comparison
   test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

   To assess the durability of induced T cell responses, we compared the
   characterization of mice splenocytes three weeks after the final
   immunization. ELISpot assays revealed a significant increase in
   virus-specific T lymphocytes secreting IFN-γ in the spleens of
   immunized mice compared to the control group, which was similar to the
   instant cellular responses (Fig. [143]6e and Fig. S[144]4b).
   Additionally, ELISpot experiments demonstrated that, compared to mice
   receiving the control group, all immunized groups led to significant
   induction of IL-4 levels in the mice splenocytes, with the inactivated
   vaccine showing the most significant increase (Fig. [145]6f).
   Furthermore, T lymphocytes secreting IL-4 in the spleen of mice
   immunized with naked DNA, DNA-LNP, and mRNA-lipo[jet] vaccines were
   significantly lower than those in the inactivated vaccine group,
   suggesting advantage in IL-4 secretion in Th response to the
   inactivated vaccine under long-term immunity. Our results indicate that
   naked DNA, DNA-LNP, and mRNA-lipo[jet] vaccines successfully induced
   predominantly Th1-driven long-lasting HTNV-specific cellular immune
   responses, while the inactivated vaccine exhibited a Th2-biased
   response under long-term immunity (Fig. [146]6g). Following long-term
   immunization, epitope prediction analysis was performed by conducting
   single peptide validations for ten peptides. The results demonstrated
   that the 15-mer peptides SYCMTGVLIEGKCFV and KKVMATIDSFQSFNT elicited
   strong T cell responses across all immunized groups (Fig. [147]6h).
   Besides, SYCMTGVLIEGKCFV and KKVMATIDSFQSFNT consistently presented as
   dominant peptides in both instant and long-term responses. Flow
   cytometry results revealed a significant increase in virus-specific
   CD4^+ Tem cells in the splenocytes of immunized mice compared to the
   control group, with no significant difference among the experimental
   groups. (Fig. [148]6i). In addition, a significant increase in CD8^+
   Tem cells was observed in naked DNA and mRNA-lipo[jet] immunized mice
   compared with PBS immunized mice (Fig. [149]6i).

Preliminary safety assessment of candidate vaccines

   We collected seven organs (heart, liver, kidney, spleen, lung,
   cerebrum, and gastrocnemius muscle) from the mice after final
   immunizations, three weeks apart, to preliminarily assess the safety of
   the vaccines in long-term effects. No significant pathological
   differences were observed in the tissues between the four candidate
   vaccine groups and the control groups (Fig. [150]7). These results
   confirm that immunization with the four candidate vaccines in mice did
   not result in pathological damage, providing preliminary evidence of
   vaccine safety.

Fig. 7. H&E analyses of major organ tissues after immunization.

   [151]Fig. 7
   [152]Open in a new tab

   Scale bar, 150 μm (gastrocnemius muscle, heart, liver, and kidney);
   300 μm (lung and spleen); and 600 μm (cerebrum).

Discussion

   Since the first isolation of the HTNV in 1978, research into vaccines
   has been relentless^[153]30–[154]32. The current HFRS vaccines in use
   are only inactivated vaccines. Additionally, DNA and subunit protein
   vaccines for HFRS are under study and show promise as effective
   preventive strategies against Hantavirus infection^[155]33,[156]34.
   However, global recognition is still pending, necessitating further
   research and development for safer and more effective vaccines to
   control HFRS, especially in areas with high incidence rates. Our study
   focuses on the development of three nucleic acid vaccine candidates
   (naked DNA, DNA-LNP, and mRNA-lipo[jet]), evaluating their
   immunogenicity and efficacy in Balb/C mice and comparing them with an
   inactivated vaccine as a benchmark. Our findings reveal that all
   candidate vaccines activated both instant and sustained immune
   responses, offering effective in vivo protection. Notably, the nucleic
   acid vaccines demonstrated distinct advantages over the inactivated
   vaccine, mainly reflected in terms of the enhanced humoral immunity
   observed with the DNA vaccine and the stronger cellular immunity with
   the mRNA vaccine. Furthermore, the DNA-LNP and mRNA-lipo[jet] vaccines
   elicited stronger splenocyte cytotoxicity and protective immunity.
   These results mark a significant step in the development of nucleic
   acid vaccines for HTNV, suggesting that sequential immunization with
   DNA and mRNA vaccines could further amplify the advantages of nucleic
   acid vaccines.

