Mucosal Immunization with a pH-Responsive Nanoparticle Vaccine Induces Protective CD8⁺ Lung-Resident Memory T Cells

Abstract

Tissue-resident memory T cells (T_RM) patrol non-lymphoid organs and provide superior protection against pathogens that commonly infect mucosal and barrier tissues, such as the lungs, intestine, liver, and skin. Thus, there is a need for vaccine technologies that can induce a robust, protective T_RM response in these tissues. Nanoparticle (NP) vaccines offer important advantages over conventional vaccines; however, there has been minimal investigation into the design of NP-based vaccines for eliciting T_RM responses. Here, we describe a pH-responsive polymeric nanoparticle vaccine for generating antigen-specific CD8⁺ T_RM cells in the lungs. With a single intranasal dose, the NP vaccine elicited airway- and lung-resident CD8⁺ T_RM cells and protected against respiratory virus challenge in both sublethal (vaccinia) and lethal (influenza) infection models for up to nine weeks after immunization. In elucidating the contribution of material properties to the resulting T_RM response, we found that the pH-responsive activity of the carrier was important, as a structurally analogous non-pH-responsive control carrier elicited significantly fewer lung-resident CD8⁺ T cells. We also demonstrated that dual-delivery of protein antigen and nucleic acid adjuvant on the same NP substantially enhanced the magnitude, functionality, and longevity of the antigen-specific CD8⁺ T_RM response in the lungs. Compared to administration of soluble antigen and adjuvant, the NP also mediated retention of vaccine cargo in pulmonary antigen-presenting cells (APCs), enhanced APC activation, and increased production of T_RM-related cytokines. Overall, these data suggest a promising vaccine platform technology for rapid generation of protective CD8⁺ T_RM cells in the lungs.

Graphical Abstrct

Tissue-resident memory T cells (T_RM) are a specialized subset of memory cells with a distinct phenotype that reside in non-lymphoid tissues and act as a first line of defense against many pathogens.^1–3 T_RM cells remain localized in their home tissues due to a combination of adhesion molecules and a lack of homing mechanisms for trafficking to distal lymphoid organs or circulating in the blood. As such, they are optimally positioned to respond more quickly than peripheral memory T cells. T_RM have been identified in several organs, including the skin, liver, kidneys, brain, and mucosal tissues such as the lungs, intestine, and female reproductive tract.^{4, 5} These cells play a key role in protection against multiple infectious diseases for which new or improved vaccines are needed, such as influenza, tuberculosis, respiratory syncytial virus, and HIV/AIDS.^6–9 T_RM are also important in immunity against tumors, including breast cancer, melanoma, and lung cancer.^10–13 For several diseases, antibodies or non-resident memory T cells alone are not sufficient for optimal protection;^14–17 however, very few currently approved vaccines have been shown to generate T_RM cells.^{18, 19} Therefore, there is a critical need to develop new vaccines that can induce protective T_RM in target tissues.

Pulmonary T_RM reside in both the lung interstitium and airways^{20, 21} and are critical in mediating protection against respiratory pathogens.^{18, 19, 22} Mucosal vaccination has garnered attention as a superior route of immunization over traditional intramuscular injections for several reasons, including its ability to mimic routes of pathogen entry and generate tissue-specific immune cells optimally positioned to fight off future infection.^{13, 23} Specifically, pulmonary immunization via intranasal (i.n.) administration is advantageous for generating T_RM in the lungs.^{24, 25} Additionally, there is evidence that pulmonary immunization can generate T cell responses in distal mucosal tissues.²⁶ Hence, the development of vaccine formulations that can be administered by mucosal routes holds great promise for a new generation of T_RM-targeted vaccines.

Protein-based subunit vaccines have been widely studied as a next-generation vaccine platform, including in the context of mucosal delivery.²⁷ A major drawback of protein-based subunit vaccines, however, is poor immunogenicity due to several drug delivery barriers, including rapid antigen clearance with poor uptake by dendritic cells and minimal accumulation in draining lymph nodes. Subunit vaccines are particularly inept at eliciting CD8⁺ T cells, which are required for immunity against many pathogens and cancers.^{28, 29} Eliciting a robust CD8⁺ T cell response requires antigen presentation on MHC-I by dendritic cells (DCs) in the context of additional molecular cues (costimulation, cytokines) that drive CD8⁺ T cell expansion and differentiation.^{28, 30} To achieve presentation by MHC-I, administered antigen must either be endocytosed by specialized cross-presenting DC subsets or delivered to the classical cytosolic MHC-I antigen processing pathway. However, the predominant fate of soluble endocytosed antigen is lysosomal degradation, with minimal presentation on MHC-I.^{31, 32}

Despite their limited capacity to generate CD8⁺ T cells, the superior safety profile of subunit vaccines has motivated strategies to improve their efficacy.³³ Toward this end, a variety of nanoparticle (NP)-based vaccine delivery systems have been developed that utilize material properties to enhance antigen uptake by DCs, promote antigen cross-presentation, and/or co-deliver immunostimulatory adjuvants in order to potentiate CD8⁺ T cell responses to immunization.^34–39 This includes NP formulations that have been administered i.n. to generate pulmonary T cell responses in mouse models of infection and cancer.^40–42 However, to date only a few reports have evaluated the ability of NP-based subunit vaccines to specifically induce CD8⁺ T_RM cells in the lungs.^{26, 43, 44} Moreover, while NP design principles for eliciting robust systemic T cell responses have largely been established, the ways in which properties of NP vaccines can be engineered to augment T_RM responses elicited by mucosal immunization have not been explored. This motivates the need for the design, optimization, and evaluation of NP vaccines for installing this unique memory T cell population in the lungs and other mucosal tissues.

While elucidation of the mechanisms underlying induction and maintenance of CD8⁺ T_RM in the lungs remains an active area of investigation, lessons in vaccine design can be taken from studies of respiratory viral infections like influenza, in which robust and durable T_RM are often generated.^{17, 45} These studies motivate the design of NP vaccines that can mimic viral infection by enhancing antigen uptake and cross-presentation in APCs, allowing for co-delivery of antigen and adjuvant, and/or increasing local antigen persistence in tissues.^46–49 Therefore, in this study we leveraged a viral-mimetic polymeric NP vaccine delivery system that utilizes a pH-dependent endosomal escape mechanism to release cargo into the cytosol, resulting in enhanced antigen delivery to the MHC-I processing pathway.³⁷ Additionally, the corona of the NP is designed to enable dual-delivery of antigen and nucleic acid adjuvant from the same particle, further augmenting its ability to mimic pathogen encounter and enhance the CD8⁺ T cell response.

Here, we demonstrate that a pH-responsive NP vaccine dual-loaded with ovalbumin protein antigen (OVA) and CpG DNA adjuvant enhanced the magnitude, functionality, and longevity of the pulmonary CD8⁺ T_RM response in mice. It also improved activation of pulmonary APCs and promoted antigen persistence in the lungs. Importantly, a single i.n. dose of the NP vaccine conferred protection against respiratory challenge with a recombinant vaccinia virus containing the CD8⁺ T cell epitope of OVA. Of significance to the design of T_RM-targeted NP vaccines, we also show that the pH-responsive activity of the NP is important for induction of lung-resident CD8⁺ T cells, demonstrating that a material property can be harnessed to install this important cell population in a mucosal tissue. Finally, we show that a clinically relevant antigen, influenza A virus nucleoprotein, can be delivered via the NP vaccine, generate a CD8⁺ T_RM response in the lungs, and confer protection in a lethal respiratory influenza virus challenge model after a single dose. Collectively, our results suggest a promising experimental vaccine platform for generating CD8⁺ T_RM against respiratory infections and offer evidence that NP properties can be modulated to augment T_RM responses elicited by pulmonary immunization.

RESULTS AND DISCUSSION

Intranasal Antigen Delivery with a pH-Responsive Nanoparticle Carrier Enhances the Lung-Resident CD8⁺ T Cell Response

The nanoparticle described in this report is formulated using a pH-responsive diblock copolymer designed to enhance cytosolic delivery of vaccine cargo and strengthen the CD8⁺ T cell response by promoting processing and presentation of antigen in the major histocompatibility complex class I (MHC-I) pathway.³⁷ The polymer is composed of two functional blocks synthesized by RAFT polymerization (Scheme S1A–S1B). The first block is a hydrophilic copolymer of dimethylaminoethyl methacrylate (DMAEMA) and pyridyl disulfide ethyl methacrylate (PDSMA). PDSMA provides pyridyl disulfide groups on the surface of the particle for conjugation to thiolated protein antigen, and DMAEMA contributes cationic charge for electrostatic complexation with a nucleic acid adjuvant. The second block is a pH-responsive, endosomolytic copolymer of propylacrylic acid (PAA), butyl methacrylate (BMA), and DMAEMA, which drives micellar nanoparticle assembly due to its hydrophobic nature. After cellular uptake and in response to endosomal acidification, the micellar structure of the NP transforms to expose the membrane-destabilizing core (PAA-BMA-DMAEMA), which promotes endosomal escape and cytosolic antigen delivery to the MHC-I pathway.

