Allergen-specific mRNA–lipid nanoparticle therapy for prevention and treatment of experimental allergy in mice

We hypothesized that the ability of allergen-specific mRNA-LNPs to induce a Th1 response and increase the level of antigen-specific IgGs (,,) could be utilized for protection against allergic responses. To test this hypothesis, we employed an allergy model using the egg antigen ovalbumin (OVA), as tools existed for detailed examination of allergen-specific immune responses. First, we tested the efficiency and specificity of the nucleoside-modified OVA-mRNA-LNPs in inducing T cell responses in mice engineered to express T cells with OVA-specific receptors (OTII cells). Mice were injected with purified naive OTII cells expressing the CD45.1 marker and treated intramuscularly (i.m.) with1-methylpseudouridine (m1Ψ)–mRNA–LNPs encoding OVA protein (OVA-mRNA-LNP), empty LNP, or phosphate-buffered saline (PBS) twice on days 0 and 7 (; supplemental material available online with this article;). The response of antigen-specific CD4T cells in lymph nodes (LNs) was analyzed 4 days later. The administration of OVA-mRNA-LNP induced marked expansion of donor OTII cells, increased the activation marker CD44, and enhanced the production of TNF-α and IL-2 cytokines in the local LNs, indicating an efficient and antigen-specific T cell response to the vaccine (). There were an increase in the proportion of donor OTII cells and host CD4T cells producing IFN-γ and expressing T-bet, which are key characteristics of Th1 cells, and an elevation in percentages of Bcl6PD-1cells, indicative of Tfh cells (), that were in line with previous reports (,,). We did not observe elevation in IL-17A, IL-4, or Foxp3 levels () in response to OVA-mRNA-LNP. No activation of host or OTII T cells was detected following treatment with PBS or empty LNP. ^{21 22 38} Supplemental Figure 1A Supplemental Figure 1B Supplemental Figure 1C ^{32 37 40} Supplemental Figure 1D N https://doi.org/10.1172/JCI194080DS1↗ + + + +

To assess the dose range of allergen-specific mRNA-LNP required for effective T cell activation, mice receiving donor OTII cells were treated with a single i.m. injection of increasing amounts of OVA-mRNA-LNP. By day 7, the analysis of local LNs showed strong OTII cell expansion at doses of 2–5 μg of OVA-mRNA-LNP, with over 80% of the cells expressing the activation marker CD44 (). These doses also correlated with increased frequencies of effector OTII cells producing TNF-α or IFN-γ or exhibiting a Bcl6PD-1phenotype (). Concurrently with the T cell response, a dose-dependent increase in OVA-specific IgG1, IgG2a, and IgG2b levels was observed in serum 18 days after treatment (). These results demonstrate the specificity and high potency of the OVA-mRNA-LNP in activating and differentiating Th cells toward Th1 and Tfh subsets. Supplemental Figure 2A Supplemental Figure 2B Supplemental Figure 2C + +

To evaluate the impact of the vaccine under allergen-induced disease conditions, mice were subsequently given different doses of OVA-mRNA-LNP, ranging from 0 to 5 μg, sensitized with OVA protein mixed with alum adjuvant (OVA+Alum), and then challenged with OVA protein via intratracheal (i.t.) and intranasal (i.n.) routes (). Notably, mice receiving a higher dose of the OVA-mRNA-LNP exhibited a substantial reduction in the eosinophil count in the bronchoalveolar lavage fluid (BALF) () and a decreased percentage of GATA3and IL-5IL-13CD4T cells in the lung upon OVA challenge (), indicating a dose-dependent effect of the OVA-mRNA-LNP against Th2 responses in the airways. Concurrently, the number of CD8T cells and neutrophils in BALF, as well as the frequency of IFN-γCD4T cells in the lungs, increased in an OVA-mRNA-LNP dose–dependent manner (). The number of CD4T cells in BALF and the percentage of Foxp3Tregs in the lungs did not show marked changes. Figure 1A Figure 1B Figure 1C Figure 1, B and C + + + + + + + + +