   In recent years, the development of mRNA vaccine technology has been a
   transformative breakthrough in vaccine research, offering substantial
   advantages over traditional methods^[157]35,[158]36. The 2023 Nobel
   Prize in Physiology or Medicine was awarded to the mRNA vaccine
   development for its critical significance in COVID-19. Significantly,
   this research heralds the inaugural creation of an mRNA vaccine
   targeting the hemorrhagic fever virus, effectively triggering
   virus-specific humoral and cellular immune responses, and ensuring
   efficacious protection in vivo. Compared with the current gold standard
   inactivated vaccine for HFRS, it induced stronger virus-specific
   Th1-type cellular immune responses, stronger splenocyte cytotoxic
   activity, and comparable protective immunity in mice. This study not
   only confirms the effectiveness of mRNA vaccines in preventing HFRS but
   also indicates their promising potential in combatting other hantavirus
   infections.

   Various delivery methods such as electroporation, gene gun, or tail
   tattoo have shown effectiveness in enhancing the immune response of DNA
   vaccines in large animals and humans^[159]37–[160]39. Their practical
   application in widespread human vaccination programs remains a
   challenge. In contrast, LNP has been extensively explored as an
   antigen-delivery system^[161]36,[162]40–[163]42. Existing studies on
   LNP as antigen delivery systems for DNA vaccines have primarily
   employed cationic lipids, known for their affinity to bind with
   negatively charged nucleic acids^[164]43–[165]46. However, delivery
   systems based on cationic lipids exhibit toxicity and immunogenicity
   both in vitro and in vivo^[166]47. Ionizable cationic lipids, designed
   to lessen the toxicity of lipid nanoparticles while preserving
   transfection efficiency, have been primarily applied in mRNA
   vaccines^[167]17. The exploration of their role in DNA vaccine delivery
   and the augmentation of immune responses remains limited. Pfeifle et
   al.’s pioneering work on a Lyme disease DNA vaccine highlights the
   potential of the DNA-LNP platform for vaccine development^[168]48.
   However, there was no direct comparison of DNA-LNP with naked DNA
   vaccines of equivalent copy numbers. In this study, ionizable lipid
   nanoparticles were used to encapsulate the HTNV DNA vaccine, and for
   the first time, evaluated against a naked DNA vaccine of the same copy
   number. While the overall immune response patterns of the DNA-LNP and
   the naked DNA vaccines were similar, DNA-LNP notably induced a stronger
   instant cellular response. GO and KEGG pathway enrichment analysis
   revealed that the DNA-LNP group showed various increased immune-related
   pathways compared to the naked DNA group. Regarding long-term cellular
   immune responses, there was an observed increase in virus-specific T
   lymphocytes secreting IFN-γ in the DNA-LNP group, albeit without a
   significant statistical difference from the naked DNA group. The
   advantage of LNP lies in their capacity to prolong the half-life upon
   entering the body, serving as a delayed-release mechanism and
   potentially reducing the frequency of immunizations required.
   Therefore, the instant cellular immune response of DNA-LNP is stronger,
   and the difference may be evened out after long-term booster
   immunization. Further research is warranted to investigate the role of
   LNP in facilitating the nuclear entry of DNA vaccines and to assess the
   cost-effectiveness of this strategy.