While this mechanism of NP-mediated antigen cross-presentation has been utilized to enhance the splenic CD8⁺ T cell response via subcutaneous administration, whether pH-responsive activity can be leveraged to generate a lung-resident CD8⁺ T cell response after intranasal antigen delivery is unknown. To evaluate this, both pH-responsive NP and non-responsive control NP (second block consisting only of BMA) were synthesized and characterized (Scheme S1A–1C, Table S1, Figure S1), and a model protein antigen, ovalbumin (OVA), was thiolated and covalently conjugated to both NP carriers (Figures S2A and S2C). Mice were immunized i.n. with antigen-NP conjugates made using either pH-responsive nanoparticles (OVA-NP_pH) or control nanoparticles (OVA-NP_ctrl). On d13 after immunization, mice were injected intravenously (i.v.) with αCD45.2 antibody to distinguish CD45⁺ blood-borne cells in the lung vasculature (“marginated vascular,” MV) from tissue-resident CD45⁻ cells in the lung interstitium (IST). By directly discriminating between cells residing in distinct lung compartments, this intravascular staining technique provides an accurate and robust method for establishing tissue residence.^{9, 50–52} Broncheoalveolar lavage fluid (BAL) was also collected to differentiate CD45⁻ cells resident in the airways (AW) from CD45⁻ IST-resident cells⁵² (Figure 1A).

Figure 1. Intravascular staining is used to determine localization of CD8⁺ T cells after intranasal delivery to the lungs.

(A) Schematic of the experimental timeline. (B) Flow cytometry was used to identify antigen-specific (Tet⁺) CD8⁺ T cells in distinct lung compartments (airway, AW; interstitium, IST; marginated vascular, MV) and the spleen. BAL was collected to discriminate AW vs. IST cells, and i.v. staining with αCD45 antibody discriminated IST (CD45⁻) vs. MV (CD45⁺) cells. Samples were stained with PE-labeled SIINFEKL/MHC-I tetramer to identify antigen-specific CD8⁺ T cells. After gating out CD11b⁺, CD11c⁺, B220⁺, and CD4⁺ cells (“dump”), CD8α⁺CD45⁻Tet⁺ events in AW and IST, CD8α⁺CD45⁺Tet⁺ events in MV, and CD8α⁺Tet⁺ events in the spleen were quantified. Dot plots are representative of the gating strategy used in multiple experiments (see Figures S3A–S3C). (C) In conjunction with i.v. staining, microscopy was used to visualize fluorescent OVA-NP conjugates in the lower airways 24 h after immunization. Lungs were stained with αCD45 antibody, which labels vascular leukocytes, and tomato lectin, which binds to capillary endothelial cells and allows for visualization of lung structure. Purple: OVA-NP; blue: vascular leukocytes; green: lung vasculature. Scale bar = 100 μm. Immunization dose: 25 μg NP, 7.5 μg OVA.

Cells obtained from lungs, BAL, and spleens were stained with fluorescent antibody against a panel of surface markers and with fluorescent MHC-I tetramer (Tet) containing SIINFEKL peptide (the immunodominant H-2K^b epitope for OVA). They were then analyzed by flow cytometry to quantify antigen-specific CD8⁺ T cells (Tet⁺CD8α⁺CD11b⁻CD11c⁻B220⁻CD4⁻) in each lung compartment and the spleen (Figures 1B and S3A–S3C). Fluorescence microscopy was also used to confirm that OVA-NP conjugates reached the lower airways after i.n. administration. Conjugates formulated with Alexa Fluor 647-labeled OVA were visible in lungs harvested 24 h after immunization (Figure 1C). We found that the pH-responsive carrier elicited a significantly greater antigen-specific (Tet⁺) CD8⁺ T cell response than the control carrier in the AW (Figure 2A) and IST (Figure 2B), while there was no significant difference in the response between carriers in the MV (Figure 2C) and spleen (Figure 2D). These data indicate that the pH-responsive property of the NP is important for generating a tissue-resident CD8⁺ T cell response in the lungs, and serve to demonstrate the importance of NP properties in development of T_RM-targeted vaccines.

Figure 2. Intranasal antigen delivery with pH-responsive nanoparticle enhances lung-resident CD8⁺ T cell response.

Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (A) AW, (B) IST, (C) MV, and (D) spleen were enumerated on d13 after i.n. administration of OVA-NP_pH or OVA-NP_ctrl. Representative dot plots are gated on viable CD8⁺ T cells. Immunization dose: 25 μg NP, 7.5 μg OVA. Data are mean ± SEM and representative of two independent experiments, with n = 6 per group. Limits of detection: 1 cell (AW), 5 cells (IST/MV), 25 cells (spleen). *p<0.05, **p<0.01 by unpaired t-test. ns, not significant. See also Figures S3A–S3C.

Intranasal Dual-Delivery of Antigen and Adjuvant with pH-Responsive Nanoparticles Enhances Magnitude and Functionality of the Lung-Resident CD8⁺ T Cell Response

After demonstrating the importance of pH-responsiveness in generating a robust lung-resident CD8⁺ T cell response, we next asked whether dual-delivery of antigen and an immunostimulatory adjuvant on the same pH-responsive NP would further augment this response. While this pathogen-mimetic property has previously been shown to enhance the systemic CD8⁺ T cell response,³⁷ to our knowledge the importance of NP dual-delivery for generating tissue-resident CD8⁺ T cells has not been clearly demonstrated. Therefore, NP were co-loaded with OVA protein and a nucleic acid adjuvant, CpG ODN 1826 (a single-stranded DNA agonist for TLR9). CpG DNA has been shown to enhance the CD8⁺ T cell response and has precedence for use in pulmonary immunization.^{40, 44, 53} CpG was electrostatically complexed to OVA-NP conjugates (Figure S2B), and this formulation is referred to henceforth as OVA-NP/CpG or the “nanoparticle vaccine.”

On d0, mice were immunized i.n. with OVA-NP/CpG, OVA-NP conjugate, a mixture of soluble OVA+CpG, a mixture of soluble OVA+NP, or NP/CpG complex mixed with soluble OVA (NP/CpG+OVA) (Figure 3A). On d13 after immunization, mice were injected i.v. with αCD45.2 antibody and lungs, BAL, and spleens were collected and analyzed by flow cytometry as described above (Figures 1A–1B and S3A–S3C). Mice immunized with a single dose of OVA-NP/CpG produced significantly more Tet⁺ CD8⁺ T cells relative to all other formulations (Figures 3B–3E). This increased response was observed in both the AW (Figure 3B) and IST (Figure 3C) lung compartments, as well as in the MV (Figure 3D) and spleen (Figure 3E). These data demonstrate that NP-mediated dual-delivery of antigen and adjuvant to the lungs enhances the CD8⁺ T cell response over immunization with OVA-NP. The effect was particularly prominent in the IST, indicating the ability of the NP vaccine to induce lung-resident CD8⁺ T cells. Simple mixing of components (OVA+CpG, OVA+NP) was ineffective, and dual-delivery on the same particle was crucial, as the formulation containing all three components without co-loading (NP/CpG+OVA) did not induce a robust response.

Figure 3. Intranasal dual-delivery of antigen and adjuvant via nanoparticle vaccine enhances magnitude and functionality of lung-resident CD8⁺ T cell response.

(A) Mice were immunized on d0 with the NP vaccine or control formulations, and lungs, spleens, and/or BAL were collected on d13 for analysis of the immune response by tetramer staining or ICCS. (B-E) Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (B) AW, (C) IST, (D) MV, and (E) spleen were enumerated on d13 after i.n. administration of OVA-NP/CpG or control formulations. (F) ICCS was used to identify % CD8⁺ T cells positive for IFNγ and/or TNFα after ex vivo restimulation of lungs and spleen with SIINFEKL peptide. Statistical differences are shown for IFNγ⁺TNFα⁺ group only. Data are mean ± SEM and representative of one to three independent experiments, with (B-E) n = 4–7 per group and (F) n = 2–4 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg CpG. Limits of detection for (B-E): 1 cell (AW), 5 cells (IST/MV), 25 cells (spleen). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by (B-E) ordinary one-way ANOVA with Tukey multiple comparisons test or (F) ordinary two-way ANOVA with Tukey multiple comparisons test. See also Figures S3A–S3D.

T cell functionality after immunization was assessed via intracellular cytokine staining (ICCS). Mice were immunized on d0 with OVA-NP/CpG, OVA-NP, OVA+CpG, or PBS, and lungs and spleens were collected on d13. Lung and spleen cells were re-stimulated with SIINFEKL peptide and analyzed by flow cytometry for production of IFNγ and TNFα (Figure S3D). Mice immunized with OVA-NP/CpG had a greater percentage of polyfunctional (IFNγ⁺TNFα⁺) antigen-specific CD8⁺ T cells in both the lungs and spleen, relative to all other formulations (Figure 3F). This further supports the importance of antigen and adjuvant dual-delivery in generating a robust and functional CD8⁺ T cell response, particularly in the lungs.

In addition, the CD8⁺ T cell response in mice immunized i.n. with OVA-NP/CpG was compared to the response to subcutaneous (s.c.) immunization with the same formulation. It has been reported that systemic administration of antigenic protein and adjuvant via the intraperitoneal (i.p.) route generates low numbers of lung-resident CD8⁺ T cells relative to i.n. administration.⁵² We hypothesized that s.c. immunization with the NP vaccine would also be ineffective. On day 13 post-immunization, tetramer staining was used to analyze the antigen-specific CD8⁺ T cell response to i.n. or s.c. administration of the NP vaccine. Responses in the AW (Figure 4A) and IST (Figure 4B) were significantly higher for mice immunized i.n.; however, there was no difference between administration routes for MV (Figure 4C) or spleen (Figure 4D). Consistent with previous reports, this demonstrates the importance of i.n. administration of the NP vaccine for generating a lung-resident CD8⁺ T cell response.^{26, 42, 43}

Figure 4. Pulmonary immunization via intranasal administration is optimal for generating a lung-resident CD8⁺ T cell response.

Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (A) AW, (B) IST, (C) MV, and (D) spleen were enumerated on d13 after i.n. or s.c. administration of OVA-NP/CpG. Data are mean ± SEM, with n = 4–5 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg CpG. Limits of detection: 1 cell (AW), 5 cells (IST/MV), 25 cells (spleen). **p<0.01, ***p<0.001, ****p<0.0001 by unpaired t-test.

Pulmonary toxicity is an important consideration when evaluating the translational potential of mucosal vaccines. NP-based i.n. or intratracheal vaccine and/or CpG administration have previously been reported to be safe in mice.^{26, 40, 41} In this work, mice immunized i.n. exhibited minimal weight loss within the first two days and recovered rapidly (Figure S4A). In addition, lung tissue harvested at d1 and d12 after immunization with either OVA-NP/CpG or OVA+CpG was evaluated for immunopathology. Mild inflammation was induced by both formulations, with no signs of pathology or tissue damage, and findings were consistent between animals in the same treatment group (Figure S4B). This mild inflammation was associated with infiltration of immune cells to the lungs (lymphocytes, macrophages, neutrophils at d1; lymphocytes, plasma cells, macrophages at d12) and is consistent with flow cytometry data showing recruitment of lymphoid cells in response to i.n. vaccine administration (Figures 3B–3E).

Taken together, these data demonstrate that i.n. administration of the NP vaccine is safe and can significantly enhance the magnitude and functionality of the antigen-specific CD8⁺ T cell response within 13 days and after a single dose. In addition to local lung-resident CD8⁺ T cells, the NP vaccine can also induce systemic immunity via i.n. administration. Importantly, pathogen-mimetic dual-delivery of antigen and adjuvant on the same particle is integral to the magnitude and functionality of this response.

Nanoparticle-Mediated Dual-Delivery Enhances Persistence and Co-Localization of Cargo and Expression of Activation Markers in Pulmonary Antigen-Presenting Cells

We next asked what characteristics of the NP vaccine formulation could account for the enhanced lung-resident CD8⁺ T cell response. Previous reports have demonstrated the importance of antigen persistence in establishing T_RM populations in the lungs and other tissues.^{2, 47, 48, 54} In addition, cross-presenting CD103⁺ DCs are involved in activation of precursor T_RM, and alveolar macrophages can promote formation of T_RM in the lungs.^55–57 We hypothesized that the NP formulation would increase vaccine residence time in the lungs, thereby promoting extended co-delivery of vaccine cargo (OVA and CpG) to pulmonary innate immune cells. We also postulated that this NP-mediated persistence would prolong activation of local APCs, thus leading to an improved downstream adaptive immune response and formation of T_RM in the lungs.⁴⁸

To evaluate this, we labeled OVA with Alexa Fluor 647 (OVA₆₄₇) and CpG with Alexa Fluor 488 (CpG₄₈₈) and immunized mice i.n. with fluorescent formulations (OVA₆₄₇-NP/CpG₄₈₈ or OVA₆₄₇+CpG₄₈₈) or PBS. First, to assess organ-level local and systemic biodistribution of the formulations, lungs and spleens were harvested at 24, 48, and 72 h post-immunization and imaged to determine whether the NP vaccine enhanced OVA retention relative to the soluble formulation. Quantification of OVA₆₄₇ average radiant efficiency in fluorescent images demonstrated that antigen remained in the lungs longer for mice immunized with the NP vaccine. At both 48 and 72 h post-immunization, mice receiving OVA₆₄₇-NP/CpG₄₈₈ had significantly more OVA₆₄₇ fluorescence in the lungs relative to OVA₆₄₇+CpG₄₈₈ (Figure 5A). In addition, there was negligible fluorescence present in spleens at all time points, suggesting that i.n. administration leads to localized pulmonary delivery with minimal systemic distribution.

Figure 5. Nanoparticle-mediated dual-delivery enhances co-localization and retention of cargo in pulmonary APCs and expression of activation markers.

(A) Lungs and spleens were imaged at 24, 48, and 72 h post-immunization to quantify uptake of OVA₆₄₇. Representative images of lungs and spleens from each treatment group at each timepoint (left) and quantification of OVA₆₄₇ fluorescence in lungs and spleens over time (right). (B) Flow cytometry was used to quantify the # of pulmonary APCs positive for both OVA₆₄₇ and CpG₄₈₈ (OVA⁺CpG⁺) at each time point (“cell count”). Cell counts were also multiplied by either OVA MFI or CpG MFI to determine “relative uptake” of each cargo (MFI × cell count). Bar graphs for CD103⁺ DC and CD11b⁺ DC are replicated for visibility. (C) Expression of CD86 was measured for several cell subsets in the lungs. Data are mean ± SEM and representative of four independent experiments, with n = 3–4 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg CpG. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by (A) unpaired t-test at each timepoint comparing OVA-NP/CpG vs. OVA+CpG, with Holm-Sidak multiple comparisons test, (B) ordinary two-way ANOVA with Sidak’s multiple comparisons test, or (C) ordinary two-way ANOVA with Tukey multiple comparisons test (significance shown for OVA-NP/CpG vs. OVA+CpG). See also Figure S5.

In addition to assessment of localization and retention at the organ level, lungs were also analyzed by flow cytometry to evaluate uptake of vaccine cargo in pulmonary APC subsets (Figure S5A), as well as expression of the costimulatory marker CD86 in these populations. The cell types analyzed were: (1) alveolar macrophages (AMφ), (2) interstitial macrophages (IMφ), (3) CD103⁺ dendritic cells (CD103⁺ DC), (4) CD11b⁺ dendritic cells (CD11b⁺ DC), (5) a population of monocyte- and macrophage-like cells not encompassed by other subsets (Mono/Mφ), (6) granulocytes (Gran), and (7) “Other” (anything not included in previous categories).⁵⁸ We expected dual-delivery of OVA₆₄₇ and CpG₄₈₈ with the NP vaccine would increase their co-localization within cells, so we analyzed cells that were double-positive for both cargoes (OVA⁺CpG⁺) (Figure S5B). Uptake was quantified for each cell subset as both “cell count” (# OVA⁺CpG⁺ events) and “relative uptake” (mean fluorescence intensity (MFI) of either OVA or CpG multiplied by # OVA⁺CpG⁺ events). Relative uptake was used to account for both the number of cells containing cargo (#) and the total amount of cargo taken up (MFI).²⁶ Initially, after 24 h, there was no difference in uptake between OVA-NP/CpG and OVA+CpG (Figures 5B and S5C, top row). Starting at 48 h and increasing by 72 h post-immunization, there was significantly more cargo co-localization from the NP vaccine in several cell types, including CD103⁺ DCs and CD11b⁺ DCs at 48 h (Figures 5B and S5C, middle row), and DCs, AMφ, IMφ, Mono/Mφ, and Gran at 72 h (Figures 5B and S5C, bottom row). These data reflect the results obtained from fluorescent organ imaging (Figure 5A) and demonstrate that NP delivery prolongs antigen and adjuvant co-localization and retention in pulmonary APCs, suggesting that an NP-mediated increase in local antigen persistence may contribute to the generation of lung T_RM.

In addition, CD86 expression was significantly upregulated in certain cell subsets after immunization with OVA-NP/CpG relative to OVA+CpG, including CD103⁺ DCs, IMφ, and Mono/Mφ (Figure 5C). The increase in expression was minimal in AMφ, CD11b⁺ DCs, and Gran (Figure S5D). This is particularly notable for CD103⁺ DCs, since this cell subset exhibited lower levels of cargo uptake compared to, e.g., Mφ populations, but the subset of CD103⁺ DCs that did internalize the formulation appears to have been strongly activated, as CD86 expression remained high even after 72 h. Importantly, this cell subset has been implicated in the development of T_RM in the lungs.⁵⁹ Taken together, these data demonstrate that the NP vaccine enhances persistence and co-localization of vaccine cargo in pulmonary innate immune cells, as well as activation of APCs that can promote a T_RM response.

Acute Cytokine Response to Nanoparticle Vaccine Is Localized to the Lungs and Supports Generation of Lung-Resident CD8⁺ T Cells

The innate immune response generated by a vaccine plays a critical role in shaping the magnitude and phenotype of the resulting adaptive immune response, and this innate response can be characterized in terms of the cytokine profile induced by the vaccine. Evidence suggests that maturation into T_RM occurs after activated T cells have migrated back to their home tissue from the lymph nodes, and that the presence of local inflammatory signals in the tissue drives this process.^{2, 60–62} A number of cytokines are important for generating CD8⁺ T cell responses (IFNγ, type I interferons (IFNα/β), IL-12, IL-1, IL-6)^63–67 and for the induction and maintenance of T_RM (IL-33, IFNα/β, TNFα, IL-12, TGFβ, IL-7, IL-15).^{2, 54, 60, 62, 68} In particular, several of these cytokines are involved in upregulating T_RM surface markers CD69 and CD103; type I IFN, IL-33, and TNFα induce CD69 upregulation on T cells, TGFβ has been shown to induce CD103 expression on T_RM precursors, and IL-12 may play a role in differentiation of CD103⁻CD69⁺ T_RM. Many of these cytokines are also important for rapid activation of pulmonary APC subsets that that have been implicated in T_RM generation.^{69, 70}