To further improve the anti-allergic outcomes, we administered a booster dose of OVA-mRNA-LNP a week after the first immunization, followed by OVA+Alum sensitization and airway challenge with OVA protein (). Mice given 2 doses of the OVA-mRNA-LNP exhibited a pronounced reduction in eosinophils in the BALF compared with those given a single dose, as indicated by flow cytometry analyses (, cell number, and, cell frequency). This finding was substantiated through histologic staining of lung tissue using hematoxylin and eosin (H&E) and anti–major basic protein (anti-MBP) staining to detect eosinophils (). While a modest elevation in eosinophils was detected in the BALF () and lung tissue () of mice pretreated with 2 doses of OVA-mRNA-LNP in response to allergen administration (compared with naive mice), eosinophil levels remained nearly 100-fold lower than those in the LNP group, aligning with the decrease in Th2 cell responses observed in the lung. Although the proportion of CD4T cells was constant between mice receiving 1 or 2 doses of the OVA-mRNA-LNP (), 2 doses led to a more notable decline in CD4GATA3T cell frequency in the lungs upon OVA challenge, alongside an increase in CD4T cells producing IFN-γ, reflecting intensified Th1 responses (). Additionally, neutrophils and CD8T cells were enriched in the BALF (and), with higher frequencies of CCR5and perforinCD8T cells in the lungs (), underscoring enhanced CD8T cell migration and cytotoxic potential. Histologic examination corroborated the presence of accumulated lymphoid cells within the lung tissue of allergen-challenged, OVA-mRNA-LNP–immunized (single dose) and boosted (second dose) mice (). Taken together, these results demonstrate that immunization with the modified allergen-specific mRNA-LNP reduces Th2 effectors and increases Th1 and cytotoxic CD8 responses after allergen challenge. Supplemental Figure 3A Figure 1D Supplemental Figure 3B Figure 1E Figure 1D Figure 1E Figure 1D Supplemental Figure 3C Figure 1D Supplemental Figure 3B Supplemental Figure 3D Figure 1E + + + + + hi + + +

We aimed to extend the OVA-mRNA-LNP protective readouts to include clinically relevant features in acute and chronic asthma models. In humans, IgE typically mediates allergen-induced symptoms, whereas anti-allergen IgGs provide protection through a blocking mechanism (,). Accordingly, allergen-specific antibody titers were measured after OVA-mRNA-LNP immunization and allergen sensitization of mice (). The OVA-mRNA-LNP–immunized group evidenced strong anti-OVA IgG1 and IgG2 responses, whereas the OVA-mRNA-LNP group compared with the LNP group had markedly reduced anti-OVA IgE upon OVA sensitization (). These findings provide early proof of concept for the potential efficacy of allergen-specific mRNA-LNP vaccination against IgE-mediated responses. ^{41 42} Figure 2A Figure 2B

We next evaluated the cytokine profile in the BALF of airway allergen–challenged mice (). OVA-mRNA-LNP immunization markedly decreased Th2 cytokines, specifically IL-4 and IL-5 (), and the chemokine eotaxin-2 (CCL24), whose level correlated with the frequencies of eosinophils in the lung (). In contrast, factors associated with Th1-skewed responses, such as MIP-1α/β, MIG, IP-10, RANTES, TNF-α, and IFN-γ, were elevated in the OVA-mRNA-LNP compared with the LNP group (,, and). These results were substantiated by quantitative PCR analysis of lung tissue, which showed downregulation of,,,, andand upregulation of,,, andin the OVA-mRNA-LNP group (), highlighting the anti-allergic effects of allergen-specific mRNA-LNP immunizations. Figure 2A Figure 2C Figure 2D Figure 2C Supplemental Figure 4A Supplemental Table 1 Supplemental Figure 4B Il4 Il5 Il13 Ccl11 Ccl24 Ifng Cxcl9 Cxcl10 Ccl5

During the acute phase of experimental asthma (), OVA-mRNA-LNP–immunized mice were protected from developing allergen-induced airway hyperresponsiveness following OVA challenge, as shown by alleviated airway resistance to the inhaled bronchoconstrictor methacholine (), and exhibited reduced lung mucus production, as demonstrated by periodic acid–Schiff (PAS) staining (), compared with the increased airway resistance and mucus secretion observed in LNP-treated mice. The protective effects of OVA-mRNA-LNP immunization were also evident in a chronic asthma model with prolonged OVA exposure (). OVA-mRNA-LNP–treated mice displayed reduced airway resistance (), decreased eosinophilia (), diminished proportion of Th2 (GATA3) cells, and increased IFN-γCD4T cell frequency (), paralleling the acute response. Although CD8T cell counts were also elevated in the chronic model, neutrophil numbers remained low and comparable to those of LNP-treated mice (). Collectively, these results indicate that the allergen-specific mRNA-LNP provides protection against clinical features of experimental asthma in both acute and chronic stages. Figure 2A Figure 2E Figure 2F Figure 2G Figure 2H Figure 2I Figure 2J Figure 2I + + + +