   Upon injection into muscle tissue, DNA and mRNA vaccines are initially
   taken up by immune cells at the injection site. DNA vaccines typically
   require a delivery system to facilitate their entry through cellular
   and nuclear membranes into the cell nucleus. While cells can also take
   up naked DNA, this process is much less efficient^[169]49,[170]50. Once
   inside the cells, DNA vaccines remain longer than mRNA, resulting in
   prolonged antigen expression and a more enduring immune memory^[171]51.
   In contrast, mRNA vaccines are typically encapsulated within lipid
   nanoparticles and enter cells via endocytosis. The mRNA is then
   released directly into the cytoplasm and quickly translated into the
   target protein without the need to enter the nucleus. This allows mRNA
   vaccines to exhibit higher uptake efficiency than DNA vaccines, leading
   to a quicker and more robust immune response, though their presence
   within the cells is shorter-lived^[172]52,[173]53. This study
   represents the first comprehensive evaluation of the immunological
   efficacy differences among hemorrhagic fever virus DNA vaccines, mRNA
   vaccines, and traditional inactivated vaccines, revealing the
   advantages of each candidate nucleic acid vaccine compared to
   traditional inactivated vaccines. Our findings demonstrate that three
   candidate nucleic acid vaccines effectively elicited a Th1-dominant,
   HTNV-specific, long-term cellular immune response, while the
   inactivated vaccine exhibited a Th2-biased response after long-term
   immunization. Additionally, either of ELISpot, flow cytometry, and
   transcriptomic techniques indicated that the mRNA and inactivated
   vaccine groups prompted notably higher IL-4 secretion levels than the
   DNA vaccine groups, both instantly or in the long term. This suggests a
   more robust Th2-type cellular immune response triggered by the mRNA and
   inactivated vaccines. While inactivated vaccines exhibit a Th2 bias
   when compared to nucleic acid vaccines, indicating a stronger
   antibody-mediated immune response, this Th2 bias does not significantly
   enhance the advantages of inactivated vaccines in humoral immune
   responses. Our findings demonstrate that three candidate nucleic acid
   vaccines showed neutralizing antibody titers and in vivo protective
   efficacy not inferior to those of inactivated vaccines. Therefore, each
   candidate vaccine could induce effective protective immunity. Moreover,
   the instant cellular immune response to the inactivated vaccine was
   less potent compared to the three candidate nucleic acid vaccines, but
   it showed improvement following a long-term booster, suggesting the
   necessity of multiple doses of inactivated vaccines for adequate
   protection. The nucleic acid vaccine exhibited obvious dominant
   peptides, while the inactivated vaccine did not show consistent
   dominant peptides. This underscores the advantage of nucleic acid
   vaccines, which focus immune responses on the target antigen, thereby
   inducing more specific immune reactions. Furthermore, the research
   revealed that SYCMTGVLIEGKCFV and KKVMATIDSFQSFNT are key epitopes of
   GP antigen presentation and can serve as candidate epitopes for a
   Hantaan virus epitope-based vaccine.

   Post-translational processing of HTNV GP yields two glycoproteins, Gn
   and Gc. Both the Gn and Gc proteins contain binding sites for
   neutralizing antibodies, making the full-length GP a richer source of
   such sites, including both linear and conformational epitopes^[174]5.
   Using full-length GP as an immunogen aids in the formation and
   maintenance of conformational epitopes, which are crucial for inducing
   neutralizing antibodies^[175]54. Additionally, the full GP offers more
   T cell epitopes, which help activate the cellular immune response
   against the virus. However, during the expression of HTNV GP, the Gn
   terminus induces autophagy and undergoes rapid degradation, leaving
   only Gc stable within the cells^[176]28,[177]29,[178]55. Given that Gn
   also contains numerous neutralizing epitopes, concerns arise regarding
   the balance between uncertain expression levels and immunological
   efficacy. Our research group previously engineered DNA vaccines
   targeting the individual Gn and Gc components of HTNV for immunological
   assessments^[179]56–[180]58. In contrast, our international
   counterparts developed a DNA vaccine based on the full-length GP, but
   did not separately assess the immunological effects of Gn and Gc under
   full-length GP immunization^[181]59,[182]60. We evaluated T cell
   responses using the dominant epitope peptides screened for Gn and Gc.
   The performance of these dominant epitopes appears to exhibit no
   significant differences between Gn and Gc, indicating that Gn and Gc
   are equivalent in antigen presentation and eliciting T cell responses.
   Therefore, the non-detection of Gn expression does not compromise its
   immunogenic efficacy and the use of full-length GP does not raise
   concerns of response imbalance between Gn and Gc.

