Chimeric hemagglutinin-based universal influenza mRNA vaccine induces protective immunity and bone marrow plasma cells in rhesus macaques

Summary

A universal influenza vaccine that elicits a strong and lasting stalk-specific antibody response is advantageous. We utilize nucleoside-modified mRNA in lipid nanoparticles (mRNA-LNP) and unmodified self-amplifying mRNA in modified dendritic nanoparticles (sam-MDNP), expressing chimeric hemagglutinin (cHA) antigens to induce stalk-specific humoral immunity in non-human primates with pre-existing influenza virus immunity. mRNA-LNP immunization induces strong stalk-specific binding antibodies capable of protecting mice from lethal heterologous influenza virus challenges and bone marrow plasma cells (BMPCs) that persist for up to 8 months. sam-MDNP vaccine induces lower humoral immunity, despite showing strong innate activation. Transcriptomic and cytokine analyses reveal a more persistent induction of interferon responses, interleukin (IL)-1β signaling, and IL-6 production in the mRNA-LNP group, correlating with the induction of serum antibody responses and BMPCs. These results identify a transcriptional signature associated with induction of BMPCs following mRNA vaccination and highlight the utility of cHA-based mRNA-LNP vaccines in inducing persistent stalk-directed protective antibody responses.

Graphical abstract

Highlights

• •mRNA-LNP chimeric HA vaccine induces strong stalk-specific antibodies in NHPs
• •mRNA-LNP chimeric HA vaccine induces plasma cells in bone marrow
• •Vaccine-induced antibody protects mice from heterologous influenza virus challenges
• •Type I interferon and IL-6 production associate with induction of humoral immunity

Styles et al. report that a chimeric influenza virus hemagglutinin-based mRNA-LNP vaccine induces a strong influenza virus stalk-specific antibody in serum and bone marrow plasma cells and define a transcriptional profile associated with induction of these responses in non-human primates.

Introduction

Seasonal influenza infections result in considerable morbidity and mortality, underscoring the urgent need for a universal vaccine that provides broad and lasting protection, regardless of the viral strain. The efficacy of the seasonal influenza vaccine is highly dependent on its match to circulating virus strains with an effectiveness ranging from 0% to 60%.¹ The influenza virus hemagglutinin (HA) stalk has long been pursued as a possible avenue for a universal influenza vaccine due to its relatively conserved nature across multiple strains.^2,3,4,5,6 Although stalk-specific antibodies can be generated by natural infection, this response is overshadowed by the highly immunogenic properties of the globular HA head.^2,7,8,9 Various methods have been utilized to target the immune response to the stalk, including sequential immunizations with chimeric hemagglutinin (cHA) proteins. cHAs are recombinant influenza virus HA proteins with the conserved stalk domain from H1 (group 1), H3 (group 2), or influenza B seasonal viruses and the globular head from exotic avian virus HA subtypes that do not circulate in the human population.^10,11,12 The strategy redirects the immune response to the immunosubdominant stalk and induces strong and durable stalk-specific antibodies.^13,14 Stalk-specific antibodies have been shown to be cross-reactive across multiple influenza virus subtypes, neutralize a range of influenza viruses, and carry out effector-mediated functions.^{5,13,15,16,17,18,19,20} Additionally, stalk-specific antibodies prevent viral budding, inhibit endosomal viral fusion, and block HA cleavage.^21,22,23 cHA-based immunization strategies could significantly lessen reliance on the unpredictable annual vaccines and provide a much more dependable alternative during a pandemic.

One highlight of the COVID-19 pandemic was the efficacy and dependability of mRNA-based vaccines.^24,25,26 Traditionally, there are two types of mRNA-based vaccine technologies: (1) conventional (non-replicating) and (2) self-amplifying. Conventional mRNA vaccines can be identified as single-stranded linear RNA with a 5′ cap, 3′ poly (A) tail and a 5′ and 3′ untranslated region, whereas self-amplifying mRNA can be described as larger in size and encodes additional viral proteins, usually alphavirus replication machinery, required for functional replication activity.^27,28 Lipid nanoparticle-encapsulated nucleoside-modified mRNA (mRNA-LNP) constructs are advantageous due to their safety and lower inflammatory profiles, enabling enhanced protein translation and expression.^{29,30,31,32,33} Longer protein expression can contribute to better recognition and activation of the host immune response, resulting in generation of long-lived antibody responses.^34,35,36 The self-amplifying modified dendritic nanoparticle (sam-MDNP) mRNA replicon contains nonstructural proteins 1–4 from the Venezuelan equine encephalitis virus and includes dendritic materials, rather than the cationic lipids used in the mRNA-LNP vaccine. sam-based influenza vaccines have been shown to have mild side effects, persist once injected resulting in more antigen expression, require a lower dose of vaccine compared to conventional mRNA vaccines, induce influenza virus-specific neutralizing antibodies, and protect from lethal viral challenge.^27,37,38,39 sam-MDNP delivery in mice has been shown to induce robust T cell responses, while exhibiting lower acute inflammatory effects than mRNA-LNPs.⁴⁰ Thus, the primary factor driving immunogenicity of a sam-MDNP vaccine is attributed to the replicative action of the sam-RNA payload. Most studies of sam-based vaccines have been done in small animal models using traditional HA immunogens, which elicit poor HA stalk-specific antibody responses.

The formation of bone marrow plasma cells (BMPCs) is necessary for the maintenance of persistent serum antibody.⁴¹ Vaccine/infection-induced BMPCs have been shown to persist for decades.^42,43 While influenza vaccination transiently increased influenza virus-specific BMPCs in humans, 70%–99% of these BMPCs were not detectable 1 year following vaccination.⁴⁴ It is unclear why these newly formed BMPCs disappeared, but these data highlight an additional hurdle for influenza vaccine development. mRNA vaccines induce de novo B cell responses in naive individuals in the case of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).⁴⁵ Thus, we closely monitored BMPC formation post-mRNA vaccination.

The preclinical efficacy of influenza vaccines is traditionally tested in mice and ferrets. However, we test the immunogenicity of a cHA mRNA-based influenza vaccine in rhesus macaques (RMs). RMs have been utilized as an experimental model for numerous human infections and diseases.^{46,47,48,49,50} There is significant value in evaluating influenza vaccine immunogenicity in RMs. Toll-like receptors, which are responsible for recognizing and alerting the body to pathogens through pathogen-associated molecular patterns, are not identical between mice and humans.^{51,52,53,54,55} This variability can contribute to an inaccurate representation of immune responses to antigens and, especially, adjuvants when translated to humans.^56,57,58 Another limitation of modeling influenza vaccines in mice and ferrets is lifespan. Mice, on average, live a maximum of 4 years, and ferrets live on average 7–10 years, while humans can live up to 120 years.^59,60,61 In contrast, RMs can live up to 40 years, and their immune responses are more like humans, making RMs a more pragmatic animal model for studying immunogenicity of vaccines and adjuvants.^62,63,64

We evaluated the immunogenicity and efficacy of an influenza virus cHA vaccine prime-boost regimen in RMs utilizing both mRNA-LNP and sam-MDNP vaccines. Both mRNA vaccine platforms elicited robust innate activation; however, the transcriptional profiles evoked by each vaccine differed significantly. sam-MDNP immunization prompted a quick and prominent type I and type II interferon (IFN) response on day 1 that swiftly subsided by day 2, while mRNA-LNP immunization prompted an equally dominant type I and type II IFN response that persisted through day 2. mRNA-LNP, in the presence of pre-existing influenza virus immunity, resulted in the generation of long-lived stalk-specific antibody and BMPCs up to 8 months post-final immunization. Additionally, the stalk-specific antibody response was protective against lethal challenge with A/Netherlands/602/2009 (H1N1) and cH6/1N5 virus when passively transferred to mice.

Results

Study design

Ten male RMs were immunized intramuscularly (IM) with the 2020/2021 quadrivalent influenza vaccine (QIV) at weeks 0 and 6 to induce influenza virus-specific immunity (Figure 1A). We used only male RMs due to the scarcity of female RMs, and the study was not powered to detect sex differences for immune responses. Eighteen weeks post-QIV booster, RMs were divided into two groups based on H1 stalk-specific binding antibody and immunized IM with 100 μg of cH8/1 mRNA-LNP or 30 μg of cH8/1 sam-MDNP vaccine. sam-MDNP immunized RMs received 1/3 the dose of mRNA-LNP immunized RMs due to the self-amplifying property of the sam-RNA. Additionally, human vaccines utilizing self-amplifying mRNA have used lower doses ranging from 1 to 10 μg.⁶⁵ Twelve weeks later, RMs were immunized IM with 100 μg of cH5/1 mRNA-LNP or 30 μg of cH5/1 sam-MDNP. mRNA vaccines expressed membrane-anchored chimeric HA proteins. Animals were followed for 40 weeks after the final vaccination.

Figure 1.

Chimeric H8 (cH8/1) mRNA-LNP immunization results in robust activation of the innate immune response

(A) Vaccine schematic.

(B) Representative gating strategy for monocyte and DC subsets.

(C) Frequency of classical monocytes and intermediate monocytes.

(D) Expression of CD80 on classical monocytes and intermediate monocytes.

(E) Frequency of pDCs, CD11c myeloid DCs, BDCA1+, BDCA1+BDCA3+, BDCA3+, and BDCA1−BDCA3− DCs.

(F) Expression of CD80 on pDCs, CD11c myeloid DCs, BDCA1+, BDCA1+BDCA3+, BDCA3+, and BDCA1−BDCA3− DCs.

(G) Representative gating strategy for NK cells and CD69 expression.

(H) Frequency of NK cells and CD69 expression on NK cells.

(I) Representative gating strategy for neutrophils.