We hypothesized that the NP vaccine would increase local production of key cytokines in the lungs, leading to an improved pulmonary CD8⁺ T_RM response. To evaluate this, mice were immunized i.n. with OVA-NP/CpG, OVA+CpG, or PBS, and multiplexed cytokine analysis was used to quantify cytokine levels in lung homogenate, BAL, and serum at 6 h, 24 h, 48 h, and 7 d post-immunization. Overall, i.n. administration of the NP vaccine generated 2.5- to 375-fold higher concentrations of cytokines in the lungs and BAL relative to concentrations in the serum, indicating local cytokine production with minimal systemic inflammation (Figure 6). The response peaked at 24 h or 48 h before returning to baseline by d7 post-immunization. The local and acute nature of this cytokine response corroborates histological analyses demonstrating a favorable safety profile for the NP vaccine (Figure S4B). At 24 h and/or 48 h post-immunization, OVA-NP/CpG generated 2- to 10-fold higher levels of cytokines associated with CD8⁺ T cells (IFNγ, IFNα/β, IL-12p70, IL-1β, IL-6) in the lungs and/or BAL, relative to OVA+CpG. Similarly, concentration of cytokines related to T_RM generation (IL-33, TNFα, IFNα/β, IL-12p70) were 2- to 15-fold higher in the lungs and/or BAL at 24 h and/or 48 h after immunization with OVA-NP/CpG, relative to OVA+CpG. We also compared the cytokine response generated by OVA-NP vs. OVA-NP/CpG. Overall, the OVA-NP formulation generated little to no response above baseline levels, whereas for the majority of cytokines tested, OVA-NP/CpG stimulated significantly higher cytokine concentrations than OVA-NP at the 48 h timepoint (Figure S6). Thus, CpG appears to be an integral component of stimulating the cytokine response profile observed after immunization with OVA-NP/CpG. Taken together, these data demonstrate that immunization with the NP vaccine generates a local cytokine milieu that supports induction of CD8⁺ T cells with a T_RM phenotype.

Figure 6. Acute cytokine response to nanoparticle vaccine is localized, transient, and supportive of lung-resident CD8⁺ T cells.

Cytokines associated with CD8⁺ T cells (IFNγ, IL-1β, IL-6, IL-12p70) and T_RM generation (TNFα, IFNβ, IL-33, IFNα) were measured in lungs, BAL, and serum obtained 6 h, 24 h, 48 h, or 7 d after i.n. administration OVA-NP/CpG or OVA+CpG. Data are mean ± SEM and representative of two independent experiments, with n = 4–5 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg CpG. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, by ordinary two-way ANOVA with Tukey’s multiple comparisons test. Statistical differences shown are for comparison of OVA-NP/CpG vs. OVA+CpG. See also Figure S6.

Nanoparticle Vaccine Generates Long-Lasting Populations of Lung-Resident Antigen-Specific CD8⁺ T Cells That Express T_RM Surface Markers

The ability of the NP vaccine to enhance the lung-resident CD8⁺ T cell response after 13 days (Figure 3) prompted us to determine whether antigen-specific CD8⁺ T cells present in murine lungs at 30 and 60 days after immunization, which are considered memory T cells,⁹ possessed a characteristic T_RM phenotype. In addition to being defined as CD45⁻ by i.v. staining, lung T_RM have been defined by surface expression of CD69—an activation marker that limits tissue egress by inhibiting expression of sphingosine-1-phosphate receptor—and CD103—an adhesion molecule that binds E-cadherin on epithelial cells and retains T_RM in their home tissues.⁶⁰ Both CD103⁻CD69⁺ and CD103⁺CD69⁺ T_RM subsets in the lungs have been reported.^{15, 52} To characterize the memory phenotype of Tet⁺ CD8⁺ T cells in the IST and AW, we immunized mice i.n. with a single dose of OVA-NP/CpG, OVA-NP, or OVA+CpG, and on d30 or d60 postimmunization, harvested lungs and spleens and quantified Tet⁺ CD8⁺ T_RM. Here, CXCR3 was used as a marker of airway residence, as previously reported.^{21, 52, 71} Staining with αCD103 and αCD69 antibody was used in conjunction with i.v. αCD45 antibody to identify T_RM, and αCXCR3 antibody was used to discriminate AW-resident cells (CXCR3^hi) from those resident in the IST (CXCR3^lo) (Figure 7A).

Figure 7. OVA-specific CD8⁺ T cells are maintained at memory timepoints and express T_RM markers CD69 and CD103.

(A) Mice were immunized i.n. with OVA-containing formulations on d0 and lungs and spleens were analyzed on d30 or d60 via tetramer and surface marker staining. CXCR3 was used as a marker of AW residence; CD103 and CD69 were used as markers of tissue residency. (B-E) Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (B) AW, (C) IST, (D) MV, and (E) spleen were enumerated on d30 after i.n. administration of OVA-NP/CpG, OVA-NP, or OVA+CpG. (F-I) Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (F) AW, (G) IST, (H) MV, and (I) spleen were enumerated on d60 after i.n. administration of OVA-NP/CpG, OVA-NP, or OVA+CpG. (J) Flow cytometry was used to quantify Tet⁺ CD8⁺ T cells expressing T_RM markers (CD103, CD69) in the airway (CXCR3^hi) and lung interstitium (CXCR3^lo). (K-L) Number (#) of Tet⁺ CD8⁺ T cells expressing CD69±CD103 in AW were enumerated on (K) d30 or (L) d60 after immunization. (M-N) Number (#) of Tet⁺ CD8⁺ T cells expressing CD69±CD103 in IST were enumerated on (M) d30 or (N) d60 after immunization. Data are mean ± SEM and representative of two to four independent experiments, with n = 5–6 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg. CpG Limit of detection: 1 cell (AW), 5 cells (IST/MV), 25 cells (spleen). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, by (B-I) ordinary one-way ANOVA or (K-N) ordinary two-way ANOVA with Tukey’s multiple comparisons test. ns, not significant. Statistical comparisons are shown for OVA-NP/CpG only.

At both d30 (Figures 7B–7E) and d60 (Figures 7F–7I) post-immunization, there were significantly more Tet⁺ CD8⁺ T cells in the lungs of mice immunized with OVA-NP/CpG relative to those receiving OVA+CpG. There were also generally more cells in the OVA-NP/CpG group relative to mice receiving OVA-NP, although in certain instances this difference was not statistically significant, consistent with the intrinsic capacity of the NP to enhance the CD8⁺ T cell response (Figure 3). At d30, OVA-NP/CpG generated a significantly higher response than OVA+CpG in the AW and IST (Figures 7B–7C), but not in the MV or spleen (Figures 7D–7E). The number of cells present in the lungs of mice immunized with OVA-NP/CpG were significantly higher in the IST relative to OVA-NP (Figure 7C), but in the AW they were not significantly different (Figure 7B). While there was no significant difference between any group in the MV (Figure 7D), OVA-NP/CpG generated a significantly higher response than OVANP in the spleen (Figure 7E). At d60, there were significantly more cells in the IST for the OVA-NP/CpG group relative to both OVA-NP and OVA+CpG (Figure 7G), while the difference between groups in the AW was less pronounced (Figure 7F). Again, there was no significant difference between any group in the MV (Figure 7H), while OVA-NP/CpG continued to produce a significantly higher response in the spleen relative to OVA-NP (Figure 7I). While in most cases the OVA-NP formulation produced a greater response than OVA+CpG, at d60 it remained inferior to OVA-NP/CpG in the IST, where the majority of CD8⁺ T_RM are located. Overall, these data show that the NP vaccine is able to enhance generation of lung-resident CD8⁺ T cells that are maintained for at least 60 days after immunization.

We next asked whether long-lasting antigen-specific CD8⁺ T cells in the lungs expressed characteristic T_RM surface markers. At these same time points, AW (CXCR3^hi) and IST (CXCR3^lo) cells were analyzed for CD103 and CD69 expression (Figure 7J). At d30, i.n. administration of OVA-NP/CpG elicited significantly more AW and IST Tet⁺ CD8⁺ T_RM with both CD103⁻CD69⁺ and CD103⁺CD69⁺ phenotypes when compared to immunization with OVA+CpG (Figures 7K and 7M). There were also significantly more AW and IST CD8⁺ T_RM of the CD103⁺CD69⁺ phenotype in the OVA-NP/CpG group relative to the OVA-NP group (green bars); for the CD103⁻CD69⁺ phenotype (blue bars), this difference was not significant, although the number of cells was still higher. At d60, in the IST, OVA-NP/CpG induced significantly more CD8⁺ T_RM of both phenotypes than either control group (OVA-NP and OVA+CpG) (Figure 7N). In the AW, this difference was only significant for OVA-NP/CpG vs. OVA+CpG in the CD103⁻CD69⁺ phenotype, although OVA-NP/CpG still produced the highest number of cells in all cases (Figure 7L). These data indicate that the NP vaccine is superior to control formulations for generating antigen-specific CD8⁺ T_RM in the lungs, particularly at 60 days after immunization.

Single-Dose Pulmonary Immunization with Nanoparticle Vaccine Protects Against Sublethal Respiratory Virus Challenge

Ultimately, an effective vaccine must generate T cells that protect against subsequent infectious challenge. Given the ability of a single i.n. dose of the NP vaccine to generate antigen-specific CD8⁺ T_RM and retain them in the lungs for up to 60 days post-immunization, we next determined whether these T_RM were protective against infection with a respiratory virus. To do this, we immunized mice i.n. with OVA-NP/CpG, OVA-NP, OVA+CpG, or PBS. On d30 or d60 post-immunization, mice were challenged i.n. with a recombinant vaccinia virus expressing influenza virus nucleoprotein, SIINFEKL peptide, and enhanced green fluorescent protein (VV.NP-S-EGFP). Since this virus expresses the SIINFEKL epitope of OVA, it provides a tool for evaluating the ability of NP vaccine-induced CD8⁺ T cells to protect against i.n. challenge. Mice were inoculated with a sublethal dose (1×10⁷ pfu) of the virus and weighed daily through d6 post-inoculation (p.i.) (Figure 8A). On d6 p.i., lungs were harvested for quantification of viral load.