   Our histological evaluation preliminarily confirms the safety of the
   candidate nucleic acid vaccines. The clinical application of both DNA
   and mRNA vaccines demonstrates their acceptable
   tolerability^[183]61–[184]67. The long-term clinical use and trials to
   date suggest that there are no major safety concerns with these
   vaccines. Given that the immune response induced by the vaccines may
   vary depending on the mouse strain and virus type used in the
   challenge, further research is needed using a variety of mouse strains
   (e.g., HLA transgenic mice) and diverse viral infection models to
   confirm our findings. Overall, our research provides promising new
   insights into the development and application of HFRS vaccines for the
   effective control of HTNV epidemics.

Methods

Sequence design

   The template for the DNA candidate vaccine is a codon-optimized version
   of the HTNV 76–118 strain M segment, which encodes the HTNV envelope
   glycoprotein, including Gn and Gc, with a length of 3616 base pairs
   (Sequence ID [185]Y00386.1). The mRNA vaccine candidate encodes a
   codon-optimized version of the HTNV 76-118 strain M segment. Besides,
   it includes 5′ and 3′ untranslated regions (UTR) as well as a poly (A)
   tail, collectively referred to as the UTR region (Table [186]S1 and
   Fig. S[187]1). The 5′ UTR and 3′ UTR sequences are derived from homo
   sapiens hemoglobin subunit beta (Sequence ID: [188]NM_000518.5). The 5′
   UTR sequence comprises a regulatory sequence followed by the 5′ UTR
   sequence of β-globin, while the 3′ UTR sequence consists of the
   concatenated sequences of two segments from β-globin’s 3′ UTR.
   Additionally, a 120-nt long poly (A) tail is included to enhance RNA
   stability and translation efficiency. A 3×Flag tag is appended to the
   C-terminus of the GP to facilitate subsequent identification of vaccine
   antigen expression.

Cell culture

   HEK293T cells were cultured in Dulbecco’s modified Eagle medium (DMEM)
   supplemented with 10% fetal bovine serum (FBS) and 1%
   penicillin/streptomycin at 37 °C in a 5% CO[2] incubator.

Construction and identification of the recombinant plasmid

   We commissioned TsingKe Biotech Ltd to construct the codon-optimized
   pVAX1-GP plasmid and pUC57-UTR plasmid, and confirmed the accuracy
   through sequencing and agarose gel electrophoresis. The pGEM-UTR
   plasmid was obtained by cloning the UTR sequence from pUC57-UTR into
   the pGEM-11Zf plasmid (Promega). The pGEM-UTR-GP plasmid was obtained
   by inserting the GP sequence obtained from double enzyme digestion of
   pVAX1-GP into the pGEM-UTR plasmid. We employed Plasmid Maxi Kits
   (TIANGEN) for plasmid purification. Subsequently, the recombinant
   plasmids underwent horizontal electrophoresis on a 1% agarose gel after
   digesting with the restriction enzymes and were visualized under UV
   light.

Lipid nanoparticle formulation of the DNA

   Lipid-nanoparticle (LNP) formulations were prepared using a modified
   procedure, previously employed for siRNA^[189]68,[190]69. In summary,
   lipids, including ionizable lipid (ALC-0315) (MCE),
   1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) (MCE), cholesterol
   (MCE) and PEG-lipid (MCE), were dissolved in ethanol at a molar ratio
   of 55:10:32.5:2.5. The 10 mM lipid mixture was mixed with 10 mM citrate
   buffer containing DNA at a 1:3 ratio through a mixer. Subsequently, the
   formulations were transferred to an aqueous buffer system, filtered
   through a 0.22 μm mesh, and concentrated to the desired concentration
   via diafiltration using a 50 kD molecular weight cutoff microsep
   (Millipore) against 30 volumes of PBS. The vaccine candidates were
   diluted in PBS containing 60 mM sucrose and stored at −20 °C at a
   0.5 mg/mL concentration.

Characterization of LNP

   We performed dynamic light scattering measurements using the Malvern
   Zetasizer Nano-ZS (Malvern). Electron microscopy imaging was conducted
   with the HT7800 (HITACHI) by depositing the sample onto a porous carbon
   grid. We utilized the PicoGreen dsDNA Quantification Assay Kit
   (Solarbio) to assess the encapsulation efficiency of DNA. In essence,
   DNA-LNP were divided into two groups: one group was subjected to lysis
   using 2% Triton X-100, while the other remained untreated.
   Subsequently, the samples were processed with PicoGreen (Solarbio)
   following the manufacturer’s instructions. The amount of DNA in the
   samples was then quantified using a microplate reader. Excitation light
   was set at 480 nm, and emission light at 520 nm.