(J) Frequency of neutrophils. n = 5 biological replicates. Statistical significance defined as a p value of < 0.05.

mRNA-LNP vaccination induced strong activation of innate immune cells in blood

Post-cH8/1 prime, we assessed innate immune activation by evaluating changes in the frequency and activation (CD80 and CD69) of monocyte and dendritic cell (DC) subsets, neutrophils, and natural killer (NK) cells on days 0, 1, 2, 4, and 7. The gating strategy used to define innate cells is shown in Figure 1B. Monocytes were characterized as classical (CD14⁺CD16⁻), intermediate (CD14⁺CD16⁺), and non-classical (CD14⁻CD16⁺) subsets (Figure 1B). mRNA-LNP immunization showed a significant increase in the frequency and activation (CD80) of intermediate monocytes on day 2 post-vaccination (Figures 1C and 1D). In addition, the frequency of classical monocytes showed a modest increase at day 4 (Figure 1C). With respect to DC subsets, their frequency did not increase, and plasmacytoid dendritic cells (pDCs) showed a decrease transiently at day 1 (Figure 1E). However, mRNA-LNP vaccination increased the activation of several DC subsets at day 1 to day 2 including CD11c+ DC, BDCA1+ BDCA3+ DC, BDCA3+ DC, and BDCA1− BDCA3− DCs (Figure 1F). Similarly, mRNA-LNP vaccination induced activation (CD69) of NK cells at day 2 (Figure 1G) and increased the frequency of neutrophils at day 1 (Figure 1H). Taken together, these data showed that mRNA-LNP vaccination induced strong innate immune activation of monocytes and different DC subsets as well as NK cells early post-vaccination.

mRNA-LNP immunizations induced a strong H1 HA stalk-specific antibody response

mRNA-based vaccines have been shown to be potent inducers of the antibody response.^{66,67,68,69,70} To assess the vaccine-specific antibody response post-cH8/1 immunization, we first monitored plasmablast responses against cH8/1, cH5/1, and cH6/1 proteins. cH6/1, which carries the same H1 stalk domain as cH8/1 and cH5/1 with an antigenically distinct head, was included to measure stalk-specific antibody responses. Representative antigen-specific plasmablast enzyme-linked immunosorbent spot (ELISpot) images after cH5/1 boost are depicted in Figure 2A. mRNA-LNP vaccination induced a strong plasmablast response against all three proteins. The cH6/1-specific response was 10-fold higher post-cH5/1 immunization compared to post-cH8/1 immunization, suggesting a strong boosting of the stalk-specific response. The responses peaked at day 7 post-cH8/1 vaccination and day 4 post-cH5/1 vaccination (Figures 2B–2D).

Figure 2.

cH8/1 and cH5/1 mRNA-LNP immunization significantly enhance HA stalk-specific antibody responses

(A) Representative ELISpot responses for total IgG, cH8/1, cH5/1, and cH6/1 from blood (in triplicate).

(B) Longitudinal cH8/1-specific plasmablast responses.

(C) Longitudinal cH5/1-specific plasmablast responses.

(D) Longitudinal cH6/1-specific plasmablast responses.

(E) Longitudinal serum-binding antibody responses against cH8/1.

(F) Longitudinal serum-binding antibody responses against cH5/1.

(G) Longitudinal serum-binding antibody responses against cH6/1.

(H) Longitudinal serum-binding antibody responses against headless HA.

(I) Longitudinal HAI titers against influenza virus strains.

(J) Top panel: matrix of negative-stain EM reconstructions of pAbs in complex with recombinant H1 (A/California/04/09) from each RM at all time points listed. Due to limited particle representation, Fab graphics with dashed outlines are predicted placements. For samples with immune complexes in low abundance, example 2D class averages with labels are shown. Bottom panel: summary of epitopes targeted by pAbs. Each circle represents a Fab specificity from the corresponding RM and time point. (NA, not available; ND, not detectable).

(K) Blocking of MEDI8852 by mRNA-LNP hyperimmune serum from 38, 48, and 64 weeks post-vaccination. See also Figure S1. n = 5 biological replicates. Statistical significance defined as a p value of < 0.05.

We next determined the binding immunoglobulin G (IgG) antibody response against cH8/1, cH5/1, cH6/1, and headless HA proteins (Figures 2E and 3H). Like plasmablast, strong binding antibody responses were observed against all four proteins tested, demonstrating a strong induction of stalk-specific immunity. Following cH8/1 immunization, the binding antibody response against cH8/1 peaked by week 6 and only contracted 2-fold by the time of cH5/1 immunization. Post-cH5/1 boost, cH8/1 antibody was further boosted 4.5-fold compared to the peak post-cH8/1 immunization at week 30 (Figure 2E). cH5/1 and stalk-specific antibody responses were boosted after cH5/1 immunization. cH5/1 binding titers rose more than 30-fold while stalk titers rose more than 24-fold post-boost (Figures 2F and 2G). Although headless HA binding titers trended lower than cH6/1 stalk response, we still observed a 24-fold increase in stalk antibody after cH5/1 boost (Figure 2H). Despite the increase in stalk antibody, we did not observe an increase in hemagglutinin inhibition (HAI) titers following cHA immunizations relative to the QIV boost (Figure 2I). Together, these data demonstrated that mRNA-LNP immunization induced strong and persistent influenza virus HA head and stalk binding antibody.

Figure 3.

cH8/1 immunization induces expansion of cTfh and non-Tfh in the blood and results in influenza-specific BMPC and MBC

(A) Representative CXCR5+ and CXCR5- CD4 T cells at week 0 and week 1 post-cH8/1 immunization.

(B) Longitudinal CXCR5+ and CXCR5− CD4 T cell frequencies post-cH8/1 immunization.

(C) Boolean pie graphs of CCR4+/CCR6/CXCR3 on Ki-67+ CXCR5^+/− CD4s week 0 and week 1 post-cH8/1 immunization.

(D) CXCR3 expression on Ki-67+ CXCR5+ and CXCR5− CD4 T cells week 0 and week 1 post-cH8/1 immunization.

(E) Representative ELISpot responses for total IgG, cH8/1, cH5/1, and cH6/1 from bone marrow of mRNA-LNP-immunized RMs (in triplicate).

(F) Longitudinal cH8/1 BMPC frequencies.

(G) Longitudinal cH5/1 BMPC frequencies.

(H) Longitudinal cH6/1 BMPC frequencies.

(I) Representative ELISpot responses for total IgG, cH8/1, cH5/1, and cH6/1 from PBMCs of mRNA-LNP-immunized RMs (in triplicate).

(J) Longitudinal cH8/1-specific MBC frequencies.

(K) Longitudinal cH5/1-specific MBC frequencies.

(L) Longitudinal cH6/1-specific MBC frequencies. See also Figure S2 and Table S1. n = 5 biological replicates. Statistical significance defined as a p value < 0.05.

We next performed electron microscopy-based polyclonal epitope mapping⁷¹ (EMPEM) using serum from pre- and post-QIV, cH8/1, and cH5/1 immunizations to map the epitopes on the HA stalk of the A/California/04/2009 (H1N1) strain. There were no detectable H1 stalk-specific antibodies prior to immunization or post-QIV booster immunization (data not shown). However, at 6 weeks (week 30) following cH8/1 prime, we detected antibodies binding to the lower stalk region and protomer in the majority of mRNA-LNP immunized RMs (Figures 2J and S1). Following the cH5/1 boost, all 5 RMs retained low stalk-binding antibodies. We also observed additional specificities, including central stalk in RKr19 and RKd19 and an anchor-like antibody in RKd19 (Figures 2J and S1). Promotor-Fab complexes were observed post-cH8/1 immunization in 4 of the 5 RMs, indicating the presence of HA-specific antibodies capable of destabilizing the HA interface and breaking apart the HA trimer (Figure 2J). To further understand the presence of MEDI8852-like responses observed by EMPEM, we performed blocking experiments with RM serum at weeks 38 (2 weeks post-cH5/1), 48 (12 weeks post-cH5/1), and 64 (7 months post-cH5/1). We found that peak serum (week 38) was capable of blocking ∼30% of MEDI8852 monoclonal antibody (mAb) binding. This inhibition was reduced to ∼10% to 7% by weeks 48 and 64, respectively (Figure 2K). These data demonstrate that cHA mRNA-LNP immunization induced antibodies binding the epitope similar to the human mAb MEDI8852, which is known to directly block the fusion of HA to host cell membranes, inhibit HA maturation, and block cell-to-cell spread of infectious virus by preventing host cell protease cleavage. Moreover, MEDI8852 mAb aids in the clearance of infectious influenza virus through host-directed Fc-mediated effector functions including antibody-dependent cellular cytotoxicity, antibody-dependent cellular phagocytosis, and complement-dependent cytotoxicity.^72,73 Cumulatively, these data indicated that mRNA-LNP vaccination was proficient at inducing antibodies to the central and MEDI8852-like low stalk epitopes of influenza virus H1.

mRNA-LNP vaccination induced ICOS+ CXCR3+ cTfh and non-Tfh cells

CD4 T cell help is critical for the generation of long-lived antibody responses. To characterize the effects of cH8/1 mRNA immunization on CD4 T cells, we monitored the frequency of proliferating cells (Ki-67+) that co-expressed inducible T cell costimulator (ICOS) (critical for B help) (Figure S2A). mRNA-LNP immunization induced a significant increase in the frequency of Ki-67+ ICOS+ CD4 T cells at week 1 (Figure S2B). Both CXCR5+ (circulating Tfh) and CXCR5− (non-Tfh) Ki-67+ ICOS+ CD4 T cells showed an increase with similar kinetics of total Ki-67+ ICOS+ cells, and the increase was greater for non-Tfh compared to Tfh cells (Figures 3A and 3B). We next investigated the co-expression of chemokine receptors CXCR3, CCR4, and CCR6 on cTfh (Tfh1, Tfh2, and Tfh17, respectively) and non-Tfh (Th1, Th2, and Th17, respectively) cells at weeks 0 and 1 post-vaccination to understand the influence of vaccination on T helper polarization (Figure 3C). Previous studies, including our own, showed a positive association between CXCR3+ Tfh and stronger antibody response.^74,75,76 The majority of Tfh and non-Tfh cells either expressed CXCR3 or did not express any chemokine receptors (Tfh0/Th0) on the day of vaccination. Following the boost, we observed a significant increase in the proportion of CXCR3+ cells among non-Tfh cells (Figure 3D). Additionally, we observed weak IFNγ+ CD4 and CD8 T cell responses in 2 of 5 immunized RMs (Figures S2C and S2D and Table S1). Collectively, these results demonstrated rapid proliferation of both cTfh1 and Th1 cells, which might have contributed to the strong antibody response in the mRNA-LNP-vaccinated RMs.

mRNA-LNP induced H1 stalk-specific plasma cells in bone marrow that persist up to 8 months

Persistent long-lived antibody responses are attributed to the establishment of BMPCs. We explored BMPC formation longitudinally. Representative ELISpot images are shown in Figure 3E. Two QIV immunizations failed to generate detectable levels of influenza virus-specific BMPCs (Figures 3F–3H). However, these responses were evident following mRNA-LNP vaccination. The cH8/1 mRNA-LNP immunization induced low frequencies of cH8/1-specific BMPCs at 4 weeks (geometric mean of 0.09%), and these responses increased by nearly 4-fold by 12 weeks (geometric mean of 0.34%) (Figure 3F). The cH5/1 mRNA-LNP immunization boosted these responses further by 2-fold at 4 weeks post-immunization, and the responses contracted by about 3-fold over 8 months (Figure 3F). The cH5/1- and cH6/1-specific BMPCs were mostly below level of detection post-cH8/1 immunization (Figures 3F–3H). However, these responses were clearly present at 4 weeks following the cH5/1 mRNA-LNP boost at frequencies (geometric mean of 0.6%–0.8%) that were comparable to cH8/1-specific BMPCs (Figures 3G and 3H). Lastly, we measured cH8/1, cH5/1, and cH6/1 memory B cell (MBC) frequencies after each immunization. Representative ELISpot images are shown in Figure 3I. Influenza virus-specific MBCs were mostly undetectable post-QIV (Figures 3J–3L). However, cH8/1 MBC frequencies peaked post-cH8/1 immunization (Figure 3J). The cH5/1 boost increased both cH5/1 and stalk-specific MBC frequencies by ∼21- and 12-fold, respectively (Figures 3K and 3L). Interestingly, these frequencies only declined by 6- and 4-fold, respectively, by 8 months post-peak (week 44) (Figures 3K and 3L). Altogether, these data demonstrate the robust plasmablast and antibody responses generated by mRNA-LNP vaccination resulted in seeding of influenza virus-specific plasma cells in the bone marrow and MBC in the blood.