Figure 8. Mice immunized with nanoparticle vaccine exhibit less weight loss and lower viral burden after intranasal challenge with recombinant vaccinia virus.

(A) Mice immunized on d0 were challenged i.n. with recombinant SIINFEKL-expressing vaccinia virus (sublethal dose of 1×10⁷ pfu/mouse) either 30 or 60 d post-immunization. Mice were weighed daily and lungs were harvested on d6 post-inoculation (p.i.). (B-C) Percent (%) weight loss in mice challenged on (B) d30 or (C) d60 after i.n. administration. Left: weight loss over time. Right: weight at d6 p.i. expressed as % initial body weight. (D-E) Lungs of mice challenged on (D) d30 or (E) d60 post-immunization were harvested on d6 p.i. for quantification of viral load. Dots show titers for individual animals. Data are mean ± SEM, with (B) n = 12–23 per group, (C) n = 5–6 per group, and (D-E) n = 5 per group. Immunization dose: 25 μg NP, 7.5 μg OVA, 1.4 μg CpG. Data are pooled from one to four independent experiments. Limit of detection for plaque assay = 6 pfu. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, by (B-C, left) repeated measures two-way ANOVA with Tukey multiple comparisons test or (B-C, right; D-E) ordinary one-way ANOVA with Tukey multiple comparisons test.

At both d30 and d60, mice immunized with the NP vaccine were significantly protected from challenge-induced weight loss relative to all other formulations. Mice immunized with OVA-NP/CpG lost ~5–7% of body weight by d6 p.i. vs. a loss of ~13–26% in other groups (Figures 8B–8C). Additionally, the OVA-NP/CpG group in the 30-day cohort began regaining weight by d5 p.i. (Figure 8B).

To further validate these findings, viral load was quantified in lungs harvested from infected mice at d6 p.i. There was a 1–2 log reduction in viral burden in the lungs of mice immunized with OVA-NP/CpG, relative to the other formulations, at both d30 and d60 post-immunization (Figures 8D–8E). Overall, these results demonstrate that a single dose of the NP vaccine provides significant CD8⁺ T_RM-mediated protection against respiratory virus challenge in an antigen-specific manner.

Nanoparticle Vaccine Containing Influenza Virus Protein Generates Antigen-Specific CD8⁺ T_RM in the Lungs

After demonstrating the ability of the NP vaccine to generate CD8⁺ T_RM in the lungs and protect against respiratory virus challenge using a model antigen (OVA), we next asked whether these findings could be applied to a more clinically relevant antigen and infection model. To this end, we shifted our focus to influenza, a respiratory infection of global importance and a significant public health challenge.⁷² It is well-established that CD8⁺ T_RM cells are important for generating heterosubtypic immunity against influenza A viruses in both mice and humans.^{15, 18, 24, 55} To evaluate the ability of the NP vaccine to protect against this pathogen, we selected nucleoprotein, a structural protein from influenza A H1N1 virus (strain A/Puerto Rico/8/1934; PR8), as the antigen, since there is precedent for using the PR8 strain in murine influenza challenge models. While surface proteins like hemagglutinin and neuraminidase are commonly used in experimental flu vaccines to generate a humoral response, internal viral antigens like nucleoprotein are known to contain the majority of influenza CD8⁺ T cell epitopes.^73–75 Thus, using nucleoprotein as a vaccine antigen lends itself to specifically studying the protective effect of an immunization-induced CD8⁺ T cell response. In addition, because CD8⁺ T cell epitopes of influenza A virus are largely conserved across strains and subtypes, they may be particularly well-suited for providing broad protection, and while CD8⁺ T cells typically do not generate sterilizing immunity, they are useful for reducing disease severity and pathogen transmission.^{24, 73, 76, 77}

We will henceforth refer to the nucleoprotein antigen as “Flu” to avoid confusion with the “NP” abbreviation for nanoparticle. Thus, formulations will be indicated as Flu-NP/CpG (nucleoprotein conjugated to nanoparticle and complexed with CpG), Flu-NP (nucleoprotein conjugated to nanoparticle), and Flu+CpG (soluble nucleoprotein mixed with CpG). We first validated that the Flu antigen could be covalently conjugated to NP via the same chemistry used to load OVA protein. Flu protein was thiolated as described for OVA and reacted with pH-responsive NP to generate Flu-NP conjugates (Figure S7A). Flu-NP conjugates were then electrostatically complexed with CpG to create Flu-NP/CpG (Figure S7B). Both Flu-NP/CpG and Flu-NP were characterized by DLS (Figure S7C).

We next assessed the ability of Flu-NP/CpG and related control formulations to generate antigen-specific CD8⁺ T cells expressing lung T_RM markers at memory timepoints (CD103⁻CD69⁺ and CD103⁺CD69⁺ CD8⁺ T_RM). To do this, we prepared a fluorescent MHC-I tetramer (Tet) containing ASNENMETM peptide, a known immunodominant H-2D^b I epitope for nucleoprotein from PR8 virus.⁷⁸ We then immunized mice i.n. with a single dose of Flu-NP/CpG, Flu-NP, or Flu+CpG, and on d30 or d60 post-immunization, harvested lungs and spleens and quantified Tet⁺ CD8⁺ T cells in the same manner as described for OVA (Figure 9A). Similarly, we also assessed expression of T_RM markers CD69 and CD103 on the antigen-specific CD8⁺ T cell populations in the AW and IST lung compartments.

Figure 9. Flu-specific CD8⁺ T cells are maintained at memory timepoints and express T_RM markers CD69 and CD103.

(A) Mice were immunized i.n. with Flu-containing formulations on d0 and lungs and spleens were analyzed on d30 or d60 via Tet and surface marker staining. CXCR3 was used as a marker of AW residence; CD103 and CD69 were used as markers of tissue residency. (B-E) Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (B) AW, (C) IST, (D) MV, and (E) spleen were enumerated on d30 after i.n. administration of Flu-NP/CpG, Flu-NP, or Flu+CpG. (F-I) Number (#) and frequency (%) of Tet⁺ CD8⁺ T cells in (F) AW, (G) IST, (H) MV, and (I) spleen were enumerated on d60 after i.n. administration of Flu-NP/CpG, Flu-NP, or Flu+CpG. (J-M) Number (#) of Tet⁺ CD8⁺ T cells expressing CD69±CD103 in AW were enumerated on (J) d30 or (K) d60 after immunization. (L-M) Number (#) of Tet⁺ CD8⁺ T cells expressing CD69±CD103 in IST were enumerated on (L) d30 or (M) d60 after immunization. Data are mean ± SEM, with n = 3–6 per group, and representative of two independent experiments. Immunization dose: 25 μg NP, 9.5 μg Flu, 1.4 μg CpG. Limit of detection: 1 cell (AW), 5 cells (IST/MV), 25 cells (spleen). *p<0.05, **p<0.01, ***p<0.001, by (B-I) ordinary one-way ANOVA or (J-M) ordinary two-way ANOVA with Tukey’s multiple comparisons test. ns, not significant.

At d30, mice immunized with Flu-NP/CpG had significantly more Tet⁺ CD8⁺ T cells in the AW, IST, and MV compartments than mice immunized with Flu+CpG (Figure 9B–9E). The difference between Flu-NP/CpG and Flu-NP groups was less pronounced, with significance only in the MV; however, a general trend of greater numbers of Tet⁺ CD8⁺ T cells produced by Flu-NP/CpG in the lungs was observed. There was no significant difference between treatment groups in the spleen. At d60, the number of cells generated by the Flu-NP/CpG, Flu-NP, and Flu+CpG groups were not significantly different in any lung compartment or in the spleen (Figure 9F–9I), though in the IST both Flu-NP/CpG and Flu-NP exhibited higher cell counts than Flu+CpG (Figure 9G). These data show that, similarly to the NP vaccine containing OVA, the NP vaccine with Flu antigen is able to generate higher numbers of lung-resident CD8⁺ T cells that are maintained for at least 30 days after immunization. However, by 60 days they have waned to similar levels as those seen in control groups.

We next examined whether these long-lasting antigen-specific CD8⁺ T cells in the lungs expressed T_RM surface markers. At 30d and 60d post-immunization, AW (CXCR3^hi) and IST (CXCR3^lo) cells were analyzed for CD103 and CD69 expression, as previously shown in Figure 7J. At d30, Flu-NP/CpG immunization generated more T_RM of both phenotypes (CD103⁻CD69⁺ and CD103⁺CD69⁺) in both the AW and IST (Figures 9J and 9L). Again, at d60, the differences between treatment groups were less pronounced, with the Flu-NP group in some instances producing equivalent or greater numbers of T_RM relative to the Flu-NP/CpG (Figures 9K and 9M). Both Flu-NP/CpG and Flu-NP were superior to Flu+CpG in the IST (Figures 9L–9M). Overall, these data indicate that immunization with a nanoparticle-containing formulation is superior to a soluble formulation for generating Flu-specific CD8⁺ T_RM in the lungs, and that addition of CpG adjuvant increases efficacy similarly to what was observed with model antigen (OVA).