To elucidate the mechanism behind allergen-specific mRNA-LNP vaccination, single-cell RNA sequencing (scRNA-Seq) was conducted on lung samples from naive mice and those pretreated with LNP or OVA-mRNA-LNP, followed by OVA+Alum sensitization and acute OVA airway challenge (). Uniform manifold approximation and projection (UMAP) clustering identified 15 distinct lung cell subpopulations with canonical markers for each subset (, A and B) (–). Transcriptomic analysis revealed that OVA-mRNA-LNP pretreatment reprograms both innate and adaptive immune responses, creating a distinct immunological landscape during allergen challenge. Among innate cells, the most prominent transcriptomic differences between LNP (asthmatic) and OVA-mRNA-LNP allergen-challenged groups were observed in neutrophils, interstitial macrophages, NK cells, dendritic cells, and plasma cells (). Pathway analysis revealed modulation of immune-regulatory pathways, including a shift toward Th1 polarization (e.g., IFN-γ regulation), regulation of adaptive activation through costimulatory checkpoints, and antigen presentation, as well as changes in cellular metabolism (). These alterations collectively indicate a less permissive environment for Th2-driven allergic inflammation and a coordinated shift away from classical allergic responses, at the level of innate immunity. Figure 2A Supplemental Figure 5 ^{43 47} Supplemental Tables 2–6 Supplemental Figure 6A

Among αβ T cells, 5 subpopulations, including naive CD4, naive CD8, activated CD4, activated CD8T cells, and Tregs, were identified by an extended panel of cell markers (). Both scRNA-Seq () and flow cytometry analysis () showed reduced frequencies of naive CD4and CD8T cells in allergen-challenged LNP- and OVA-mRNA-LNP–treated mice compared with naive mice. Activated CD8T cells were enriched in the allergen-challenged OVA-mRNA-LNP group, and activated CD4cells and Tregs were elevated in the LNP group (). + + + + + + + + Figure 3, A and B Supplemental Figure 6B Figure 3C Figure 3C

Further transcriptional analysis of activated CD4T cells revealed a Th2/Th17 signature in LNP-treated (asthmatic) mice and a distinct Th1 signature in OVA-mRNA-LNP–treated mice (). These differentially expressed genes included key lineage-specific transcriptional factors (,,,), receptors (,,,,,,), functional molecules (,,), and cytokines (,,,,,). Flow cytometry analysis corroborated these findings, showing a reduction of IL-5IL-13, GATA3, IL-17A, and ST2CD4T cells and elevation of IFN-γCD4T cells in the OVA-mRNA-LNP group compared with the LNP group (). + + + + + + + + + Figure 3D Figure 3E Irf4 Gata3 Eomes Tbx21 Il17rb Il1rl1 Ccr4 Il2ra Cxcr3 Ccr5 Cd160 Areg Cish Nkg7 Il17a Il5 Il13 Ifng Ccl4 Ccl5

We evaluated the transcriptional profile of activated CD8T cells by calculating module scores () based on the gene signature of human CD8T cells recently identified as the most potent responders following SARS-CoV-2 mRNA vaccination (). This human gene set included 197 genes and was enriched with genes related to cytotoxic and effector functions of T cells, TCR signaling, antigen processing, and metabolism () (). Notably, activated CD8T cells from the OVA-mRNA-LNP group exhibited, on average, higher module score values compared with those of the naive and LNP groups, indicating close resemblance to the human CD8T cells induced by SARS-CoV-2 mRNA vaccination (). Furthermore, flow cytometry analysis revealed that, similar to human CD8T cells induced by the SARS-CoV-2 mRNA vaccine (), the allergen-challenged OVA-mRNA-LNP group showed an elevated frequency of CD38KLRG1CD8T cells, which demonstrated enhanced perforin production and expressed the activation marker CXCR6 (). These results demonstrate that the CD8T cell response following allergen-specific mRNA-LNP vaccination is conserved at least in part across species and vaccine immunogens. + + + + + + – + + ^{48 49} Supplemental Table 7 ⁴⁹ Figure 3F ⁴⁹ Figure 3G

Although CD8T cells may exert anti-allergic effects (), their cytotoxic functions may be undesirable and could contribute to non-allergic inflammation. Previous studies suggest that mTOR inhibitors suppress CD8T cell effector function when coupled with a potent antigenic signal (,). Therefore, we hypothesized that coadministering the OVA-mRNA-LNP with the mTOR inhibitor everolimus may modulate CD8T cell cytotoxic activity in the allergic model (). Indeed, mTOR inhibition attenuated OVA-mRNA-LNP–induced CD8T cells in the lung and the CD38KLRG1CD8T cell subset, including perforin production (), while preserving the anti-allergy effect of the vaccination. This was evidenced by decreased frequencies of Th2 cells in the lungs, lower eosinophil counts in the BALF (), normalized airway hyperresponsiveness (), and reduced inflammatory infiltration in lung tissues (). + + + + + – + ^{50 51 52} Figure 4A Figure 4B Figure 4C Figure 4D Figure 4E