Passive transfer of serum from mRNA-LNP-vaccinated RMs protected against lethal influenza virus challenge in mice

To determine the protective effect of the antibody response generated through mRNA-LNP immunization, we performed passive transfer of sera from vaccinated RMs into female BALB/c mice followed by lethal virus challenge. Serum from influenza virus naive RMs, week 8 post-QIV, week 2 post-cH5/1 (peak serum), week 48, and 7 months (memory serum) post-cH5/1 immunization, were separately pooled and used for passive transfer. Twenty-four hours post transfer, mice were challenged with 10 times the lethal dose 50 (LD₅₀, titered using pooled week 8 post-QIV serum) of A/Netherlands/602/2009 H1N1 or cH6/1N5 virus (Figure 4A) with the latter virus representing a virus, which could only be recognized by stalk-specific antibodies. All mice passively immunized with naive RM serum lost significant body weight and succumbed to infection by day 6 post-A/Netherlands/602/2009 H1N1 infection and day 7 post-cH6/1N5 infection (Figures 4B–4E). Mice passively immunized with week 38 serum (week 2 post-cH5/1 immunization) experienced minimal weight loss and completely survived lethal challenge by both A/Netherlands/602/2009 H1N1 and cH6/1N5 viruses (Figures 4B–4E). Passive transfer of week 48 and week 64 memory serum pools resulted in significant weight loss in all mice, and 4 of the 5 challenged mice per group did not survive (Figures 4C and 4E and S3). This could be attributed to a 6- to 11-fold drop in circulating stalk-specific antibodies at weeks 48 and 64, respectively, relative to the peak titer (Figure 2G). The circulating stalk-specific antibodies produced by resident stalk-specific BMPCs were not sufficient in quantity to protect against a 10-fold lethal challenge by either virus strain. Collectively, these data demonstrate that the antibody response induced by mRNA-LNP vaccination exhibits specificity that can protect against a heterologous lethal influenza virus challenge.

Figure 4.

Chimeric mRNA-LNP serum completely protects mice against lethal influenza challenge

(A) Passive antibody transfer schematic.

(B) Weight loss of mice passively immunized with serum from naive RMs, week 14 serum (8 weeks post-QIV boost), week 38 serum (2 weeks post-cH5/1 immunization), week 48 (12 weeks post cH5), and week 64 serum (7 months post-cH5/1 boost) and challenged with A/Netherlands/602/2009.

(C) Survival curve of mice passively immunized with serum from naive RMs, week 14 serum, week 38 serum, week 48, and week 64 serum and challenged with A/Netherlands/602/2009.

(D) Weight loss of mice passively immunized with serum from naive RMs, week 14 serum, week 38 serum, week 48, and week 64 serum and challenged with cH6/1N5.

(E) Survival curve of mice passively immunized with serum from naive RMs, week 14 serum, week 38 serum, week 48, and week 64 serum and challenged with cH6/1N5. See also Figure S3. n = 5 mice/group. Statistical significance defined as a p value < 0.05.

sam-MDNP immunization induced strong innate activation but resulted in a weaker humoral response

We next tested the immunogenicity of sam-MDNP vaccine. Sam-MDNP immunized RMs followed the same immunization schedule as mRNA-LNP immunized RMs (Figure 5A), and similar innate and adaptive immune measurements were performed. Like mRNA-LNP vaccination, sam-MDNP vaccination was proficient at inducing strong innate immune activation. Within the monocyte populations, we observed intermediate monocytes increased in frequency (Figure S4A) and activation (CD80) (Figure 5B) at day 1 post-cH8/1 vaccination. There was no change in classical or non-classical monocyte frequencies or activation (Figures S4A and S4B). Within the DC populations, we found no subsets increased in frequency post-vaccination (Figure S3A). However, sam-MDNP vaccination induced activation of CD11c+, BDCA1+BDCA3+, and BDCA3+ DC subsets on day 1 or day 2 (Figure 5B). There was no significant change in frequency or activation of NK cells nor neutrophil frequencies (Figures S4A and S4B). Overall, the innate activation profile following sam-MDNP was comparable to mRNA-LNP activation but with a faster kinetics for intermediate monocytes and some of the DC subsets.

Figure 5.

sam-MDNP vaccine responses

(A) sam-MDNP vaccine schematic.

(B) Longitudinal CD80 expression on intermediate monocytes, CD11c+ myeloid DCs, BDCA1+BDCA3+ mDCs, and BDCA3+ mDCs.

(C) Longitudinal serum-binding antibody responses against cH8/1.

(D) Longitudinal serum-binding antibody responses against cH5/1.

(E) Longitudinal serum-binding antibody responses against cH6/1.

(F) Longitudinal serum-binding antibody responses against headless HA.

(G) Longitudinal expression of ICOS on total Ki-67+ CD4 T cells, CXCR5+ Ki-67+ CD4-T cells, and CXCR5- Ki-67+ CD4 T cells.

(H) CXCR3 expression on Ki-67+ CXCR5+ and CXCR5− CD4 T cells week 0 and week 1 post-cH8/1 immunization.

(I) IFNγ+ CD4 T cell responses before and after cH8/1 and cH5/1 immunization.

(J) IFNγ+ CD8 T cell responses before and after cH8/1 and cH5/1 immunization.

(K) Longitudinal cH6/1 MBC frequencies.

(L) Longitudinal cH6/1 BMPC frequencies. See also Figures S4 and S5 and Table S1. n = 5 biological replicates. Statistical significance defined as a p value < 0.05.

Next, we evaluated the humoral immune response. cH8/1 immunization resulted in a 3.3-fold increase in binding titers to cH8/1, and cH5/1 did not further boost these responses (Figure 5C). The cH8/1 immunization also resulted in about 3-fold increase in cH5/1 (Figure 5D)-, cH6/1 (Figure 5E)-, and headless HA (Figure 5F)-specific binding titers, and these were marginally boosted (less than 2-fold) following cH5/1 immunization. EMPEM analysis revealed the presence of antibodies binding the HA anchor, similar in angle of approach to the anchor-binding antibodies in humans^77,78 by 10 weeks post-cH8/1 immunization in 3 of the 5 sam-MDNP immunized RMs (Figures S5A and S5B). These data indicated that sam-MDNP immunization is a potent activator of the innate immune response; however, the binding antibody responses appeared to be blunted.

The T cell response was next evaluated. The sam-MDNP vaccination induced a significant increase in total proliferating ICOS+ CD4 T cells at 2 weeks post-cH8/1. This increase was observed in both cTfh and non-Tfh CD4 T cell subsets (Figure 5G). However, the CXCR3 expression did not increase in either the cTfh or non-Tfh populations (Figure 5H). In determining the cH8/1 or cH5/1-specific IFNγ+ T cell response, sam-MDNP vaccination induced low frequency of CD4 T cells in one monkey but a robust CD8 T cell response in the same monkey. These responses were not recalled to the same magnitude following cH5/1 immunization, indicating that these responses were likely directed to the head and not the stalk of cH8/1 (Figures 5I and 5J; Table S1).

The sam-MDNP vaccination induced cH8/1, cH5/1, and stalk-specific MBCs in 2–3 animals (Figures 5K and S5C and S5D); however, we were unable to detect any cH8/1, cH5/1, or stalk-specific BMPCs (Figures S5E and S5F and 5L). Additionally, we found that sam-MDNP immune sera were unable to protect mice from lethal challenge by A/Netherlands/602/2009 H1N1 or cH6/1N5 viruses (Figures S5G–S5J). It is important to note that influenza virus-specific MBCs were established in sam-MDNP immunized RMs, which could provide some protection in the host species.

mRNA-LNP vaccination results in longer induction of IFN, IL-1β, and IL-6 signaling compared to sam-MDNP vaccination

We next evaluated the similarities and differences in transcriptional profiles induced by each vaccine regimen to understand changes in global gene expression profiles early post-vaccination. Blood collected on days 0, 1, 2, 4, and 7 post-cH8/1 immunization was utilized to perform RNA sequencing analysis (Figure 6A). Principal-component analysis revealed distinct clustering of genes on days 1 and 2 compared to day 0 in the mRNA-LNP group (Figure 6B). However, in the sam-MDNP group, only day 1 formed a distinct cluster demonstrating better persistence of changes in transcriptome in the mRNA-LNP group (Figure 6B). Accordingly, mRNA-LNP immunization led to an upregulation of 1,078 genes and downregulation of 332 genes on day 1 compared to only 588 upregulated and 166 downregulated genes in the sam-MDNP group. By day 2, 316 genes remained upregulated and 47 remained downregulated in the mRNA-LNP group compared to 43 upregulated and 19 downregulated genes in the sam-MDNP group (Figure 6C).

Figure 6.

Immune activation post-sam-MDNP immunization peaked at day 1 post-immunization and subsided rapidly, while activation in the mRNA-LNP immunized group was more persistent

(A) Vaccine schematic.

(B) Principal-component analysis (PCA) for mRNA-LNP and sam-MDNP groups day 0–7 post-cH8/1 immunization.

(C) Number of upregulated and downregulated genes days 0, 1, and 2 post-cH8/1 immunization.

(D) Gene ontology enrichment analysis (GSEA) of differentially altered pathways at day 1 (D1) and day 2 (D2) post-cH8/1 immunization.

(E) Heatmaps of the IFNα response, IL-6/JAK/STAT3 signaling, and E2F targets genes at day 1 and day 2 post-cH8/1 immunization.

(F) Longitudinal gene expression of IRF7, IRF9, ISG15, MX1, and IFIT2.

(G) Longitudinal gene expression of IL-1β, IL-1R1, IFNGR1, CXCL10, and IL-15RA.

(H) Plasma concentration of IL-6, IL-1RA, and CXCL10. See also Figure S6. n = 5 biological replicates. Statistical significance defined as a p value < 0.05.