Single-Dose Pulmonary Immunization with Nanoparticle Vaccine Protects Against Lethal Challenge with Influenza A H1N1 Virus

As with previous experiments using OVA protein antigen, the ultimate test of efficacy for a vaccine is its ability to protect against infection. Thus, we sought to determine whether a single-dose of NP vaccine with Flu antigen could similarly protect against respiratory challenge. To this end, we immunized mice i.n. with Flu-NP/CpG, Flu-NP, Flu+CpG, or PBS, and on d30 or d60 post-immunization, challenged i.n. with PR8 virus. In this lethal challenge model, mice were inoculated with 200 FFU of PR8, a dose at which untreated animals experience severe weight loss and disease symptoms. Mice were monitored daily through d14 p.i. for weight loss, morbidity, and mortality (Figure 10A).

Figure 10. Mice immunized with nanoparticle vaccine exhibit improved survival after intranasal challenge with influenza A H1N1 virus.

(A) Mice immunized on d0 were challenged i.n. with influenza A H1N1 PR8 virus (lethal dose of 200 FFU/mouse) either 30 d or 60 d post-immunization. Mice were weighed daily and evaluated for morbidity/mortality. (B-C) Percent (%) weight loss in mice challenged on (B) d30 or (C) d60 after i.n. administration. (D-E) Survival of mice challenged on (D) d30 or (E) d60 post-immunization. Mice that exceeded 30% weight loss were considered deceased. Data are mean ± SEM, with n = 4–6 per group, from two independent experiments. Immunization dose: 25 μg NP, 9.5 μg Flu, 1.4 μg CpG. Statistical significance in survival curves was determined with a Mantel-Cox log-rank test (**p<0.01; ns, not significant).

At d30, mice immunized with either Flu-NP/CpG or Flu-NP were protected from challenge-induced weight loss relative to Flu+CpG and naïve (PBS-treated) mice (Figure 10B). Mice in the Flu-NP/CpG and Flu-NP groups lost less weight overall, and around d7 p.i., began to recover instead of continuing to lose weight. In addition, 83% of Flu-NP/CpG mice and 67% of Flu-NP mice survived challenge, while 100% of Flu+CpG mice and PBS mice succumbed to infection (>30% weight loss) by d7 or d8 p.i., respectively (Figure 10D).

At d60, the protective effect of the Flu-NP/CpG formulation was maintained (Figure 10C), consistent with results seen in the vaccinia challenge model. In addition, the survival rate for the Flu-NP/CpG group was the same (83%), while for the Flu-NP group it dropped to 33% (Figure 10E). As before, 100% of mice in the Flu+CpG and PBS groups perished by d9 p.i. Overall, these results demonstrate that a single dose of the NP vaccine formulated with a clinically relevant antigen (influenza A nucleoprotein) offers protection against lethal respiratory virus challenge. Nucleoprotein is an internal viral protein, and so it is a target for protective T cell responses. Given that T_RM specific for the immunodominant CD8⁺ T cell epitope of nucleoprotein are detected at the highest frequency in lungs of mice immunized with Flu-NP/CpG, our data suggest these CD8⁺ T_RM cells contribute to protection.

CONCLUSION

As the importance of tissue-resident memory T cells in defense against disease becomes increasingly clear, efforts are turning toward developing mucosal vaccines that can induce a strong and protective T_RM response against infectious pathogens and cancers.^{13, 26, 43, 44, 79–82} Here, using a mouse model of pulmonary immunization, we demonstrated that a single dose of pH-responsive NP vaccine provided extended co-delivery of antigen and adjuvant to pulmonary APCs, produced a cytokine milieu supportive of tissue-resident CD8⁺ cells, induced a CD8⁺ T_RM response that persisted for up to 60 days after immunization, and was protective against both lethal and sublethal challenge with respiratory viruses (influenza and vaccinia, respectively). Antigen-specific CD8⁺ T cells displayed surface markers characteristic of T_RM (CD103, CD69), and intravascular staining showed them to be resident in the lung interstitium (CD45⁻CXCR3^lo) and airways (CD45⁻CXCR3^hi). Both the pH-responsive functionality of the carrier and its capacity for dual-delivery were crucial for inducing a CD8⁺ T_RM response, which highlights the significance of engineering nanomaterial properties in the design of T_RM vaccines. We also showed that the route of administration was important for generating lung-resident cells; i.n. administration of the NP vaccine induced more antigen-specific CD8⁺ T cells in the IST and AW than did s.c. immunization. Notably, unlike previous reports of T_RM vaccines in which multiple doses are typically administered, these results were obtained with a single dose, which offers potential translational advantages such as increased compliance and dose reduction.

Our findings suggest that several key properties of the NP vaccine are linked to the resulting CD8⁺ T_RM response. It has been proposed that local antigen recognition and persistence in non-lymphoid tissues, including the lungs, can promote formation of T_RM cell populations.^{47, 54, 83, 84} Takamura et al. found that CD8⁺ T cell encounter with cognate antigen in the lung, but not in the lung-draining lymph node, was critical to conversion from circulating to resident cells in the lungs.²⁰ Notably, the NP we describe here increased retention of vaccine cargo in pulmonary APCs for up to three days after pulmonary immunization. This finding suggests that NP-mediated delivery can extend antigen residence in the lungs over soluble antigen formulations and enhance the pulmonary CD8⁺ T_RM response. This could be due to the cationic surface charge of the NP, which may confer mucoadhesive properties that extend residence time in the lung.^{69, 85} It could also reflect the capacity of pH-responsive NP carriers to prolong intracellular residence time by avoiding endosomal recycling and lysosomal degradation.⁴¹ Longer-term studies that further examine the duration of antigen persistence and its effects on the T_RM response are warranted, and may motivate the use of pH-responsive materials as particle depots to control antigen release and delivery.

It has also been reported that certain pulmonary APC subsets, including CD103⁺ DCs and alveolar macrophages, can promote establishment of lung T_RM.^55–57 In our experiments, alveolar macrophages took up large amounts of OVA and CpG over the course of 72 h and therefore may have contributed to the T_RM response. More importantly, we observed uptake in CD103⁺ dendritic cells, the predominant cross-presenting DC subset in mucosal tissues.^{86, 87} The pH-responsive nature of the NP, which allows it to transport antigen to the cytosol, potentially facilitated cross-presentation and priming of CD8⁺ T_RM responses by CD103⁺ DCs. The relative contributions that NP-mediated cytosolic antigen delivery vs. intrinsic mechanisms of DC cross-presentation provide in the induction of a CD8⁺ T_RM response in the lungs merit further investigation. We speculate that both mechanisms contribute to the response. In addition, we found that CD103⁺ DCs were the cell type most potently activated by the NP vaccine, with CD86 expression sustained for at least three days post-immunization. NP delivery enhanced co-localization of CpG with OVA in CD103⁺ DCs during this time, which likely contributed to robust CD86 expression.⁸⁸ Methods of guiding vaccine delivery specifically to CD103⁺ DCs (e.g., antibody targeting) may be a promising approach for increasing vaccine uptake in this APC subset that is important for promoting T_RM formation.⁵⁵

Several surface markers have been used to describe T_RM cells in various non-lymphoid tissues—in particular, CD103 and CD69.^{18, 80, 89} However, these markers are not found on all T_RM; many T_RM populations located outside epithelia do not express CD103.⁹⁰ CD69 is more universally expressed and has been suggested to influence accumulation and retention of CD8⁺ T cells in the lungs during the early stages of infection.²⁰ Both CD103⁻CD69⁺ and CD103⁺CD69⁺ T_RM subsets have been shown in the lungs.^{52, 91, 92} In our experiments, we found the CD103⁻CD69⁺ phenotype to be somewhat more prevalent than CD103⁺CD69⁺ in the IST and AW. It has been suggested that infiltration of tumors with CD8⁺ T_RM expressing CD103 correlates to longer survival of patients with a variety of cancers, including breast, lung, ovarian, cervical, and bladder.⁶⁰ Notably, TGFβ is needed to induce expression of CD103 on T_RM,^{62, 93, 94} and we were unable to detect this cytokine at several timepoints after immunization. In light of this, methods to tailor the NP vaccine to stimulate TGFβ production or otherwise generate more CD103-expressing CD8⁺ T_RM cells may warrant further study. Nevertheless, the NP vaccine provided significant protection against challenge with respiratory virus, demonstrating its translational promise.

In conclusion, this report demonstrates generation of protective CD8⁺ T_RM in the lungs with a single mucosally-administered dose of a pH-responsive nanoparticle vaccine. Intranasal dual-delivery of antigen and adjuvant with pH-responsive NP was shown to enhance the magnitude and functionality of the lung-resident CD8⁺ T cell response, as well as increase antigen persistence and activation in pulmonary APCs. Efficacy of the nanoparticle vaccine was demonstrated with both a model antigen (ovalbumin) and a clinically relevant antigen (influenza A nucleoprotein). Antigen-specific CD8⁺ T cells in the lungs displayed characteristic T_RM markers at memory timepoints, and mice were protected against respiratory virus infection in both sublethal and lethal challenge models. The use of intravascular staining to identify lung-resident cells, in conjunction with staining for markers of tissue resident memory, has enabled relationships between NP vaccine properties and the generation of T_RM to be established. This NP vaccine provides a modular platform technology for delivery of any number of clinically-relevant protein or peptide antigens, as well as other nucleic acid adjuvants. Additionally, there are several practical advantages to this system that lend themselves to the possibility of scale-up and translation, such as the ability to synthesize NP on a large scale and sterilize them by filtration.⁹⁵ It would also be feasible to develop this vaccine as a needle-free aerosol formulation, to facilitate clinical translation and simple administration without the need for skilled healthcare workers. Overall, this NP system represents a promising technology for the development of T_RM vaccines against respiratory infections such as influenza, other pathogens that target non-lymphoid tissues, and mucosal cancers.