In vitro transcription and purification of RNA

   The mRNA was produced in vitro through T7 RNA polymerase-mediated
   transcription, using the plasmid pGEM-UTR-GP linearized by the
   restriction enzyme Xba I as a template. The DNA template was purified
   and quantified using spectrophotometry, followed by in vitro
   transcription using T7 High Yield RNA Transcription kit (Novoprotein),
   with N1-methylpseudouridine-5’-triphosphate (m1ΨTP) substituting for
   uridine-5’-triphosphate. The RNA was then purified using lithium
   chloride, capped enzymatically with cap1 (Novoprotein), and further
   purified after capping reaction completion. Subsequently, the
   solution’s concentration was determined, and the integrity of the RNA
   was assessed through microfluidic capillary electrophoresis. Besides,
   we employed a double-antibody sandwich NovoFast dsRNA ELISA
   (Novoprotein) to quantify dsRNA residuals in the mRNA vaccine
   candidate.

Transfection of HEK293T cells

   HEK293T cells were transfected with plasmid pVAX1-GP, mRNA, and
   transfection reagent lipofectamine 3000 (Thermo Fisher Scientific)
   separately, following the manufacturer’s instructions, or transfected
   with DNA-LNP vaccine candidates.

Western blot analysis of transfected cells

   After 24 h post-transfection, the cell lysates were collected by
   centrifugation at 1000 × g using RIPA buffer (Beyotime). The collected
   lysates were then mixed with 5×SDS loading buffer and subjected to
   SDS-PAGE electrophoresis. The antigen proteins expressed by the cells
   were detected by western blotting with a flag antibody (Proteintech).
   Blots were developed with High-sig Western ECL Substrate (Tanon) and
   imaged with a Fusion FX Imager (Vilber) using the Image Lab software
   version 6.0.

RNA isolation and quantitative real-time PCR

   The total RNA from transfected 293 T cells was isolated using TRIzol
   reagent (Sigma). cDNA was prepared using PrimeScript RT Master Mix
   (TaKaRa), and quantitative real-time PCR (qPCR) was performed using
   SYBR-green PCR Master Mix (Qiagen). The qPCR experiments were carried
   out on a Bio-Rad thermal cycler (Bio-Rad Laboratories Inc., CA, USA
   CFX96). The PCR conditions were as follows: an initial incubation at
   50 °C for 30 min, followed by a denaturation step at 95 °C for 5 min.
   Subsequently, 40 cycles of amplification were performed, each
   consisting of a 20 s denaturation step at 95 °C and a 1-min
   annealing/extension step at 55 °C. For the detection of vaccine
   antigens, we designed two pairs of antigen-specific primers based on
   the Gn and Gc gene regions. The human beta-actin gene was employed as
   the reference gene. Gene expression levels were calculated using the
   comparative Ct method. All primers were synthesized by TsingKe Biotech
   Ltd (Table [191]S2).

Immunofluorescence analysis of eukaryotic protein expression

   The transfected HEK293T cells were fixed with 4% paraformaldehyde and
   permeabilized in 0.2% Triton X-100. After blocking with 3% bovine serum
   albumin-PBS solution to prevent non-specific binding, the cells were
   incubated overnight at 4 °C with Flag antibody (1:2000, Proteintech).
   Subsequently, 594-conjugated goat anti-rabbit antibody (1:200,
   Proteintech) was used as the secondary antibody and incubated in the
   dark at 37 °C for 2 h. DNA was stained with DAPI (Solarbio). Images
   were obtained using a confocal microscope (Fluoview FV3000, Olympus,
   Japan).