To further elucidate cellular functions associated with vaccine-induced gene modulation, we performed pathway analysis on day 1 and day 2 differentially expressed genes (Figure 6D). Both vaccines induced a strong IFNα, IFNγ, and overall inflammatory response on day 1 as indicated by increased expression of RSAD2, IFIT2, IF144L, ISG15, CMPK2, OASL, MX1, EPSTI1, IRF7, and IRF9. However, by day 2, expression of these genes decreased in the sam-MDNP group while they remained elevated in the mRNA-LNP group (Figures 6D–6F). Genes associated with the interleukin (IL)-6/JAK/STAT3 signaling, IL-2/STAT5 signaling, complement activation, tumor necrosis factor alpha (TNF-α) signaling via nuclear factor κB (NF-κB), and hypoxia were upregulated through day 2 in the mRNA-LNP group but declined significantly after day 1 in the sam-MDNP group (Figures 6D and 6E). Interestingly, genes related to protein secretion, angiogenesis, and coagulation were more upregulated in the mRNA-LNP group, while genes related to apoptosis, the p53 pathway, and MTORC1 signaling were more elevated in the sam-MDNP group on day 1. Genes associated with E2F targets and MYC_targets V1 and V2 were downregulated on days 1 and 2 in the mRNA-LNP group, but not in the sam-MDNP group (Figures 6D and 6E). Collectively, these data showed that both vaccines induced strong type I and type II IFN signaling and IL-6/JAK/STAT3 signaling; however, this upregulation persisted longer in the mRNA-LNP group. Additionally, genes related to protein secretion were upregulated in the mRNA-LNP group, while the genes related to cell-cycle arrest and cell death were upregulated in sam-MDNP group.

IL-1β, through the downstream regulation of IL-6, plays an important role in the antibody response.⁷⁹ IL-1β and its receptor IL-1R1 expression were upregulated on day 1 post-mRNA-LNP immunization relative to the sam-MDNP group (Figure 6G). Expression of the IFNα receptor IFNAR1 was also higher in the mRNA-LNP group on day 1 post-vaccination (Figure 6G). Since IL-1β mRNA expression was reduced in the sam-MDNP group, which indirectly impacts the antibody response, we examined IL-6 protein production in the plasma using the Meso Scale Discovery platform. We found that IL-6 production increased in the mRNA-LNP group at day 1, but not in the sam-MDNP group (Figure 6H). A previous study showed that IL-1 (IL-1α and IL-1β) and IL-1ra (IL-1 receptor antagonist) play an important role in regulating the inflammatory response to RNA vaccines.⁸⁰ IL-1ra negatively regulates the proinflammatory response by competing for receptor occupancy with IL-1β.^81,82 Therefore, we measured the concentration of IL-1ra in the plasma on days 1 and 2. Interestingly, sam-MDNP immunization resulted in significantly higher IL-1ra concentrations in the plasma at day 1 post-vaccination compared to the mRNA-LNP group (Figure 6H). We noted that transcripts for CXCL10 (IP-10, the ligand for CXCR3) and IL-15RA (the receptor for IL-15) were induced in both groups but were elevated in the sam-MDNP group on day 1 compared to day 2 in the mRNA-LNP group (Figure 6G). Accordingly, we noticed an increase in the concentration of IP-10 in the plasma, which increased in the sam-MDNP group at day 1 but not in the mRNA-LNP group (Figure 6H). Both CXCR3 and IL-15 have been shown to be associated with favorable B and T cell responses in vivo.^74,83 Despite this, we did not see induction of a strong antibody response in the sam-MDNP group or increased CXCR3 expression on proliferation ICOS+ cTfh or non-Tfh (Figures 5G and 5H). Cumulatively, these data suggested that increased IL-1β expression could have positively modulated IL-6 cytokine production, leading to a better humoral immune response in the mRNA-LNP group.

To further understand the contribution of the type of nanoparticle and/or RNA used for the induction of cytokines, we stimulated rhesus peripheral blood mononuclear cells (PBMC) in vitro using MDNP and LNP containing either sam-RNA or mRNA and measured IL-6, IL-1ra, and IP-10 in the supernatants (Figure S6). With MDNPs, they induced faster cytokine production compared to LNPs (as in vivo), and the type of RNA had minimal effect on cytokine production except more IL-6 was produced with sam-RNA. However, with LNPs, sam-RNA induced higher concentrations of cytokines compared to mRNA. When we compared sam-MDNP and mRNA-LNP in vitro, like in vivo, the sam-MDNP induced higher levels of IL-1ra and IP-10 compared to mRNA-LNP. These results suggested that both the lipid composition and the type of mRNA contributed to the induction of IL-1ra and IP-10 in vivo.

Induction of type I and II IFN pathways is important for persistent BMPC responses

Lastly, we sought to identify specific transcriptional signatures associated with induction of strong humoral immune responses ultimately resulting in generation of BMPCs. We took two approaches. First, we determined the associations between some of the genes in the IFN signaling and IL-6/Jak/STAT3 signaling pathways. We performed correlations between normalized gene expression on days 1 and 2 with the cH6/1-specific plasmablast response, peak binding antibody responses, or BMPCs. On day 1, we observed a direct association between expression of IFNRA1, IL-1β, and IL-1R1 and inverse association between expression of CXCL10 and IL-15RA with all three humoral measurements (Figure 7A). A strong IFN response on day 2, as evidenced by MX1, IFIT2, ISG15, IRF7, IRF9, STAT3, STAT2, and STAT1, also strongly correlated with a better humoral immune response (Figure 7A). Interestingly, a strong type I IFN response on day 2 highly correlated with the BMPC response 2 months post-final cH5/1 immunization. In addition to transcriptomic expression, we observed IL-6 cytokine production directly correlated with cH6/1-specific plasmablasts, binding antibody, and BMPCs, while IL-1ra protein inversely correlated with these responses (Figure 7B).

Figure 7.

Induction of type I and II interferon pathways is important for persistent BMPC responses

(A) Association between humoral immune responses and specific transcriptional profiles day 1 and 2 post-cH8/1 immunization.

(B) Correlation between IL-6 and IL-1ra protein expression and cH6/1-specific peak plasmablasts, binding antibody, and BMPC responses.

(C) Graphical representation of genes that correlated with vaccine responses and over-representation analysis (ORA) of the pathways most represented by those correlated genes.

(D) List of pathways that significantly correlated with cH6/1-specific plasmablast, binding antibody, and BMPC responses without prior down-selection. n = 5 biological replicates. Statistical significance defined as a p value of < 0.05.

In the second approach, we determined correlations between the expression of every transcript on day 1 or day 2 with humoral immunity to identify transcriptional signatures associated with induction of a strong antibody response and BMPCs. This analysis allowed us to define transcriptional signature without prior down-selection of genes based on differential gene expression. This analysis identified 4,777; 3,012; and 3,177 genes that correlated with cH6/1-specific plasmablasts, binding antibody, and BMPCs, respectively. We took all genes that correlated with the cH6/1 humoral immune response from both vaccine groups and performed over-representation enrichment analysis to identify common immunological pathways (Figure 7C). Again, we identified the type I IFN response to be highly associated with the cH6/1-specific BMPC response (Figure 7D). Moreover, we identified the type II IFN response, TNF-α signaling via NF-κB, and heme metabolism as pathways associated with BMPC formation (Figure 7D). It is notable that both the type I and II IFN response pathways highly correlated with the overall humoral immune response. Collectively, these data signify that a persistent type I and II IFN response coupled with TNF-α signaling via NF-κB in innate cells early post-vaccination may be important to generate BMPCs.

Discussion

Induction of potent and persistent stalk-specific antibodies in the presence of pre-existing immunity to influenza virus is a crucial step toward a universal influenza virus vaccine. In this study, we targeted the H1 stalk through sequential cH8/1 and cH5/1 vaccination expressed by mRNA-LNP or sam-MDNP vaccines. Our data revealed that mRNA-LNP vaccination can induce a potent stalk-specific antibody response with protective potential against heterologous lethal influenza virus challenge. Importantly, the mRNA-LNP vaccine-induced antibody persisted up to 10 months and was associated with induction of strong BMPCs. Sequential immunization with mRNA-LNP vaccines shows promise as an effective strategy for eliciting long-lasting protective antibody responses against heterologous influenza virus infections in humans.

Interestingly, we observed similar innate immune activation between sam-MDNP and mRNA-LNP vaccines; despite differences in lipid carrier composition and the self-amplification nature of RNA, both vaccines were potent activators of intermediate monocytes. A previous mRNA-LNP-based vaccination in non-human primates (NHPs) showed rapid migration of classical, intermediate, and non-classical monocyte subsets to the site of immunization within 24 h,⁸⁴ in line with some of our findings in the blood. Another observation was that innate activation typically occurred on day 1 with sam-MDNP immunization and day 2 with mRNA-LNP immunization.

sam-MDNP vaccine induced a lower antibody response compared to mRNA-LNP vaccine. Previous studies using self-amplifying vaccine approaches for COVID-19 and other infectious diseases showed induction of strong antibody responses in mice and NHPs.^38,40,85 However, in NHPs, this result was generated using LNP delivery. While the self-amplifying platform used here also induced a potent CD8 T cell response in mice,⁴⁰ it is notable that, in these studies, the MDNP delivery formulation was observed to be less inflammatory when compared head to head against conventional LNPs. It may be that MDNP delivery material induces a less pronounced or shorter-lived initial local cytokine response in NHPs than LNPs and that the downstream adaptive immune response is dependent upon this initial material-based response. This would be in line with reports that LNPs possess strong inherent adjuvanting activity in the absence of an mRNA payload.^86,87 We observed low levels of HA-specific CD4s in 2 of 5 RMs and high to moderate CD8s in those same RMs. Importantly, we used a 3-fold lower dose of sam-MDNP, expecting a dose advantage with the sam-RNA, though this dose is considerably larger than what has been administered in human trials using LNPs,⁶⁵ making it difficult to make firm conclusions about inherent immunogenicity differences between the two mRNA platforms in the current study. Regardless, this study provides an opportunity to understand the molecular and immunological determinants of inducing long-lived humoral immunity in NHPs and highlights the complexity of the interplay between immunogenicity of the nucleic payload and the carrier formulation.