To better approximate physiological conditions, we evaluated the impact of allergen-specific mRNA-LNP vaccination on the immune response to house dust mite (HDM). Mice were immunized with an allergen-specific mRNA-LNP encoding the major HDM allergen Der p1, followed by sensitization with HDM and challenges with naturally purified Der p1 (np-Der p1) protein or HDM extract (). Der p1–mRNA–LNP vaccination increased production of Der p1–specific IgG1 antibodies (), indicating active immunogenicity. Following challenge with np-Der p1 protein, immunized mice exhibited reduced allergic responses, including lessened eosinophilia, decreased Th2 (GATA3and IL-5IL-13) and Th17 (IL-17A) cell frequencies, and lowered mucus production (). Additionally, elevated CD8T cell frequency and enhanced IFN-γ production by CD4T cells were observed in Der p1–mRNA–LNP–vaccinated mice compared with LNP-treated mice (). A reduction in Th2 and Th17 cell populations, along with an increase in IFN-γ–producing cells, was also detected in Der p1–mRNA–LNP mice challenged with HDM extract (), although the anti-allergic effect was not complete (). These results highlight the potential of the Der p1–mRNA–LNP vaccines to counteract HDM-induced allergic inflammation, warranting further research to optimize their efficacy and broader applications. Figure 5A Figure 5B Figure 5, C–E Figure 5, C and D Figure 5F Supplemental Figure 7 + + + + + +

To investigate the therapeutic potential of the allergen-specific mRNA-LNP as an immunotherapy for established allergies, mice were sensitized by application of the allergen to the skin in the presence of a vitamin D analog, calcipotriol (MC903), which mimics the atopic dermatitis condition (). This sensitization was followed by i.m. treatment with either OVA-mRNA-LNP or empty LNP and subsequent repeated airway challenges with OVA (). Sensitization with OVA+MC903 resulted in an increased production of OVA-specific IgE and IgG1, but not IgG2a, antibodies (, day 15). Both OVA-specific IgG1 and IgG2a levels were remarkably amplified following treatment with OVA-mRNA-LNP and remained elevated beyond levels seen following treatment with LNP, even after the OVA challenges (, day 72). The IgE level decreased after sensitization (day 50) regardless of OVA-mRNA-LNP treatment, indicating that mRNA does not worsen humoral allergic responses in allergen-challenged mice. Upon repeated OVA airway challenges, OVA-mRNA-LNP–treated mice exhibited notable reductions in the asthma-related features, including eosinophilia, Th2 response, mucus secretion, and airway hypersensitivity (). Additionally, these mice demonstrated increased frequencies of IFN-γCD4T cells and perforinCD8T cells expressing the CD38KLRG1phenotype (), as was observed in the prophylaxis model (). In contrast, LNP-treated mice developed lung eosinophilia (), a rise in Th2 cells (), excessive mucus production in the tissue (), and heightened airway resistance to methacholine (). Notably, OVA-mRNA-LNP–treated mice showed no signs of anaphylaxis, either immediately or at any point during the hours or days following OVA-mRNA-LNP injections (). In contrast, in a parallel cohort of mice, subcutaneous administration of OVA () was associated with a high rate of systemic adverse events as manifested by a severe drop in body temperature following the first dose of OVA immunotherapy (, OVA s.c.). Collectively, these findings underscore the potential safety and efficacy of antigen-specific mRNA-LNP to treat allergic inflammation and asthma symptoms. ⁵³ Figure 6A Figure 6B Figure 6B Figure 6, C–F Figure 6G Figure 3G Figure 6C Figure 6D Figure 6E Figure 6F Supplemental Figure 8 ⁵⁴ Supplemental Figure 8 + + + + + –

Both female and male mice aged 6–12 weeks were included in this study, and similar findings are reported for both sexes.

All mouse strains used were on the C57BL/6 genetic background (B6J, 000664) and were purchased from The Jackson Laboratory. OTII mice (004194) were bred with CD45.1 mice (002014). Mice were bred and housed under specific pathogen–free conditions. All experiments complied with protocols approved by the Cincinnati Children's Hospital Medical Center (CCHMC) Animal Use and Care Committee.

The amino acid sequences of the OVA and Der p1 were obtained from GenBank accession numbers MF321513.1 and P08176.2, respectively. The sequences underwent codon optimization and GC enrichment using our proprietary algorithm to improve expression and reduce potential immunogenicity of the in vitro–transcribed mRNA. The codon-optimized allergen sequences were gene-synthesized by GenScript and cloned into an in vitro transcription template containing an optimized T7 promoter, 3′-untranslated region (UTR), 5′-UTR, and 100-adenine tail. The allergen nucleoside–modified mRNAs were prepared using the MegaScript Transcription Kit (Thermo Fisher Scientific), cotranscriptionally capped using the CleanCap system (TriLink Biotechnologies), purified using a modified cellulose base chromatography method (), precipitated, eluted in nuclease-free water, and quantified using the NanoDrop One system (ThermoFisher Scientific). Length and integrity were determined using agarose gel electrophoresis. Endotoxin content was measured using the GenScript Toxisensor chromogenic assay, and values were below detection levels (0.1 EU/mL). mRNA was frozen at –20°C until formulation. ⁶⁵