Animals and immunization

   Female BALB/c mice (6 weeks old) were procured from the Laboratory
   Animal Center of the Fourth Military Medical University and were
   assigned to six groups using a random number generator. All mice were
   in good health and had never participation in any research previously.
   They were acclimated for one week prior to immunization. Before all
   invasive procedures, mice were anesthetized with 2% isoflurane
   inhalation to ensure they remained pain-free during the operations.
   Each group received two injections of candidate vaccines, administered
   into the gastrocnemius muscle under isoflurane anesthesia, on day 0 and
   day 28, respectively. The six groups of candidate vaccines were as
   follows: PBS (n = 21), empty lipid nanoparticle (LNP) without
   encapsulated nucleic acid (n = 21), inactivated vaccine (Inact)
   (n = 21), endotoxin-free DNA plasmid pVAX1-Gp (n = 21),
   LNP-encapsulated endotoxin-free DNA-LNP (n = 21), and mRNA-lipo[jet]
   vaccine (n = 21). The inactivated vaccine, referenced to the
   manufacturer’s instructions, consisted of 10 μg of inactivated bivalent
   HFRS vaccine (HANPUWEI) diluted in 0.9% NaCl, with an injection volume
   of 50 μL. Both DNA and DNA-LNP candidate vaccines contained 30 μg of
   DNA with an injection volume of 50 μL. The mRNA-lipo[jet] candidate
   vaccine comprised 10 μg mRNA, which is delivered using in vivo mRNA
   transfection reagent (Polyplus). The injection volume is 100 μL,
   administered via intramuscular injection in two separate points. Each
   group of mice received a final booster vaccination on day 182 with the
   same dose as the initial two vaccinations. Serum samples were collected
   from mouse eye blood on days 52 and 206 after the initial immunization
   to detect neutralizing antibody responses. Additionally, serum samples
   were collected from mouse tail vein blood on day 0 (before the initial
   immunization) and on days 7, 14, 21, 28, 35, 42, 49, 56, 63, 70, 84,
   98, 182, 196, and 203 after the initial immunization to assess specific
   antibody responses. Spleen tissues from the immunized mice in each
   group were collected on days 52 and 206 after the initial immunization
   for the assessment of cellular immune responses using ELISpot and flow
   cytometry. After the experiment, all mice were anesthetized with 2%
   isoflurane inhalation. After confirming adequate anesthesia, trained
   personnel humanely euthanized the animals by cervical dislocation,
   following standard procedures.

Sera antibody titer evaluation by enzyme-linked immunosorbent assay (ELISA)

   Determination of the titer of HTNV-specific IgG antibodies was
   performed by ELISA. In brief, ELISA plates (Costar) were coated with
   inactivated HTNV virus diluted in sodium carbonate buffer (100 µL/well)
   and incubated overnight at 4 °C. After coating, the plates were washed
   once with wash buffer (PBST, PBS with 0.05% Tween 20, 200 µL/well) and
   then blocked with 1% bovine serum albumin diluted in wash buffer
   (200 µL/well) for 40 min at 37 °C. Subsequently, mouse sera were
   serially diluted, starting from 1:100, and 100 µL of each dilution was
   added to the wells as the primary antibody and incubated for 1 h at
   37 °C. The plates were then washed five times and each well received
   100 µL of horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG
   as the secondary antibody (1:2000, Proteintech), followed by an
   incubation of 30 min at 37 °C. After washing six times, the plates were
   incubated with TMB substrate (Solarbio) for 15 min at 37 °C. Finally,
   50 µL of ELISA stop solution (Solarbio) was added to terminate the
   reaction, and the absorbance was measured at 450 nm using a microplate
   reader (TECAN).

Serum neutralization test

   Serum samples were serially diluted in a 2-fold gradient, starting from
   1:10 to 1:320. Subsequently, they were mixed with a viral solution at a
   dose of 100 PFUs, followed by incubation at 37 °C for 1 hour. This
   mixture was then introduced to a 96-well plate containing a monolayer
   of Vero E6 cells, which were allowed to adsorb for 2 hours in a cell
   culture incubator. After discarding the mixture, DMEM containing 1.6%
   carboxymethyl cellulose and 2% FBS was applied to each well and the
   plate was incubated at 37 °C in a 5% CO[2] atmosphere for 5 days. The
   cells were subsequently fixed with 4% paraformaldehyde and
   permeabilized with 0.5% Triton X-100. A homemade monoclonal antibody,
   mAb1A8 against HTNV-NP, was added and incubated overnight at 4 °C.
   Followed by incubation with HRP-conjugated goat anti-mouse antibody
   (1:2000, Proteintech) at room temperature for 1 hour. Color development
   was performed using a Maxi-Blue precipitating TMB substrate, and blue
   plaques were counted after drying. The neutralizing titers were
   determined and calculated according to the Karber method.