Importantly, we determined a transcriptional signature associated with induction of BMPCs following mRNA vaccination in NHPs. Our results highlighted the importance of type I IFN signaling, IL-1β signaling, and IL-6 production as key factors in inducing a potent antibody response and BMPCs. mRNA-LNP vaccines have been shown to induce BMPCs in humans and NHPs; however, the associated transcriptional profile was not discussed.^45,88,89 mRNA-LNP immunization resulted in persistent upregulation of certain IFN-stimulated genes (ISGs) through day 2, which play an important role in the type I and II IFN response and influence the humoral immune response.^90,91,92,93 Genes in the IL-6/JAK/STAT3 signaling pathway related to innate cell chemotaxis and adhesion, inflammatory response, antiviral response, and anti-inflammatory response were upregulated in the mRNA-LNP group.^{94,95,96,97,98} Taken together, these data demonstrated that, while both vaccines induced a robust inflammatory response, only mRNA-LNP vaccination resulted in prolonged expression of ISGs, which has been previously shown with LNP-formulated mRNA vaccines.^99,100

IL-1β and its receptor IL-1R1 were upregulated in the mRNA-LNP group at day 1 post-cH8/1 immunization. IL-1β induces and regulates the expression of IL-6 and activates the transcription factor NF-κB. Upon T cell receptor engagement, NF-κB induces differentiation, maturation, and proliferation of activated T cell, which can provide B cell help. In agreement with increased IL-1β expression, we observed a significant increase in IL-6 expression on day 1. The role of IL-6 in CD4 T cell differentiation and B cell responses has been well documented. IL-6 enhances Tfh differentiation, activation, migration, plasmablast proliferation, and survival and promotes growth and maturation of plasma cells.^{101,102,103,104,105,106} We found that IL-6 protein secretion directly correlated with the HA stalk-specific plasmablast, binding antibody, and BMPC responses when including both vaccine groups. Type I IFN signaling, IL-1β signaling, and IL-6 production were quickly downregulated in the sam-MDNP group. Multiple reasons could have contributed to this. Expression of alphaviral nonstructural genes suppresses IFN signaling and, eventually, cellular translation through nonstructural gene action and cellular sensing of viral replication,^107,108 further reviewed in Liu et al.¹⁰⁹ IL-1ra was significantly lower on day 1 in the mRNA-LNP group relative to the sam-MDNP group. Higher IL-1ra induction could have resulted in decreased IL-1β signaling, leading to decreased IL-6 expression. Additionally, we found that IL-1ra protein expression was indirectly associated with the magnitude and frequencies of plasmablasts, binding antibody, and BMPCs in RMs. The IL-1 response has been previously reported to be triggered by the ionizable lipid components used in LNPs and is likely a major contributor to the associated adaptive immune response.⁸⁰ Higher levels of IL-1Ra were also observed in vitro utilizing mRNA-LNP, mRNA-MDNP, sam-LNP, and sam-MDNP constructs expressing luciferase. sam-LNP, followed by mRNA-MDNP and sam-MDNP, expressed more IL-1Ra relative to mRNA-LNP in RM PBMCs. Differentiating leukocyte cytokine signaling in vivo at the site of injection or draining lymph nodes in the context of vaccination with different RNA cargo types with equivalent delivery formulation chemistry would be valuable to further understanding the mechanisms determining the potency of nucleic acid vaccines.

In conclusion, we show that a cHA-based nucleoside-modified mRNA-LNP prime-boost regimen induces influenza virus stalk-directed antibodies in the presence of pre-existing influenza virus immunity. These responses were protective against lethal influenza virus challenge. mRNA-LNP vaccination led to the establishment of stalk-specific MBCs and BMPCs. Vaccines capable of inducing strong stalk-directed antibodies would be invaluable as universal influenza virus vaccines to protect against drifted seasonal, zoonotic, and emerging pandemic influenza virus strains. Additionally, we believe we have identified a transcriptional profile for generating influenza virus-specific BMPCs through vaccination. These data imply that a strong type I and II IFN response, strong IL-1β expression coupled with lower IL-1ra, and strong IL-6 production are important to achieve BMPC formation in vivo.

Limitations of the study

A limitation of this study was that we did not determine the durability of protection mediated by the immunity induced by the cHA mRNA vaccination approach. Influenza virus challenges were not conducted in the vaccinated RMs, as the influenza virus typically induces a mild or subclinical illness rather than severe disease in RMs.^110,111 We note that the failure to see protection with memory sera in passive antibody transfer studies does not indicate the lack of long-term protection, since many other factors, including the frequency of MBCs that will rapidly generate more antibody-secreting cells and BMPCs following infection in the host, can contribute to protection. The contribution of the MBCs will not be captured in an adaptive transfer setting. In addition, adaptive transfer experiments used 10 times the lethal challenge dose, which was determined using serum collected after two QIV vaccinations. The dose chosen was very stringent and would also have contributed to lack of protection with memory sera.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Rama R. Amara (ramara@emory.edu).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •All data underlying the figures can be found via ImmPort identifier SDY2537. All EM maps were deposited to the Electron Microscopy Data Bank (ebi.ac.uk/emdb) under accession IDs EMD-43686, EMD-43687, EMD-43689, EMD-43690, EMD-43691, EMD-43692, EMD-43693, EMD-43694, EMD-43695, EMD-43696, EMD-43697, EMD-43698, EMD-43699, and EMD-43704. All RNA sequence data were deposited into Gene Expression Omnibus (ncbi.nlm.nih.gov/geo) under accession IDs GEO: GSE304256, GSM9145993, GSM9145994, GSM9145995, GSM9145996, GSM9145997, GSM9145998, GSM9145999, GSM9146000, GSM9146001, GSM9146002, GSM9146003, GSM9146004, GSM9146005, GSM9146006, GSM9146007, GSM9146008, GSM9146009, GSM9146010, GSM9146011, GSM9146012, GSM9146013, GSM9146014, GSM9146015, GSM9146016, GSM9146017, GSM9146018, GSM9146019, GSM9146020, GSM9146021, GSM9146022, GSM9146023, GSM9146024, GSM9146025, GSM9146026, GSM9146027, GSM9146028, GSM9146029, GSM9146030, GSM9146031, GSM9146032, GSM9146033, GSM9146034, GSM9146035, GSM9146036, GSM9146037, GSM9146038, GSM9146039, GSM9146040, GSM9146041, and GSM9146042. All accession numbers are listed in the key resources table.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This study was funded in part by the Collaborative Influenza Vaccine Innovation Centers (CIVICs contract # 75N93019C00051 to F.K., R.A., R.R.A., J.S.C., A.B.W., and Y.K.), UM1AI169662 (R.R.A) Office of Research Infrastructure Programs (ORIP)/NIH base grant P51 OD011132 (Emory National Primate Research Center [ENPRC]), and National Institute of Allergy and Infectious Diseases (NIAID)/NIH R01AI146101 and R01AI153064 (N.P.). The authors thank Kheyanna Davis, Zeba Momin, and Julia Boll for help with sample processing; the ENPRC research resources and the vaterinarian team specifically Stephanie Ehnert, Stacey Weissman, Jennifer Wood, Joe Jenkins, and Rachelle Stammen for help with primate care, proceedures, and sample collection; ENPRC/CFAR immunology core, Kiran Gill, and Ankur Saini for training and maintenance of flowcytometry instruments.

Author contributions

Conceptualization, T.M.S., A. Akhtar., F.K., R.A., and R.R.A.; data acquisition, T.M.S., A. Akhtar., C.G., G.N., H.M., J.S.M., P.T., K.S., H.L.T., J.A.F., J.H., and J.M.C.; data analysis, T.M.S., A. Akhtar., C.G., G.N., D.G., H.L.T., J.A.F., M.L., B.F., A. Abbad., Z.S.Q., J.H., J.S.C., and J.M.C.; methodology, T.M.S., A. Akthar., C.G., G.N., J.S.M., P.T., D.G., K.S., H.L.T., J.A.F., G.C., Y.K.T., Z.S.Q., J.H., and J.M.C.; supervision, T.M.S., A. Akhtar., G.N., J.H., J.M.C., A.B.W., J.S.C., N.P., Y.K., F.K., R.A., and R.R.A.; manuscript writing, T.M.S., A. Akhtar., and R.R.A.; manuscript editing, T.M.S., A. Akhtar., C.G., G.N., D.G., G.C., Z.S.Q., J.H., J.M.C., A.B.W., J.S.C., C.W.M., N.P., Y.K., F.K., and R.R.A.