Cellulose-purified RNAs containing1-methylpseudouridine (m1Ψ) were encapsulated in LNPs using a self-assembly process as previously described wherein an ethanolic lipid mixture of ionizable cationic lipid, phosphatidylcholine, cholesterol, and polyethylene glycol–lipid was rapidly mixed with an aqueous solution containing mRNA at acidic pH (). The LNP formulation used in this study is proprietary to Acuitas Therapeutics; the proprietary lipid and LNP compositions are described in US patent US10,221,127. Additional LNPs were formulated using SM-102, distearoylphosphatidylcholine cholesterol, and 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol-2000 (BroadPharm) in a molar ratio of 50:10:38.5:1.5. Lipids were dissolved in ethanol and combined with aqueous mRNA with a flow rate ratio of 3:1 using a NanoAssemblr Ignite (Cytiva). mRNA-LNPs were diluted 40:1 in PBS containing 10% sucrose, concentrated to 1 mg/mL using centrifugal filters (MilliporeSigma), and stored at –80°C. The hydrodynamic size, polydispersity index, and ζ-potential of mRNA-LNPs were measured using a Zetasizer Nano ZS90 (Malvern Instruments). The mRNA encapsulation efficiency was determined using a modified Quant-iT RiboGreen RNA assay (Invitrogen). Particles had a size of 80 nm, a polydispersity index of 0.02, and an encapsulation efficiency above 95%. N ⁶⁶

Cells were collected from LNs and spleens of OTII mice expressing CD45.1, forced through a 70 μm cell strainer, and treated with ammonium-chloride-potassium (ACK) lysing buffer to remove red blood cells. Naive CD4T cells were isolated using the EasySep Mouse Naive CD4T Cell Isolation Kit. A total of 3 × 10cells per mouse were intravenously injected in 200 μL PBS into sex-matched C57BL/6 recipients, which were immunized 3 days after cell injection. LNs were collected at the specified time points. + + 6

In the preventive protocol, naive wild-type C57BL/6 mice were injected i.m. in the lower left leg with increasing doses (0.1, 0.3, 1, 2, and 5 μg) or a consistent amount of the m1Ψ-mRNA-LNPs encoding an ovalbumin protein (2 μg; OVA-mRNA-LNP), Der p1 protein (5 μg; Der p1–mRNA–LNP), or empty LNP on days 0 and 7, as indicated. Intraperitoneal (i.p.) treatment with everolimus (5 mg/kg of body weight) commenced 2 days before the first immunization and continued daily for 14 days.

In the treatment protocol, mice sensitized by MC903 (1 nmol in 25 μL per ear; Tocris) and OVA (50 μg in 5 μL per ear; MilliporeSigma) were injected with 10 μg of OVA-mRNA-LNP or empty LNP twice, with a 7-day interval between injections. For standard immunotherapy, mice were given 3 subcutaneous (s.c.) injections every third day with 1 mg of endotoxin-free OVA (Ovalbumin EndoFit, InVivoGen) in 100 μL endotoxin-free saline.

A complete list of reagents is provided in. Supplemental Table 8

LNP- or OVA-mRNA-LNP–immunized mice were sensitized through i.p. injection with 100 μg OVA (MilliporeSigma) emulsified in 100 μL of Imject ALUM (Thermo Fisher Scientific) or aluminum hydroxide gel (InVivoGen) in a total volume of 200 μL per mouse on days 24 and 36 after first immunization. Subsequently, on days 48 and 49, mice were challenged i.t., followed by i.n. challenge on days 50 and 51 with 50 μg of OVA. In the chronic model, mice were challenged i.n. with OVA (50 μg in 30 μL) 3 times daily, followed by 6 additional challenges every other day.

In the HDM model, mice were sensitized i.n. with 80 μg of HDM extract (Greer Laboratories Inc.) in 40 μL. Ten days later, mice were challenged i.n. with HDM (10 μg in 30 μL) or np-Der p1 protein (10 μg in 30 μL; InBio) for 4 consecutive days.

In the treatment model, mice were sensitized with MC903 (1 nmol in 25 μL per ear; Tocris) plus OVA (50 μg in 5 μL per ear) applied to both ears daily for 14 days, followed by immunization with OVA-mRNA-LNP (10 μg, i.m.) or LNP 2 weeks later. Two weeks after the last immunization, mice were challenged i.t. with OVA (50 μg in 30 μL) 3 times daily, followed by 5 additional i.n. OVA challenges every other day.

Forty-eight hours after the final OVA or HDM challenge, mice were euthanized with pentobarbital and subsequently examined. Sex- and age-matched naive mice served as untreated controls.