METHODS

RAFT Synthesis of (PDSMA-co-DMAEMA)-b-(PAA-co-DMAEMA-co-BMA)

RAFT copolymerization of pyridyl disulfide ethyl methacrylate (PDSMA) and dimethylaminoethyl methacrylate (DMAEMA) was conducted under a nitrogen atmosphere in dioxane (40 wt % monomer) at 30 °C for 18 h, as previously described.³⁷ PDSMA monomer was synthesized according to a previously reported procedure.⁹⁶ The RAFT chain transfer agent (CTA) used was 4-cyano-4-(ethylsulfanylthiocarbonyl) sulfanylpentanoic acid (ECT) and the initiator used was 2,2’-azobis(4-methoxy-2,4-dimethylvaleronitrile) (V-70) (Wako Chemicals, Richmond, VA). The initial molar ratio of DMAEMA to PDSMA was 92:8, and the initial monomer ([M]₀) to CTA ([CTA]₀) to initiator ([I]₀) ratio was 100:1:0.05. The resultant poly(PDSMA-co-DMAEMA) macro-chain transfer agent (mCTA) was isolated by precipitation (6×) into pentane. A schematic of the mCTA polymerization reaction can be found in Scheme S1A.

Purified mCTA was dried in vacuo for one week and used for block copolymerization with DMAEMA, propylacrylic acid (PAA), and butyl methacrylate (BMA) to create a pH-responsive polymer, as described previously.^{37, 97} DMAEMA (30%), PAA (30%), and BMA (40%) ([M]₀/[mCTA]₀ = 450) were added to the mCTA dissolved in dimethylacetamide (DMAc) (40 wt % monomer and mCTA) along with V-70 initiator ([mCTA]₀/[I]₀ = 2.5). Polymerization took place under a nitrogen atmosphere for 24 h at 30 °C. The resultant diblock copolymer was isolated by dialysis against acetone using a 3.5 kDa MWCO membrane, followed by dialysis against deionized water. The purified diblock copolymer was lyophilized for 72 h prior to use. A schematic of the pH-responsive polymerization reaction can be found in Scheme S1B.

Polymer composition and monomer conversion of both the mCTA and diblock copolymer were characterized by ¹H NMR spectroscopy (CDCl₃) on a Bruker AV400 spectrometer (Figures S1A and S1B). Gel permeation chromatography (GPC, Agilent) with DMF containing 0.1 M LiBr as the mobile phase and in-line light scattering (Wyatt) and refractive index (Agilent) detectors was used to determine molecular weight (MW) and polydispersity indices (PDI) of both the mCTA and diblock copolymer (Figure S1D, Table S1). Molecular weights were determined using dn/dc values calculated previously (0.071 for mCTA and 0.065 for diblock). Characterization was done according to previously published methods.³⁷ Representative NMR spectra, GPC traces, and a summary of polymer properties can be found in Figure S1 and Table S1.

RAFT Synthesis of (PDSMA-co-DMAEMA)-b-(BMA)

Purified mCTA used for the synthesis described above was also used for block copolymerization with poly(butyl methacrylate) (pBMA) to create a non-pH-responsive control polymer. Monomer was added to mCTA ([M]₀/[mCTA]₀ = 300) and dissolved in dioxane (40 wt % monomer and mCTA) along with V-70 initiator ([mCTA]₀/[I]₀ = 20), then polymerized under a nitrogen atmosphere for 24 h at 35 °C. The resultant diblock copolymer was isolated by dialysis as described above. The purified polymer was then lyophilized, and its composition, molecular weight, and polydispersity index were analyzed using ¹H NMR (CDCl₃) spectroscopy and GPC (Figures S1C and S1D, Table S1), according to previously published methods.⁹⁸ The control polymerization reaction can be seen in Scheme S1C.

Preparation and Characterization of Nanoparticles

Self-assembled micellar nanoparticles (NP) were obtained by first dissolving lyophilized polymer at 50 mg/ml in 100% ethanol, then rapidly pipetting dissolved polymer into 100 mM phosphate buffer (pH 7) to a final concentration of 10 mg/ml. Nanoparticles were formulated in the same way for both pH-responsive (NP_pH) and control (NP_ctrl) polymers. For in vivo studies, ethanol was removed by buffer exchange into PBS (pH 7.4) via 3 cycles of centrifugal dialysis (Amicon, 3 kDa MWCO, Millipore), and NP solutions were then sterilized via syringe filtration (Whatman, 0.22 μm, GE Healthcare). Final polymer concentration was determined with UV-Vis spectrometry (Synergy H1 Multi-Mode Reader, BioTek) by measuring absorbance of aromatic PDS groups at 280 nm. Size of the NP was measured via dynamic light scattering (DLS). NP solutions were prepared at a concentration of 0.1–0.2 mg/ml in PBS (pH 7.4) and the hydrodynamic radius was measured using a Malvern Instruments Zetasizer Nano ZS Instrument (Malvern, USA). Representative DLS data for both polymers at physiological pH (7.4) can be found in Figure S1E. In addition, size change of NP_pH but not NP_ctrl at pH 5.8, as measured by DLS, can be seen in Figure S1F (left).

Erythrocyte Lysis Assay

The degree to which the pH-responsive polymer was able to induce pH-dependent lysis of lipid bilayer membranes (thus leading to cytosolic delivery) was assessed via a red blood cell hemolysis assay as previously described.⁹⁹ Briefly, polymers (10 μg/ml) were incubated for 1 h at 37 °C in the presence of human erythrocytes in 100 mM sodium phosphate buffer. Buffers in the pH range of the endosomal processing pathway (7.4, 7.0, 6.6, 6.2, and 5.8) were used. Extent of cell lysis (i.e., endosomolytic activity) was determined via UV-Vis spectrometry by measuring the amount of hemoglobin released (Abs = 541 nm) (Figure S1F, right). Absorbances were normalized to a 100% lysis control (1% Triton X-100). Samples were run in quadruplicate.

Preparation of Antigen-Nanoparticle Conjugates

A model antigen, ovalbumin protein (OVA), was conjugated to pendant PDS groups on NP via thiol-disulfide exchange. For conjugate characterization, OVA from chicken egg white (MilliporeSigma) was used; for in vivo studies, endotoxin-free (<1 EU/mg) EndoFit™ OVA (Invivogen) was used. In some experiments, OVA was labeled with fluorescein isothiocyanate isomer (FITC; Sigma) for evaluating conjugation efficiency via fluorescent imaging of SDS-PAGE gels, or with Alexa Fluor 647-NHS ester (AF647; Thermo Fisher Scientific) for tracking conjugates after in vivo administration. Following manufacturer’s instructions, dye was added to OVA for a degree of labeling of ~1 FITC/OVA or ~0.5 AF647/OVA.

To prepare OVA for conjugation, free amines on the protein were thiolated by incubation with ~25 molar excess of 2-iminothiolane (Traut’s Reagent, Thermo Fisher Scientific) in reaction buffer (100 mM phosphate buffer, pH 8, supplemented with 1 mM EDTA) as previously described.³⁷ Unreacted 2-iminothiolane was removed by buffer exchanging thiolated OVA into 1X PBS (pH 7.4) using Zeba™ Spin desalting columns (0.5 ml, 7 kDa MWCO, Thermo Fisher Scientific). For in vivo studies, thiolated OVA was sterilized via syringe filtration (0.22 μm, Millipore). Following manufacturer’s instructions, the molar ratio of thiol groups to OVA protein was determined with Ellman’s reagent (Thermo Fisher Scientific) to be ~3–5 thiols/OVA. Polymer NP solutions were reacted with thiolated OVA at various molar ratios of pH-responsive NP:OVA (5:1, 10:1, 20:1) or control NP:OVA (3.5:1, 7:1, 14:1) to make OVA-NP_pH and OVA-NP_ctrl conjugates, respectively. The conjugation ratio for the control polymer was adjusted to maintain a constant dose of antigen for both carriers. Conjugation was done overnight, in the dark, at room temperature, and under sterile conditions (when needed), as previously described.³⁷ Antigen conjugation was verified via non-reducing SDS-polyacrylamide gel electrophoresis (SDS-PAGE) using 4–20% Mini-Protean TGX Precast Protein Gels (Bio-Rad) (Figure S2A). Gels were run at 130 V for 1 h and imaged with a Gel Doc™ EZ System (Bio-Rad). A conjugation ratio of 5:1 (pH) or 3.5:1 (ctrl) was used for all in vivo formulations in order to maximize the amount of antigen delivered. DLS was used to measure the size of OVA-NP conjugates, as described above (Figure S2C, top and bottom left).

Influenza A H1N1 nucleoprotein (Flu) formulated in sterile phosphate buffer was obtained from Sino Biological (Beijing, China). To prepare Flu antigen for conjugation, free amines were thiolated by incubation with ~250 molar excess of 2-iminothiolane in reaction buffer. Unreacted 2-iminothiolane was removed by buffer exchange into sterile 1X PBS. Following manufacturer’s instructions, the molar ratio of thiol groups to Flu protein was determined with a Measure-iT™ Thiol Assay Kit (Thermo Fisher Scientific). The concentration of Flu protein after thiolation and purification was measured using the Pierce™ Rapid Gold BCA Protein Assay Kit (Thermo Fisher Scientific). In some cases, Flu was labeled with AF647 prior to thiolation for evaluating conjugation efficiency via fluorescent imaging of SDS-PAGE gels. Nanoparticles were reacted with thiolated Flu at a molar ratio of 5:1 pH-responsive NP:Flu to make Flu-NP conjugate. Antigen conjugation was verified via SDS-PAGE (Figure S7A) and gels were imaged with an IVIS Lumina III Imaging System (PerkinElmer, Waltham, MA). DLS was also used to measure the size of Flu-NP conjugates (Figure S7C).