Declaration of interests

The Icahn School of Medicine at Mount Sinai has filed patent applications relating to SARS-CoV-2 serological assays, Newcastle Disease Virus-based SARS-CoV-2 vaccines, influenza virus vaccines, and influenza virus therapeutics, which list Florian Krammer as a co-inventor. Mount Sinai has spun out a company, Kantaro, to market serological tests for SARS-CoV-2, and another company, CastleVax, to develop SARS-CoV-2 vaccines. F.K. is a co-founder and scientific advisory board member of CastleVax. F.K. has consulted for Merck, CureVac, Seqirus, and Pfizer and is currently consulting for Third Rock Ventures, GSK, Gritstone, and Avimex. The Krammer laboratory is also collaborating with Dynavax on influenza vaccine development. N.P. is named on patents describing the use of nucleoside-modified mRNA in lipid nanoparticles as a vaccine platform. He has disclosed those interests fully to the University of Pennsylvania, and he has in place an approved plan for managing any potential conflicts arising from the licensing of these patents. N.P. served on the mRNA strategic advisory board of Sanofi Pasteur in 2022 and the advisory board of Pfizer in 2023 and 2024. N.P. is a member of the Scientific Advisory Board of AldexChem and BioNet Asia. C.W.M. is a co-founder and CSO for Tiba Biotech and Accurius Therapeutics, is an advisor to Adjuvant Capital and the International Vaccine Institute, and works as a consultant for Vaccines and Viral Vectors. J.S.C. is a co-founder of Tiba Biotech and Revela Bio. C.W.M., J.S.C., J.S.M., and P.T. hold equity in Tiba Biotech. The laboratory of A.B.W. received unrelated sponsored research agreements from Third Rock Ventures during the conduct of the study.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Total Rhesus Ig	LifeSpan Bioscineces Inc	LS-347340
Biotin-conjugated anti-monkey IgG antibody	Rockland	617-106-012; RRID:AB_218723
Horseradish peroxidase (HRP) Avidin D	Vector Laboratories	A-2004; RRID:AB_2336507
Anti-monkey IgG/IgA/IgM antibody-heavy and light chain	Rockland	617-101-130; RRID:AB_218708
GAMon/IgG(Fc)/PO	Nordic Mubio	GAMon/IgG(Fc)/7S
Goat anti-Human IgG1 (Heavy chain) Recombinant Secondary Antibody, HRP	ThermoFisher	A55739; RRID:AB_2925764
Anti-CD3 (clone SP34-2)	BD Biosciences	562877; RRID_AB_2737860
Anti-CD3 (clone SP34-2)	BD Biosciences	566517; RRID:AB_2744378
Anti-CD4 (clone L200)	BD Biosciences	563737; RRID:AB_2687486
Anti-CD8(clone SK1)	BD Biosciences	563919; RRID:AB_2722546
Anti-Interferon gamma (clone B27)	BD Biosciences	557995; RRID:AB_396977
Anti-CD20 (clone 2H7)	BD Biosciences	749954; RRID:AB_2874186
Anti-CXCR5 (MU5UBEE)	Thermofisher	12-9185-42; RRID:AB_11219877
Anti-CXCR3 (clone 1C6)	BD Biosciences	562451; RRID:AB_11153118
Anti-PD-1 (Eh12.2H7)	Biolegend	329920; RRID:AB_10960742
Anti-CD71 (clone L01.1)	BD Biosciences	745747; RRID:AB_2743216
Anti-CD38 (clone OKT10	Nonhuman Primate Reagent Resources	AB_2819278; RRID:AB_
Anti-ICOS (clone C398.4A)	Biolegend	313538; RRID:AB_2687079
Anti-CCR7 (clone 150503)	BD Biosciences	561271; RRID:AB_10561679
LIVE/DEAD™ Fixable Near-IR Dead Cell Stain Kit	Fisher	L34976
Anti-Ki67 (clone B56)	BD Biosciences	561277; RRID:AB_10611571
Anti-CD14 (clone M5E2)	Biolegend	301842; RRID:AB_2561946
Anti-CD16 (clone 3G8)	BD Biosciences	563785
Anti-HLA DR (clone G46-6)	BD Biosciences	745782; RRID:AB_2743240
Anti-CD1c (clone L161)	Biolegend	331536; RRID:AB_2629760
Anti-CD141 (clone 1A4)	BD Biosciences	565084; RRID:AB_2739058
Anti-CD11c (clone 3.9)	BD Biosciences	748288; RRID:AB_2872716
Anti-CD159 (clone NKG-2A)	Miltenyi Biotec	130-113-563; RRID:AB_2726170
Anti-CD69 (clone FN50)	BD Biosciences	612817; RRID:AB_2870141
Anti-CD86 (clone IT2.2)	Biolegend	305430; RRID:AB_2563824
Anti-CD80 (clone L307.4)	BD Biosciences	557227; RRID:AB_396606
Anti-CD28	BD Pharmingen	555725; RRID:AB_396068
Anti-CD49d	BD Pharmingen	555501; RRID:AB_2130052
Anti-CD45RA (clone 5H9)	BD Biosciences	556626; RRID:AB_396198
Anti-CD40L (clone TRAP1)	BD Biosciences	748983; RRID:AB_2873383
Anti-OX40 (clone L106)	BD Biosciences	340420; RRID:AB_400027
Anti-4-1BB (clone 4B4-1)	BD Biosciences	550890; RRID:AB_398477
MEDI8852	Andrew Ward's laboratory stock	N/A
Bacterial and virus strains
H5/1Cal09(H1N1)N1Cal09(H1N1)	Florian Krammer’s laboratory stock	N/A
cH6/1Cal09(H1N1)N1Cal09(H1N1)	Florian Krammer’s laboratory stock	N/A
cH8/1Cal09(H1N1)N1Cal09(H1N1)	Florian Krammer’s laboratory stock	N/A
Quadrivalent 2020/2021 seasonal influenza vaccine	Sanofi Pasteur	Lot #UT7011KA
A/Guangdong-Maonan/SWL1536/2019 (H1N1) pdm09-like	Florian Krammer’s laboratory stock	N/A
A/Hong Kong/2671/2019 (H3N2)-like	Florian Krammer’s laboratory stock	N/A
B/Washington/02/2019-like	Florian Krammer’s laboratory stock	N/A
B/Phuket/3073/2013-like	Florian Krammer’s laboratory stock	N/A
A/Netherlands/602/2009 (H1N1)	Florian Krammer’s laboratory stock	N/A
cH6/1Cal09(H1N1)N5	Florian Krammer’s laboratory stock	N/A
cH8/1Cal09(H1N1)N1Cal09(H1N1)	Florian Krammer’s laboratory stock	N/A
HA ectodomain from A/California/4/2009 with E47K stabilizing mutation in HA2	Andrew Ward'slaboratory stock	N/A
Biological samples
PBMCs	Rhesus Macaques	Emory National Primate Reseach Center macaque colony
BMPC	Rhesus Macaques	Emory National Primate Reseach Center macaque colony
Serum	Rhesus Macaques	Emory National Primate Reseach Center macaque colony
Fine needle Aspirates (FNAs)	Rhesus Macaques	Emory National Primate Reseach Center macaque colony
Chemicals, peptides, and recombinant proteins
R848	InvivoGen	tlrl-r848
IL-2	Peprotech	200-02
TMB 2-Component	KPL	5120-0047
Meso Scale Discovery U-PLEX Viral Combo 1 (NHP)	Percision for Medicine	# K15344K-
GolgiStop	BD Pharmingen	554724
GolgiPlug	BD Pharmingen	555029
Cytofix/Cytoperm	BD Biosciences	554722
Perm wash buffer	BD Biosciences	554723
Receptor destroying enzyme	Hardy Diagnostics	370013
DSPC	Avanti Polar Lipids	860368P-10mg
DMG-PEG2000	Avanti Polar Lipids	880151P-5g
Cholesterol	Sigma	C8667-5G
Protein A resin	Cytiva	17127903
Superdex200 10/300GL Column	GE Healthcare Life Sciences	28990944
Uranyl Formate	Electron Microscopy Sciences	D310 25 GM
N1 -methylpseudouridine-5′-triphosphate (m1ΨTP)	TriLink	N-1081
Cholesterol (MDNP)	Sigma Aldrich	C8667
CleanCap reagent AG (3′ OMe)	TriLink	N-7413
1,2-distearoyl-sn-glycero-3-phosphocholine	Avanti	850365
polyethylene glycol-lipid	Avanti	770159
Cholesterol (LNP)	Avanti	700100
Critical commercial assays
Quant-iT Ribogreen assay	Life Technologies	R11491
Deposited data
EMD-43686	Electron Microscopy Data Bank	N/A
EMD-43687	Electron Microscopy Data Bank	N/A
EMD-43689	Electron Microscopy Data Bank	N/A
EMD-43690	Electron Microscopy Data Bank	N/A
EMD-43691	Electron Microscopy Data Bank	N/A
EMD-43692	Electron Microscopy Data Bank	N/A
EMD-43693	Electron Microscopy Data Bank	N/A
EMD-43694	Electron Microscopy Data Bank	N/A
EMD-43695	Electron Microscopy Data Bank	N/A
EMD-43696	Electron Microscopy Data Bank	N/A
EMD-43697	Electron Microscopy Data Bank	N/A
EMD-43698	Electron Microscopy Data Bank	N/A
EMD-43699	Electron Microscopy Data Bank	N/A
SDY2537	ImmPort	N/A
GSE304256Chimeric hemagglutinin-based universal influenza mRNA vaccine induces protective immunity and bone marrow plasma cells in rhesus macaques	Gene Expression Omnibus (GEO)	N/A
GSM9145993 Sample-1 Day-0 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145994 Sample-2 Day-0 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145995 Sample-3 Day-0 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145996 Sample-4 Day-0 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145997 Sample-5 Day-0 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145998 Sample-6 Day-0 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9145999 Sample-7 Day-0 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146000 Sample-8 Day-0 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146001 Sample-9 Day-0 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146002 Sample-10 Day-0 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146003 Sample-1 Day-1 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146004 Sample-2 Day-1 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146005 Sample-3 Day-1 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146006 Sample-4 Day-1 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146007 Sample-5 Day-1 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146008 Sample-6 Day-1 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146009 Sample-7 Day-1 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146010 Sample-8 Day-1 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146011 Sample-9 Day-1 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146012 Sample-10 Day-1 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146013 Sample-1 Day-2 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146014 Sample-2 Day-2 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146015 Sample-3 Day-2 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146016 Sample-4 Day-2 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146017 Sample-5 Day-2 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146018 Sample-6 Day-2 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146019 Sample-7 Day-2 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146020 Sample-8 Day-2 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146021 Sample-9 Day-2 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146022 Sample-10 Day-2 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146023 Sample-1 Day-4 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146024 Sample-2 Day-4 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146025 Sample-3 Day-4 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146026 Sample-4 Day-4 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146027 Sample-5 Day-4 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146028 Sample-6 Day-4 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146029 Sample-7 Day-4 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146030 Sample-8 Day-4 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146031 Sample-9 Day-4 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146032 Sample-10 Day-4 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146033 Sample-1 Day-7 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146034 Sample-2 Day-7 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146035 Sample-3 Day-7 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146036 Sample-4 Day-7 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146037 Sample-5 Day-7 LNP-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146038 Sample-6 Day-7 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146039 Sample-7 Day-7 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146040 Sample-8 Day-7 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146041 Sample-9 Day-7 sam-RNA	Gene Expression Omnibus (GEO)	N/A
GSM9146042 Sample-10 Day-7 sam-RNA	Gene Expression Omnibus (GEO)	N/A
Experimental models: Cell lines
Baby hamster kidney (BHK) cells	ATCC	CCL-10
Experimental models: Organisms/strains
Male Rhesus Macaques	Emory National Primate Research Center	N/A
Software and algorithms
Flowjo	Flowjo	https://www.flowjo.com/
Graphpad	GraphPad	https://www.graphpad.com/
Leginon	Potter et al.112	N/A
Appion	Lander et al.113	N/A
Relion/3.0	Scheres et al.114	N/A
UCSF Chimera	Pettersen et al.115	N/A
UCSF ChimeraX	Pettersen et al.115	N/A
Mabtech Apex 2.0	Mabtech	N/A
Immunospot CTL counter and Image Acquisition 4.5 software	Cellular Technology	N/A
Other
Paxgene tubes	BD Biosciences	761165
GLOBINclear™-Human Kit, for globin mRNA depletion	ThermoFisher Scientific	AM1980
Clontech SMART-Seq v4 Ultra Low Input RNA kit	Takara	634894
Nextera XT DNA Library Preparation kit	Illumina	FC-131-1024
400 mesh copper grids	Electron Microscopy Sciences	0400-Cu
4200 TapeStation System	Agilent	G2991BA
NanoAssemblr Ignite microfluidic mixer	Precision Nanosystems	N/A
Malvern Zetasizer Nano ZS	Malvern Panalytical	N/A
Tecnai F20 electron microscope	FEI	ZEN3600
AKTA Pure chromatography system	Cytiva	N/A
TemCam F415 CMOS camera	TVIPS	29018224

Experimental model and study participant details

Rhesus macaques

Ten male Indian rhesus macaques (RMs) around three years of age from the Emory National Primate Research Center breeding colony were obtained in accordance with the Animal Welfare Act and the National Institute of Health (NIH, Bethesda, MD) Guide for the Care and Use of Laboratory Animals using protocols approved by Emory University Institutional Animal Care and Use Committee (IACUC).

Cell lines

Baby hamster kidney (BHK) cells were purchased from ATCC (cat # CCL-10) and were not tested for mycoplasma contamination.

Mice

Six to eight-week-old female BALB/c mice were used for passive transfer studies. All animal experiments and procedures were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin–Madison School of Veterinary Medicine (protocol no. V006426-A04). Animals were acclimated to the facility conditions (25–28°C and 35–45% humidity) before the start of experiments, allowed access to food and water (ad libitum), kept on a 12/12 h light/dark cycle, and provided with enrichment. Humane endpoints for euthanasia included body weight loss greater than 25% or inability to remain upright.