Synthesis and preparation of GP peptide

   The peptide pool used for ELISpot (30 μg/mL of each peptide) and flow
   cytometry (1 mg/mL of each peptide) consists of 10 sets of 15-mer
   peptides, all of which were obtained from Shanghai Apeptide Co (Table
   [192]S3).

Enzyme-linked immunospot assay (ELISpot)

   The ELISpot experiment was conducted following the manufacturer’s
   protocol to evaluate the cellular immune response in vaccinated mice
   using the mouse IFN-γ/IL-4 ELISpot kit (BD Pharmingen). In essence, the
   plates were initially incubated with capture antibodies at 4 °C
   overnight, followed by blocking with RPMI 1640 medium containing 10%
   FBS for 2 h. Subsequently, we plated 1,000,000 cells/well and ex vivo
   reactivated immune mouse splenocytes using either a treatment group
   consisting of a 15-peptide mixture (each peptide at a final
   concentration of 0.1 μg/mL, with a total concentration of 2 mg/mL, as
   detailed in Table [193]S3) or control groups (negative control: RPMI
   1640 medium containing 10% FBS; positive control: Concanavalin A,
   2 μg/mL (Sigma)). After incubation for 24 h (37 °C, 5% CO[2]), these
   cells were collected for subsequent transcriptome sequencing as
   described below. Then the plates were subjected to a washing step,
   followed by the addition of biotinylated anti-mouse IFN-γ or IL-4
   antibodies, and incubation at room temperature for 2 h. After rinsing
   the plates thoroughly, avidin-horseradish peroxidase and
   3-amino-9-ethylcarbazole peroxidase substrate kit were introduced.
   Following this, an automated ELISpot plate reader (CTL) was employed to
   read and count the spots on the air-dried plates. Finally, we
   calculated the number of spot-forming cells per 1 million splenocytes.

Cellular immune response by flow cytometry

   To perform intracellular cytokine staining of T cells from immunized
   mice, 1,000,000 splenocytes were ex vivo stimulated for 4 h using a
   cell stimulation cocktail (Invitrogen). After two washes with staining
   buffer, splenocytes were stained for cell surface antigens using
   directly labeled antibodies, including CD3 (FITC) (BioLegend), CD4 (PE)
   (BioLegend), and CD8 (Pacific Blue) (BioLegend). Subsequently, cells
   were adequately fixed and permeabilized using Cytofix/Cytoperm solution
   (BD) and incubated in the dark at 4 °C for 20 min. The permeabilized
   cells were then washed twice and subjected to intracellular cytokine
   staining using antibodies specific for IL-2, IL-4, IFN-γ, and Granzyme
   B (APC) (BioLegend) within Perm/Wash buffer (BD). Incubation was
   carried out in the dark at 4 °C for 30 min. Afterward, cells were
   washed twice with the staining buffer and subsequently resuspended in
   the same buffer before being subjected to flow cytometric analysis.
   Data analysis was conducted using NovoExpress software. Gating strategy
   to identify cytokine secretion in the T cells is present in Fig.
   [194]3e (take the IL-2 release from CD4^+ T cells for example). From
   all events, lymphocytes can be distinguished by their FSC/SSC
   properties. Single cells are then isolated by their relationship
   between FSC-H versus FSC-A. From single cells, CD4^+ T cells are gated,
   from which IL-2^+ CD4^+ T cells can be identified.