For BALF collection, lungs were lavaged once with 0.8 mL of PBS through cannulation of the tracheal tube. Total cell numbers in the collected fluid were counted using a hemocytometer. BALF was centrifuged, and supernatant was stored at –80°C for enzyme-linked immunosorbent assay (ELISA) or cytokine/chemokine multiplex assays. Cells were resuspended in PBS with 2% fetal bovine serum (FBS) and subjected to live flow cytometry staining on the same day.

One lobe of the lungs was minced and subjected to Liberase TL digestion (0.25 mg/mL) in the presence of DNase I (0.5 mg/mL) at 37°C for 1 hour. The resulting suspension was forced through 70 μm cell strainers, washed, and resuspended in RPMI plus 10% FBS with DNase I (0.5 mg/mL). These cells were used for live or intracellular staining or subjected to scRNA-Seq or quantitative PCR.

Two days after the final OVA challenge, mice were euthanized with pentobarbital for histologic examination. Lungs were inflated through the tracheal tube with 0.7 mL of 10% neutralized buffered formalin, removed, fixed overnight in 10% formalin, and dehydrated in 70% ethanol. Lung tissues were embedded in paraffin and cut into 5-μm-thick sections, which were subsequently deparaffinized. The sections were stained with hematoxylin and eosin (H&E) or periodic acid–Schiff (PAS) or subjected to immunohistochemical stain against murine eosinophil major basic protein (MBP) as previously reported (). ⁶⁷

Mucus-containing goblet cells were detected by PAS staining. PAS-stained goblet cells in the airway epithelium were quantified using a scoring system (0: <5% goblet cells; 1: 5%–25%; 2: 25%–50%; 3: 50%–75%; 4: >75%), as previously described (). Twenty to fifty airways per mouse were examined, and the average score was calculated. Histologic analyses were performed in a blinded manner by the same person. ⁶⁸

The FlexiVent system (SCIREQ Scientific Respirator Equipment Inc.) was used to evaluate airway hyperresponsiveness 48 hours after final OVA exposure. Mice were anesthetized with a mixture of ketamine (90–120 mg/kg) and xylazine (10–20 mg/kg) and paralyzed with pancuronium bromide (0.8–1.2 mg/kg). Tracheae were cannulated with an 18-gauge blunt cannula. Mice were ventilated at 150 breaths/min and 3.0 cm water positive and expiratory pressure and allowed to stabilize on the machine for 2 minutes. Mice were then exposed to methacholine (0, 6.25, 12.5, 25, and 50 mg/mL) aerosolized in PBS for 15 seconds and ventilated for an additional 10 seconds. Ventilation cycle measurements were taken until resistance peaked. Airways were then re-recruited by deep inflation, and the next methacholine dose was administered. Analyses were performed in a blinded manner by the same person.

To monitor anaphylaxis, the anal temperature was measured before the injection of endotoxin-free OVA or OVA-mRNA-LNP and subsequently every 10–15 minutes during the first hour after injection. Thereafter, mice were visually assessed hourly for 6 hours and then twice daily for 3 days.

Freshly obtained cells from BALF and lungs of naive or allergen-sensitized/challenged mice were stained for surface markers, including anti–mouse CD45, CD11c, CD64, CD11b, CD3, CD4, CD8, Gr1, or Ly6G, and Siglec-F, at 4°C for 60 minutes.

For surface and intracellular staining of T cells, cells from the lungs or LNs were incubated in PBS at 4°C for 10 minutes with the fixable viability dye eFluor 780. After washing with PBS containing 2% FBS, cells were fixed and permeabilized using the Foxp3/Transcription Factor Fixation/Permeabilization Kit according to the manufacturer's instructions (Thermo Fisher Scientific). Subsequently, cells were stained with anti-mouse fluorescent antibody cocktails. For lungs, anti–mouse CD44, CD4, CD8, GATA3, Foxp3, KLRG1, CD38, CXCR6, ST2, and CCR5 antibodies were used (). For LNs, anti–mouse CD44, CD4, CD45.1, CD45.2, GATA3, Foxp3, CD25, T-bet, Bcl6, and PD-1 antibodies were used. The staining process occurred at room temperature for 60 minutes. Supplemental Table 8

For intracellular cytokine detection, cells from lungs or LNs were stimulated with phorbol 12,13-dibutyrate (500 ng/mL) and 1 μM ionomycin in the presence of brefeldin A in an incubator at 37°C for 4 hours. Cells were stained in PBS for 10 minutes with the fixable viability dye eFluor 780 in the presence of brefeldin A and then were fixed and permeabilized with the Foxp3/Transcription Factor Fixation/Permeabilization Kit. T cells were stained with fluorescent antibodies against mouse CD4, CD8, CD44, GATA3, Foxp3, IL-2, IL-4, IL-5, IL-13, IL-17A, IFN-γ, TNF-α, CCR5, KLRG1, CD38, and perforin.