Method details

mRNA-LNP production

mRNAs were designed and produced as previously described.¹¹⁶ Briefly, codon-optimized cH8/1 and cH5/1 gene containing plasmids were synthesized (Genscript). Plasmids were then linearized, and a T7-driven in vitro transcription reaction (Life Technologies) was performed to generate mRNA with 101 nucleotide long poly(A) tails. Capping of mRNA was performed in concert with transcription through addition of a trinucleotide cap1 analog CleanCap and N1-methylpseudouridine-5′-triphosphate (TriLink) was incorporated into the reaction instead of UTP. Cellulose-based purification of mRNA was performed as described.¹¹⁷ mRNAs were then checked on an agarose gel before storing at −20°C. Purified mRNAs were encapsulated in lipid nanoparticles using a self-assembling ethanolic lipid mixture of an ionizable cationic lipid, 1,2-distearoyl-sn-glycero-3-phosphocholine, cholesterol, and a polyethylene glycol-lipid as previously described.¹¹⁸ This mixture was rapidly combined with an aqueous solution containing mRNA at acidic pH. The ionizable cationic lipid (pKa in the range of 6.0–6.5), proprietary to Acuitas Therapeutics (Vancouver, Canada) and LNP composition are described in the patent application WO 2017/004143. The average hydrodynamic diameter was ∼80 nm with a polydispersity index of 0.02–0.06 as measured by dynamic light scattering using a Zetasizer Nano ZS (Malvern Instruments Ltd, Malvern, UK) and an encapsulation efficiency of ∼95% as determined using a Quant-iT Ribogreen assay (Life Technologies).

Self-amplifying mRNA formulation

The cH8/1 and cH5/1 coding sequences were kindly provided to Tiba Biotech by Florian Krammer, codon optimized for humans, and cloned into a DNA plasmid encoding a self-amplifying mRNA template based on the genome of Venezuelan equine encephalitis virus. RNA was synthesized by run-off in vitro transcription from an upstream T7 promoter and subsequent enzymatic capping essentially as reported previously to generate a Cap1 structure.¹¹⁹ Expression potency of each lot of self-amplifying mRNA produced for this study was validated by transfection of baby hamster kidney (BHK) cells and immunoblot using the appropriate antibodies (anti-H8 and anti-H5), also provided by the Krammer laboratory and as previously described.^120,121 The capped self-amplifying RNA was dissolved in 10mM citrate buffer and mixed using a NanoAssemblr Ignite microfluidic mixer (Precision Nanosystems) with a proprietary ethanolic solution of an ionizable modified dendron provided by Tiba Biotech (at a mass ratio of 7.5), plus cholesterol, 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), and 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (DMG-PEG2000) (Avanti Polar Lipids) in a molar ratio 1: 2.88: 0.6: 0.065 respectively. Nanoparticles were dialyzed against sterile, endotoxin-free phosphate-buffered saline (PBS) using 20,000 Da molecular weight cutoff dialysis cassettes and sterile filtered using 0.2 micron poly(ether sulfone) filters. Nanoparticle diameter and polydispersity index were assessed by dynamic light scattering and encapsulation efficiency measured by RiboGreen (Thermo) dye-exclusion assay. The average hydrodynamic diameter was ∼95 nm with a polydispersity index of ∼0.07 as measured by dynamic light scattering using a Malvern Zetasizer Nano ZS and an encapsulation efficiency of ∼96% as determined using RiboGreen (Thermo) dye-exclusion assay.

Immunizations

Ten RMs were inoculated intramuscularly (IM) with the full human dose of quadrivalent 2020/2021 seasonal influenza vaccine (QIV, Sanofi Pasteur) in the deltoid on weeks 0 and 6. Eighteen weeks post QIV prime, RMs were immunized IM with either 100μg of mRNA-LNP or 30 μg of sam-MDNP expressing cH8/1 followed by 100μg of mRNA-LNP or 30 μg of sam-MDNP expressing cH5/1 twelve weeks later.

Plasmablast and bone marrow plasma cell assays

ELISpot assays were performed as previously described with certain modifications.¹²² Total Ig (LS-347340 LifeSpan Bioscineces Inc) was diluted to 5 mg/mL in 1X PBS, and the QIV, cH5/1, cH6/1, cH8/1 antigens were diluted to 1 mg/mL in PBS. Each well of the ELISpot plates (MSHAN4B50, Millipore-Sigma) was coated with 100 μl of the above antigens and incubated overnight at 4°C. The plates were then washed with wash buffer I (1X PBS, 0.05% Tween 20) one time and three times with wash buffer II (1X PBS). The plates were then blocked with 100 μl per well of complete Roswell Park Memorial Institute (RPMI) medium {10% Fetal bovine serum (FBS) from Corning (cat# 35-011-CV), 1:100 dilution of penicillin G and streptomycin (Pen/Strep) from Lonza (cat# 17-602E) in 1X RPMI 1640) and incubated in a 5% CO2 incubator at 37°C for two hours. The (PBMCs) or bone marrow (BM) cells were re-suspended at 107cells/ml. Then, 50mL (total of ∼5x10∧5 cells/well) of cell suspension per well was added to the first well of each column with successive 3-fold dilutions thereafter. After 16–18 h incubation, the plates were washed four times with each wash buffer I and II. The plates were incubated with 100 μl biotin-conjugated anti-monkey IgG antibody (in 1:2000, Cat# 617-106-012, Rockland) for two h at room temperature (RT). After washing the plates four times with wash buffer I containing 1% fetal calf serum (FCS), the plates were incubated with horseradish peroxidase (HRP) Avidin D (in 1:5000, Cat# A-2004, Vector Laboratories) for two hours at room temperature (RT). The secondary antibody and the HRP Avidin D were diluted using wash buffer I containing 1% fetal calf serum (FCS). Then, the plates were washed four times with wash buffer I and developed using 3-amino-9-ethylcarbazole AEC (0.3 mg/mL AEC in 0.1M Na-acetate buffer, pH 5.0, and 0.03% hydrogen peroxide). The plates were dried and detected using the Immunospot CTL counter and Image Acquisition 4.5 software (Cellular Technology). Spots were counted manually. Antibody secreting cells ASCs were reported as per million PBMCs or percentage of total bone marrow plasma cells.

ELISpot assay for memory B cell assay

PBMCs were thawed. The cells were then cultured in complete RPMI medium containing R848 (2.5 mg/mL) (Cat# tlrl-r848, InvivoGen) and IL-2 (1000 U/ml) (Cat# 200-02, Peprotech) at a concentration of 1 million cells/ml of the medium. On day 3, half of complete medium with IL-2 and R848 was replaced with fresh medium. On day 6 cells were washed and counted. Cells were re-suspended in complete RPMI medium. Five hundred thousand cells were seeded on the ELISpot plates coated with antigens. The cells were incubated for 8 h. ELISpot was performed same as the ELISpot assay performed above. All the reagents were same as the above ELISpot assay except the total IgG (anti-monkey IgG/IgA/IgM antibody-heavy and light chain; cat # 617-101-130, Rockland Inc.) The images of the plates were taken on Mabtech IRIS fluorospot/ELISpot reader and spots were counted manually.

Binding antibody responses

Binding antibody response to seasonal QIV, cH8/1, cH5/1, cH6/1, and mini-H1 were determined by enzyme-linked immunosorbent assay (ELISA) as previously described with some modifications.^47,123 Briefly, Costar high-binding microtiter plates (Corning Life Sciences) were coated with 1 μg/mL of either QIV, cH8/1, cH5/1, cH6, or mini H1 in PBS and incubated overnight at 4°C. The next day, plates were washed with 0.05% PBS-Tween (PBST) and blocked with 5% whey diluted in PBST for 30 min at room temperature (RT). RM serum was serially diluted 3-fold, added to the plates, and incubated for 1 h at RT. Plates were again washed with PBST and bound serum antibody detected using a 1:10,000 dilution of anti-rhesus IgG (Nordic MUbio, GAMon/IgG(Fc)/7S) conjugated to horseradish peroxidase (HRP). Plates were again washed with PBST and tetramethylbenzidine substrate (KPL) added. The reaction was stopped after 20 min by adding 100 μL of 2N H₂SO₄. Plates were read at 450 nm wavelength using a Varioskan Lux (Thermoscientific) reader. Fifty percent maximal binding (EC50) values were calculated in prism using log(inhibitor) vs. normalized response (variable slop) with a cutoff of 0.1.

Meso scale discovery

The Meso Scale Discovery (MSD) U-PLEX Viral Combo 1 (NHP) kit (cat# K15344K-1) was used to measure cytokine concentration in plasma and cell supernatant. The kit measures 19 analytes: G-CSF, GM-CSF, IFN-α2a, IFN-γ, IL-1RA, IL-1β, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12p70, IP-10, MCP-1, MIP-1α, TNF-α, and VEGF-A. The MSD-provided protocol for U-PLEX assays was followed with no changes. Briefly, after bringing all reagents to room temperature, the two multiplex coating solutions were added to the appropriate 96-well plate (50 μL per well), and the plates were covered and incubated for one hour while shaking at 600 rpm on a Lab-Line Instruments Titer plate shaker. After washing three times with 1x MSD wash buffer (cat# R61AA-1), two replicates of serial 5-fold dilutions of calibrator mixes were added to the appropriate plates, along with two blank wells. Plasma was diluted 2-fold and cell supernatants were diluted 3-fold and added to appropriate wells (50 μL per well). Plates were incubated while shaking for one hour and washed again, then detection antibody master mix was added to the appropriate plates. After another one-hour incubation and wash, MSD GOLD Read Buffer B was added to each well and the plate was immediately read using the MESO QuickPlex SQ 120 MM instrument.

Stimulation of RM PBMCs with conventional and self-amplifying luciferase expressing mRNA encapsulated in LNP and MDNP

PBMCs were isolated from four different RMs and seeded in a 96-well plate at a concentration of 250K cells/well. PBMCs were then incubated with 1 μg of four different mRNA constructs: mRNA-LNP, sam-MDNP, sam-LNP, and mRNA-MDNP encoding luciferase mRNA for 0, 6, 24, 48, and 72 h. At each indicated time point, cells were spun down, supernatants collected and stored at −20°C for MDS analysis.

Competition ELISAs

Similarly to how binding antibody responses were determined, costar high-binding microtiter plates (Corning Life Sciences) were coated with 1 μg/mL of mini-HA in PBS overnight at 4°C. Next, plates were washed with 0.05% PBS-Tween (PBST) and blocked with 5% whey diluted in PBST for 30 min at room temperature (RT). RM serum was serially diluted 3-fold, added to the plates, and incubated for 1 h at RT. Plates were again washed with PBST and MEDI8852 mAb (kindly provided by Andrew Ward) was added at a saturating concentration of 0.125 μg/mL for 1 h at RT. Plates were washed with PBST and incubated with Goat anti-Human IgG1 (Heavy chain) Recombinant Secondary Antibody, HRP (ThermoFisher, cat#A55739) at 1:2000 for 1 h at RT.). Plates were again washed with PBST and tetramethylbenzidine substrate (KPL) added. The reaction was stopped after 20 min by adding 100 μL of 2N H₂SO₄. Plates were read at 450 nm wavelength using a Varioskan Lux (Thermoscientific) reader. Inhibition determined by dividing MEDI8852 binding in the presence of RM serum by MEDI8852 binding in the absence of RM serum (MEDI8852 saturation).