   For the detection of cell surface antigens, 500,000 mouse splenocytes
   were stimulated with a GP peptide pool (1 mg/mL for each peptide, see
   Table [195]S3) for 6 h at 37 °C under 5% CO[2]. Following two washes
   with staining buffer, we applied fluorescently conjugated antibodies,
   including CD3 (APC-Cyanine7) (BioLegend), CD4 (FITC) (BioLegend), CD8
   (FITC) (BioLegend), CD44 (PE) (BD Biosciences), and CD62L (APC)
   (BioLegend), to label the splenocytes, which were then incubated at
   4 °C in the dark for 30 min. After two additional washes with a
   staining buffer, the samples were analyzed using a flow cytometer. Data
   analysis was performed using NovoExpress software. Gating strategy to
   identify effector memory T (Tem) cells in the spleen is present in Fig.
   [196]3g (take the CD4^+ Tem cells in splenocytes for example). From all
   events, lymphocytes can be distinguished by their FSC/SSC properties.
   Single cells are then isolated by their relationship between FSC-H
   versus FSC-A. From single cells, CD4^+ T cells are gated, from which
   Tem (CD44^hi CD62L^lo) cells can be identified.

Transcriptome analysis

   We collected 5,000,000 mouse splenocytes from immunized mice in each
   group, which were stimulated with peptide pools for 24 h. Each sample
   was derived from three immunized mice and preserved in TRIzol reagent
   (Sigma). Subsequently, the samples were sent to Seqhealth Ltd (Wuhan,
   China) for transcriptome analysis.

HTNV challenge of mice

   The challenge model using the HTNV 76–118 strain for viral infection
   has been extensively described^[197]32,[198]52. Each group of candidate
   vaccines underwent immunization. Subsequently, on day 53, six mice in
   each group received the challenge or via intramuscular injection with
   5 × 10^5 PFUs of HTNV 76–118 strain per mouse. Five days before the
   challenge, the mice were relocated from the immunization facility to a
   biosafety level 3 (BSL-3) animal facility. On the 3rd day
   post-challenge, all animals were euthanized, and major organs (heart,
   liver, spleen, lung, and kidney) were collected for subsequent
   analysis, including viral RNA level determination, histopathological
   examination, and immunofluorescence staining. All experiments were
   conducted under BSL-3 conditions and adhered to the international
   laboratory biosafety guidelines.

Quantification of viral RNA in challenged mice tissues

   After the challenge, we conducted qPCR to detect viral RNA in the major
   organs of challenged mice. In brief, tissue samples were weighed, and
   total RNA was extracted using TRIzol reagent (Sigma). We quantified
   HTNV RNA using the S segment as the target and utilized 18S rRNA as an
   internal reference gene. All primers were synthesized by TsingKe
   Biotech Ltd (Table [199]S2).

Immunofluorescence staining of challenged mice tissues

   Tissue samples from each group were collected for paraffin embedding
   and sectioning. After slide preparation, the tissue sections were
   subjected to a series of processes including sealing, incubation with
   the primary antibody, incubation with the secondary antibody, and
   nuclear staining, as detailed in “Immunofluorescence analysis of
   eukaryotic protein expression”. The primary antibody was a homemade
   anti-HTNV-NP mouse antibody, and the secondary antibody employed was a
   FITC-labeled goat anti-mouse antibody. The results were observed and
   analyzed using a fluorescence microscope.

Hematoxylin and eosin (H&E) staining

   In terms of histopathology, we collected samples from six major organs
   and muscle tissue on the 24th day following long-term immunity. These
   samples were initially fixed using a 4% paraformaldehyde solution,
   subsequently embedded in paraffin, and then sectioned and stained with
   hematoxylin and eosin (H&E). The slides are scanned and photographed by
   a slide scanner (Winmedic). The staining results are certified by more
   than two pathologists.

Statistical analysis

   All data analyses were conducted using GraphPad Prism 9.0 software.
   Unless specified, experimental dates are presented as mean ± SEM.
   Statistical significance among different groups was assessed using
   one-way ANOVA with multiple comparison test (*P < 0.05, **P < 0.01,
   ***P < 0.001, ****P < 0.0001). And the data analysts were blinded to
   the experimental group assignments.

Animal ethics statement

   All animal experiments were conducted strictly under the regulations of
   the Chinese Regulations of Laboratory Animals and Laboratory
   Animal-Requirements of Environment and Housing Facilities. The
   procedures for the care and use of animals were approved by the Animal
   Ethics Committee of the Fourth Military Medical University (No.
   FMMU-DWZX-20221201).

Supplementary information

   [200]Supplemental Information^ (2.9MB, pdf)

Acknowledgements