All staining was performed in the presence of Fc Block (anti–mouse CD16/CD32, BD Biosciences). A complete list of reagents and antibodies is provided in. Data acquired with a BD LSR Fortessa flow cytometer (BD Biosciences) were analyzed by FlowJo software (Tree Star Inc.). The absolute number of cells in each population in the BALF was calculated by multiplication of the total cell numbers in the collected BALF by the percentage of flow cytometry–gated cells. Supplemental Table 8

Mouse lungs were collected and homogenized in Trizol with QIAGEN beads. mRNA was isolated following the Trizol protocol and purified using the Quick-RNA Miniprep Kit (Zymo Research). For real-time PCR analysis, mRNA was reverse-transcribed with the ProtoScript cDNA Synthesis Kit (New England Biolabs). Primers specific for mouse,,,,,,,,, andwere obtained from Integrated DNA Technologies (). SYBR Green Real-Time PCR was performed and analyzed using the ABI Quant-Studio 7 Flex Real-Time PCR system (Thermo Fisher Scientific). The transcripts of interest were normalized tocDNA. Il4 Il5 Il13 Ifng Cxcl9 Cxcl10 Ccl5 Ccl11 Ccl24 Eif3k Eif3k Supplemental Table 9

BALF was collected 2 days after the last challenge and analyzed with the Mouse Cytokine/32-Plex Discovery Assay (Eve Technologies Corp.). The Mouse CCL24/Eotaxin-2/MPIF-2 DuoSet ELISA Kit (R&D Systems) was used to measure the concentration of eotaxin-2 in the BALF.

To measure OVA-specific antibody secretion, blood was collected from the tail 2 weeks after the last dose of OVA-mRNA-LNP or Der p1–mRNA–LNP immunization, 7 days after the last OVA+Alum sensitization, or 1 day after OVA+MC903 sensitization. Serum was separated by centrifugation at 2,236for 5 minutes at room temperature. Ninety-six-well ELISA plates were coated with 10 μg/mL of OVA overnight at 4°C. Serum samples were added in different dilutions. Detection of OVA-specific antibodies from serum was performed using horseradish peroxidase–conjugated (HRP-conjugated) anti-IgG1, anti-IgG2a, or biotinylated anti-IgG2b followed by HRP according to the manufacturer's instructions. For substrate, 3,3′,5,5′-tetramethylbenzidine (TMB) was used. The colorimetric reaction was stopped with 10% HPO, and the optical density was quantified using an ELISA plate reader at 450 nm, with subtraction of background absorbance at 570 nm. OVA-specific monoclonal antibodies IgG1, IgG2a, and IgG2b were used as standards (). To prevent nonspecific binding, SuperBlock Blocking Buffer (Thermo Fisher Scientific) was used for blocking and dilutions. For the detection of OVA-specific IgE in murine serum, an ELISA kit from BioLegend was used. For the detection of Der p1–specific IgG, an ELISA kit from Chondrex was used. g 3 4 Supplemental Table 8

Single-cell transcriptome analysis of the murine esophagus was performed using the BD Rhapsody Single-Cell Analysis System (BD Biosciences). To obtain a single-cell suspension, the left lobe of the lungs was minced and incubated with Liberase TL (0.25 mg/mL) in the presence of DNase I (0.5 mg/mL) for enzymatic digestion. Cells from 2 mice per group were pooled, treated with ACK lysing buffer for 1 minute, washed, passed through a 70 μm cell strainer, and counted using the TC20 Automated Cell Counter (Bio-Rad). Subsequently, cells were processed on the Rhapsody platform following the manufacturer's protocol. Cells from each treatment group were labeled with individual sample tags (BD Mouse Single-Cell Multiplexing Kit, catalog 633793). Before loading into the cartridges, cells were quantified on the Rhapsody scanner after staining with Vybrant DyeCycle Green (Invitrogen, V35004) for 5 minutes at room temperature. Barcoded samples were pooled in equal amounts for each of the 3 treatment groups (naive, LNP, OVA-mRNA-LNP), and 60,000 cells were loaded into each cartridge, with 2 cartridges used for sequencing. Single cells were isolated with the BD Rhapsody Express Single-Cell Analysis System according to the manufacturer's recommendations; after the final bead wash step, 40,000 cells were detected as singlets.

Libraries were prepared following the BD Rhapsody System mRNA Whole Transcriptome Analysis (WTA) and Sample Tag Library Preparation Protocol (BD Biosciences, 633801) and sequenced at the CCHMC Genomics Sequencing Facility on the NovaSeq 6000 using the PE (paired end) 100 sequencing format. After sequencing and preprocessing by the Rhapsody WTA Analysis Pipeline using exact cell count, 28,000 cells per cartridge were used for the downstream analysis.