Phenotype response

T cell staining. Whole blood was stained with a cocktail of surface antibodies specific for CD3 (clone SP34-2; BD Biosciences), CD4 (clone L200; BD Biosciences), CD8(clone SK1; BD Biosciences), CD20 (clone 2H7; BD Biosciences), CXCR5 (MU5UBEE, Thermofisher), CXCR3 (clone 1C6, BD Biosciences), CCR4 (clone 1G1, BD Biosciences), CCR6 (clone G034E3, Biolegend), PD-1 (Eh12.2H7; Biolegend), CD95 (clone DX2, BD Biosciences), CD71 (clone L01.1; BD Biosciences), CD38 (clone OKT10, Nonhuman Primate Reagent Resources), ICOS (clone C398.4A; Biolegend), CCR7 (clone 150503; BD Biosciences) and LIVE/DEAD Near-IR Dead Cell stain (Fisher) for 20 min in the dark at RT. Red blood cells were lysed for 10 min with Lysing Solution (BD Biosciences, Franklin Lakes, NJ) in the dark. Cells were spun down at 2000 rpm for 5 min and lyses buffer decanted. Cells were washed with 2mL of FACS (PBS with 2% FBS and 0.05% sodium azide) wash buffer and again spun down at 2000 rpm for 5 min. FACS washed was discarded and cell were fixed and permeabilized with FoxP3/Transcription Factor staining buffer set (eBioscience) following the manufacturer’s instructions and stained with Ki67 (clone B56; BD Biosciences) for 30 min in the dark at RT. Cells were again washed with FACS buffer and cell acquired on LSR Fortessa (BD, Biosciences).

Innate immune cell staining: Plasma was removed from whole blood by spinning at 2000 rpm from 10 min to prevent plasma antibody from binding to the CD16 receptor and blocking anti-CD16 binding. Blood was reconstituted with PBS back to the original volume. Blood was then stained with LIVE/DEAD Near-IR Dead Cell stain and antibodies specific for CD3, CD20, CD14 (clone M5E2, Biolegend), CD16 (clone 3G8, BD Biosciences), HLA DR (clone G46-6, BD Biosciences), CD1c (clone L161, Biolegend), CD141 (clone 1A4, BD Biosciences), CD123 (clone 7G3, BD Biosciences), CD11c (clone 3.9, Biolegend), CD159 (clone NKG-2A, Miltenyi Biotec), CD69 (clone FN50, BD Biosciences), CD86 (clone IT2.2, Biolegend), and CD80 (clone L307.4, BD Biosciences) and incubated at RT for 20 min in the dark. From this point cells were treated the same as for T cell staining.

Intracellular cytokine staining

Intracellular cytokine staining (ICS) was performed as described previously¹²⁴ with some modifications. PBMCs we isolated from blood as previously described and stimulated with influenza cH8 head (peptides 1–76), cH5 head (peptides 1–73), or H1 stalk (peptides 77–139) in the presence of 1 μg/mL anti-CD28 and anti-CD49d (BD Pharmingen, San Diego, CA) in RPMI 1640 complete media (containing 10% FBS, HEPES, Gentamycin, and Penicillin-Streptomycin).

The Influenza HA cH8 head, cH5 head, and H1 stalk peptide pools contained overlapping peptides (15mers overlapping by 11). After 2 h of stimulation at 37°C in 5% CO₂, GolgiStop (0.5 mg/mL; BD Pharmingen) and GolgiPlug (0.5 mg/mL; BD Pharmingen) was added to the cells. After an additional 4 h of incubation, the cells were placed at 4°C overnight. The next day, cells were washed in FACS wash buffer and surface stained with CD4-BV650 (clone L200; BD Pharmingen), anti-human CD8-AmCyan (clone SK1; BD Biosciences), and Live/Dead Fixable Near-IR APC-Cy7 stain (Fisher) for 20 min at 4°C. Cells were again washed with FACS wash buffer and permeabilized with Cytofix/Cytoperm (BD Biosciences) for 25 min at 4°C. Next, cells were washed twice with Perm wash buffer (BD Biosciences) and intracellular staining carried out using a combination of anti-human CD3-BV421 (clone SP34-2; BD Biosciences) and anti-human interferon gamma (IFN-g)-Alexa 700 (clone B27; BD BiosciencesCells). After a 25 min incubation at 4°C, cells were again washed with Perm wash buffer followed by FACS wash buffer. Cells were resuspended in FACS wash buffer and acquired on an LSRII (BD Immunocytometry Systems, San Jose, CA) and analyzed using FlowJo software (Tree Star, Ashland, OR).

Activation-induced markers (AIM) assay

Antigen-specific T cell responses in the lymph nodes (LN) were determined though the AIM assay. Cells isolated from lymph nodes (LN) fine needle aspirates (FNA) were stimulated with cH8/1 head, cH5/1 head, and H1 stalk peptides overnight at 37°C. The following day cells were washed with FACS wash and incubated with antibodies specific for CD3, CD4, CD8, CD20, CXCR5, CD69, PD-1, CD25 (clone 2A3; BD Biosciences), CD45RA (clone 5H9; BD Biosciences), CD40L (clone TRAP1; BD Biosciences), OX40 (clone L106; BD Biosciences), and 4-1BB (clone 4B4-1; BD Biosciences) for 45 min are RT in the dark.

Hemagglutination inhibition (HAI) assay. Non-human primate (NHP) sera was assessed for antibodies with HAI activity to A/Guangdong-Maonan/SWL1536/2019 (H1N1) pdm09-like, A/Hong Kong/2671/2019 (H3N2)-like, B/Washington/02/2019-like, B/Phuket/3073/2013-like, A/Netherlands/602/2009 (H1N1) and cH6/1N5 and cH8/1N1 viruses. Briefly, sera were treated prior to testing with receptor destroying enzyme (RDE, Denka-Seiken) to inactivate unspecific viral inhibitors. 25μL of sample were incubated with 75μL of RDE at 37°C for 18–20 h. Reaction was stopped by addition of 75μL of sodium citrate 2.5% (w/v) (Fisher Scientific), and incubation at 56°C for 60 min. For a final serum dilution of 1:10, 75μL of PBS were added. Samples were serially diluted (2-fold) in PBS up to 1:5,120 using V-bottom microtiter plates.

Each corresponding virus was added to serially diluted sera (4 hemagglutination units per well (HAU)). Plates were incubated at room temperature (RT) for 30 min. Chicken red blood cells (RBCs, Viromed Laboratories) were added to the serum–virus mix at a concentration of 0.5% (v/v) in PBS and incubated at RT for 40 min. Plates were tilted and read. HAI titers were reported as the reciprocal of the last serum dilution to inhibit RBC agglutination. Non-reactive sera were assigned a value of 5 for plotting purposes. Sera with titers ≥5,120 were retested at a higher initial dilution to get the final titer. As positive control of hemagglutination, a polyclonal sera with high breadth of reactivity was used.

Passive antibody transfer

Serum from mRNA-LNP and sam-MDNP groups were pooled by equal volumes for each time point. 6–8-week-old female BALB/c mice were administrated 200 μL of serum by intraperitoneal injection (n = 5/group). Twenty-four hours later, prior to challenge, small amounts of blood were collected via submandibular route to confirm successful serum transfer (via ELISA), and then the mice were challenged with 10 LD₅₀ virus A/Netherlands/602/2009 or cH6/1N5 (tittered in the context of animals passively transferred with RM serum 8 weeks post QIV booster immunization). The mice are monitored daily for survival and body weight loss over a period of 14-day post infection. Mice showing more than 25% of body weight loss are considered to have reached the experimental endpoint and were euthanized.

Transcriptomics

Blood was collected into PAXgene Blood RNA tubes (BD Biosciences) and the RNA was extracted using the MagMAX for Stabilized Blood Tubes RNA Isolation Kit, compatible with PAXgene Blood RNA Tubes (ThermoFisher Scientific). RNA quality was assessed using a TapeStation 4200 (Agilent) and then one microgram of total RNA was subjected to globin transcript depletion using the GLOBINclear Kit, human (ThermoFisher Scientific). Ten nanograms of the globin-depleted RNA was used as input for cDNA synthesis using the Clontech SMART-Seq v4 Ultra Low Input RNA kit (Takara Bio) according to the manufacturer’s instructions. Amplified cDNA was fragmented and appended with dual-indexed bar codes using the Nextera XT DNA Library Preparation kit (Illumina). Libraries were validated by capillary electrophoresis on a TapeStation 4200 (Agilent), pooled at equimolar concentrations, and sequenced with PE100 reads on an Illumina. Sequencing data was aligned to Mmul10 Ensembl release 100 using STAR 2.7.3a.¹²⁵ Transcript abundance estimates were calculated using the STAR ‘;--quantMode GeneCounts' option. DESeq2 was used for transcript normalization and differential expression analysis.¹²⁶ Genes were considered significantly differentially expressed if the Benjamini-Hochberg adjusted p-value (fdr = 0.05) was below 0.05. Gene set enrichment analysis was performed using the fGSEA R package¹²⁷ with pathway definition defined in the Hallmark collection of MSigDB.^128,129 Pathways were considered significantly altered if the Benjamini-Hochberg adjusted p-value (fdr = 0.05) was below 0.05.

EMPEM

As previously described in Turner et al.⁷¹ NHP serum was heat inactivated for 1 h at 56°C and protein A was added to 1mL of serum and gently mixed at 4°C for 3 days. An average of 6mg of IgG was isolated from each sample and then digested into Fab and Fc products. Digested Fab/Fc was purified over an S200i column on an AKTA (GE). 10 μg of H1 HA (A/California/04/2009 (H1N1) with E47K stabilizing mutation in HA2) was added to clean Fab/Fc and incubated overnight at 4°C and was purified over an S200i column the following day. The size exclusion chromatography (SEC) peak corresponding to the polyclonal complex was combined and immediately added to a negative stain electron microscope (nsEM) grid, and stained with 2% uranyl formate. Negative stain EM data was collected on a Tecnai 200kv microscope equipped with a Tietz video and image processing systems (TVIPS) camera. Automated acquisition was performed with Leginon¹¹² and processed on Appion.¹¹³ DoG Picker¹³⁰ was used to choose particles, 2D and 3D classification was done in Relion version 3.0.¹¹⁴ Figures were made with UCSF Chimera.¹¹⁵

Quantification and statistical analysis

All experiments were conducted with 5 biological replicates per group. Statistical analysis was performed using GraphPad Prism v9 software. For statistical comparisons between mRNA-LNP and sam-MDNP groups, a non-parametric Mann-Whitney U test was utilized. For comparisons between time points within the same group, a two-way AVOVA was used. The Spearman rank test was used for correlations. Survival curves were analyzed using the Kaplan-Meier method and the log rank (Mantel-Cox) test. Statistical significance was defined as a p value of <0.05. For transcriptome analyses, we used R package (version 3.6.1) to calculate Pearson correlation coefficients along with p-values (with null hypothesis of no association) between normalized gene expression data and vaccine response. Normalized gene expression values were obtained using DESeq2.¹²⁶ Figures 1A, 4A, 5A, and 7C were created with BioRender.com.

Published: September 25, 2025