Raw scRNA-Seq data from 3 experimental conditions (naive, LNP, OVA-mRNA-LNP) were analyzed with v1.12.1 of the BD Rhapsody WTA Analysis Pipeline (BD Biosciences) on the Seven Bridges Platform (). Sequencing reads in the FASTQ files were aligned to the GRCm38.p6 reference genome (GENCODE, mouse release M19). Cell calling was performed with the Refined Putative Cell Calling parameter disabled. Instead, the Exact Cell Count input was set to 28,000 cells. Generated feature-barcode matrices from 2 batches were merged, preserving the experimental condition and the batch information of each cell within its barcode. The merged feature-barcode matrix was then uploaded to SciDAP () for all subsequent data analysis steps. https://www.sevenbridges.com/platform↗ https://scidap.com↗

Low-quality cells were removed with the scRNA-Seq Filtering Analysis pipeline () using the following quality control thresholds: a minimum of 500 transcripts per cell, 500–5,000 genes per cell, and maximum 5% of transcripts mapped to mitochondrial genes. Doublets were identified and removed with the scDblFinder R package () as part of this pipeline. https://github.com/Barski-lab/workflows-datirium/blob/b210f5a5d3b303e3627582963a0b33b289f9a40e/workflows/sc-rna-filter.cwl↗ ⁶⁹

The remaining high-quality cells were processed by the scRNA-Seq Dimensionality Reduction Analysis pipeline (). Molecular count data were first corrected for technical variability and then integrated using the pairs of cells sharing a matched biological state as integration anchors. Other configuration parameters included (a) setting the normalization method to "sctglm," which resulted in using the glmGamPoi R package within the SCTransform function (); (b) limiting the number of highly variable genes for normalization and integration to 3,000; and (c) selecting the first 50 principal components (PCs) for principal component analysis and uniform manifold approximation and projection (UMAP) dimensionality reduction algorithms. https://github.com/Barski-lab/workflows-datirium/blob/b210f5a5d3b303e3627582963a0b33b289f9a40e/workflows/sc-rna-reduce.cwl↗ ⁷⁰

Dimensionally reduced single-cell data were clustered with the scRNA-Seq Cluster Analysis pipeline () with 50 PCs and 1.0 clustering resolution. For each of the 27 obtained clusters, gene markers were identified using the FindAllMarkers function with default parameters (). The cluster with damaged cells (low transcripts-per-cell counts and nonspecific gene markers) was removed, and both scRNA-Seq Dimensionality Reduction Analysis and scRNA-Seq Cluster Analysis pipelines were rerun with 50 PCs and 1.0 clustering resolution. The resulting clusters produced cell types that were used in the following steps of the analyses. https://github.com/Barski-lab/workflows-datirium/blob/b210f5a5d3b303e3627582963a0b33b289f9a40e/workflows/sc-rna-cluster.cwl↗ https://satijalab.org/seurat/reference/findallmarkers↗

Cells belonging to the T cell cluster were separated and reclustered by application of scRNA-Seq Dimensionality Reduction Analysis and scRNA-Seq Cluster Analysis pipelines. Dimensionality was reduced to 20 PCs, and the number of highly variable genes was limited to 1,000. The clustering resolution was set to 0.5.

Differentially expressed genes between experimental conditions were identified with the scRNA-Seq Differential Expression Analysis pipeline () run on the selected subsets of cells using the DESeq2 package (). Gene expression data were aggregated to the pseudobulk level per sample. https://github.com/Barski-lab/workflows-datirium/blob/b210f5a5d3b303e3627582963a0b33b289f9a40e/workflows/sc-rna-de-pseudobulk.cwl↗ ⁷¹

Module scores for activated CD8T cells were calculated at the single-cell level, as previously described (), using 197 genes identified in human CD8T cells following SARS-CoV-2 mRNA immunization () (). + + ⁴⁸ Supplemental Table 7 ⁴⁹

The results are presented as mean ± SEM or displayed with box-and-whisker plots. Normality tests were applied to validate Gaussian distribution. Multiple-group comparisons were conducted using 1-way or 2-way analysis of variance (ANOVA) with Tukey correction applied. All statistical analyses, except scRNA-Seq analyses, were performed with Prism 10 software (GraphPad Software Inc.).less than 0.05 was considered statistically significant (–). P ^{65 68}

All experiments complied with protocols approved by the Cincinnati Children's Hospital Medical Center (CCHMC) Animal Use and Care Committee.

All data are available in the main text or the supplemental materials. Individual data point values can be found in thedocument. The single-cell RNA sequencing data were deposited at the NCBI's Gene Expression Omnibus (GEO) repository under the accession number GSE259439. Supporting Data Values