Formation of Nanoparticle/Adjuvant Complexes

NP/adjuvant complexation was carried out by combining CpG ODN 1826 (Invivogen) with NP, OVA-NP, or Flu-NP in PBS at room temperature for at least 30 min. Theoretical charge ratios (+/−) of 4:1 and 6:1 were tested. The charge ratio was defined as the molar ratio between protonated DMAEMA tertiary amines in the first block of the copolymer (positive charge; assuming 50% protonation at physiological pH) and phosphate groups on the CpG backbone (negative charge).³⁷ The charge ratios at which complete complexation of CpG to the polymer occurred were determined via an agarose gel retardation assay (Figure S2B for OVA and Figure S7B for Flu). Free CpG, OVA+CpG, Flu+CpG, and NP/CpG, OVA-NP/CpG, and Flu-NP/CpG complexes prepared at various charge ratios were loaded into lanes of a 4% agarose gel and run at 90 V for 30 min. Gels were stained with GelRed® Nucleic Acid Gel Stain (Biotium, Fremont, CA) for 20 min and visualized with a Gel Doc™ EZ System (Bio-Rad). A charge ratio of 6:1 was used for all in vivo formulations in order to maximize the stability of the formulation. DLS was used to measure the size of the OVA-NP/CpG formulation (Figure S2C, top right) and Flu-NP/CpG formulation (Figure S7C, bottom), as described above.

Animals

Male or female C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME), maintained at the animal facilities of Vanderbilt University under either conventional, specific pathogen-free (SPF barrier facility), or animal biosafety level 2 (ABSL-2) conditions, and experimented upon in accordance with the regulations and guidelines of Vanderbilt University Institutional Animal Care and Use Committee (IACUC).

Intranasal Immunization

Endotoxin-free OVA (<1 EU/mg, EndoFit™), sterile buffer solutions (1X PBS, pH 7.4), and sterile polymer solutions with ethanol removed were used for vaccine formulations. Experimental groups were: (1) nanoparticles loaded with covalently-conjugated OVA and complexed with CpG DNA (OVA-NP/CpG); (2) nanoparticles conjugated to OVA (OVA-NP); (3) a mixture of OVA (non-thiolated) and nanoparticles (OVA+NP); (4) a mixture of CpG-complexed nanoparticles and non-thiolated OVA (NP/CpG+OVA); (5) a mixture of non-thiolated OVA and CpG (OVA+CpG); (6) OVA conjugated to non-pH-responsive control polymer (OVA-NP_ctrl); and (7) PBS for sham mice. For all groups containing “NP,” this denotes the pH-responsive polymer. Conjugates were prepared 1–2 days before use and stored at 4 °C. OVA was thiolated and used immediately for conjugation to NP at a molar ratio of 5:1 (NP:OVA), as described above. On the day of use, CpG was complexed to conjugates at a 6:1 charge ratio via rapid pipetting of CpG DNA (~0.5 mg/ml) into the conjugate solution, as described above. The formulation was allowed to react for at least 30 min at room temperature for complete complexation of CpG before administration to mice.

In experiments using OVA antigen, male mice (8–12 weeks old) were anesthetized with ketamine/xylazine (10 mg/ml ketamine hydrochloride, Vedco; 1 mg/ml xylazine hydrochloride, Vanderbilt Pharmacy) by intraperitoneal (i.p.) injection (~200 μl anesthesia/22 g mouse weight). Anesthetized mice were immunized intranasally (i.n.) on day 0 with formulations containing 7.5 μg OVA and/or 1.4 μg CpG with or without 25 μg polymer. In dosing pilot studies, doses of 50 μg polymer (15.1 μg OVA, 2.8 μg CpG) and 12.5 μg polymer (3.8 μg OVA, 0.7 μg CpG) were also tested; the 25 μg polymer dose was ultimately selected for its ability to induce a robust CD8⁺ T cell response with minimal toxicity. Vaccine formulations in a total volume of 80 μl PBS were delivered via pipette through the nostrils into the lungs of mice; inoculation with this volume allows formulations to reach the lower airways.¹⁰⁰ The dose was applied at the center of the nose to allow inhalation into both nostrils, at a rate of ~8 μl/s. In some cases, anesthetized mice were instead immunized with a subcutaneous (s.c.) injection at the base of the tail. Animals were monitored either daily or thrice weekly for weight loss and signs of morbidity.

In experiments using Flu antigen, female mice (10 weeks old) were anesthetized as described and immunized i.n. on day 0 with formulations containing 9.5 μg Flu and/or 1.4 μg CpG with or without 25 μg polymer. Vaccine formulations in a total volume of 80 μl PBS were delivered through the nostrils as described.

Measurement of Antigen-Specific CD8⁺ T Cell Response and Tissue-Resident Memory Markers

On day 13, 30, or 60 after immunization, mice were anesthetized and intravenously (i.v.) injected with 200 μl of anti-CD45.2-APC antibody (clone 104; Tonbo) at 0.01 mg/ml (2 μg αCD45 antibody per mouse), as previously described.⁵² This was done to stain marginated vascular leukocytes (MV; CD45⁺) and differentiate them from those resident in the lung interstitium (IST; CD45⁻).⁵¹ To allow for circulation of αCD45 antibody, mice were rested for 3–5 min after i.v. injection and prior to CO₂ euthanasia. For experiments done at d13, lungs of euthanized mice were perfused with PBS to collect broncheoalveolar lavage fluid (BAL) from the airway compartment (AW) while maintaining IST and MV populations in the lung parenchyma.⁵² For d30 and d60 experiments, staining with αCXCR3 antibody was used to define airway residence of CD8⁺ T cells, and lung perfusion was omitted.⁵² Lungs and spleens were then collected from each mouse. Organs were harvested and processed as previously described.¹⁰¹ Briefly, lungs were minced with a scalpel and incubated for 1 h at 37 °C in complete RPMI medium (cRPMI [RPMI+10% FBS]; Gibco) supplemented with 2 mg/ml collagenase (Sigma) and 50 nM dasatinib (LC Laboratories, Woburn, MA). Lungs and spleens were treated with ACK lysing buffer (Gibco) and passed through 70 μm cell strainers to generate single cell suspensions.

Cell suspensions from BAL, lungs, and spleens were stained for 1 h at 4 °C with anti-B220-FITC (clone RA3–6B2; BD Biosciences), anti-CD4-FITC (clone H129.19; BD Biosciences), anti-CD11b-FITC (clone M1/70; Tonbo), anti-CD11c-FITC (clone N418; Tonbo), anti-CD8α-Pacific Blue (clone 53–6.7; BD Biosciences), and 1.5 μg/ml PE-labeled OVA_257–264 (SIINFEKL)-H-2K^b tetramer (Tet) prepared according to a previously reported procedure.¹⁰² In experiments utilizing mice immunized with Flu antigen, cells were instead stained with PE-labeled Flu_366–374 (ASNENMETM)-H-2D^b Tet prepared using the same method. Antibodies labeled with FITC (B220/CD4/CD11b/CD11c) were referred to as the “dump” channel and were used to exclude B cells, CD4⁺ T cells, dendritic cells, and macrophages from gating. In experiments evaluating tissue-resident memory markers on d30 and d60, cells from lungs and spleens were also stained with anti-CD69-PE/Cy7 (clone H1.2F3; Tonbo), anti-CD103-Brilliant Violet 510 (clone 2E7; BioLegend), and anti-CXCR3-PerCP/Cy5.5 (clone CXCR3–173; BioLegend). In this case, BAL was not collected, as the number of antigen-specific CD8⁺ T cells present in BAL samples at these timepoints is too low for accurate flow cytometric quantification.

After staining, cells were washed with FACS buffer (PBS supplemented with 2% FBS and 50 nM dasatinib) and stained with propidium iodide (BD Biosciences) or Ghost Dye™ Red 780 (Tonbo) to discriminate live vs. dead cells. AccuCheck counting beads (Thermo Fisher Scientific) were included in samples to allow for calculation of absolute cell counts. The frequency of antigen-specific CD8⁺ T cells was determined by flow cytometry on a 3-laser LSR-II flow cytometer (BD). All data were analyzed using FlowJo Software (version 10.4.2; Tree Star, Inc., Ashland, OR). Cells were gated by forward and side scatter to exclude debris and doublets. Viable antigen-specific CD8⁺ T cell populations were defined as follows: AW = CD8α⁺CD45⁻Tet⁺ cells in BAL samples (CXCR3^hi in lung samples for select experiments); IST = CD8α⁺CD45⁻Tet⁺ cells in lung samples (CXCR3^lo in select experiments); MV = CD8α⁺CD45⁺Tet⁺ cells in lung samples; SPL = CD8α⁺Tet⁺ cells in spleen samples. All cells in the CD8α⁺ gate were also B220⁻CD4⁻CD11b⁻CD11c⁻ (“dump channel”). T_RM cells were defined as either CD103⁺CD69⁺ or CD103⁻CD69⁺ CD8⁺ T cells in lung samples. Representative gating for each sample type can be found in Figure S3A–3C.

Intracellular Cytokine Staining of Antigen-Specific CD8⁺ T Cells

On day 13 after immunization, lungs and spleens were harvested and processed to obtain single-cell suspensions. Cells were plated in 96-well V-bottom plates at 3×10⁶ cells/well (lung) or 2×10⁶ cells/well (spleen) in cRPMI and re-stimulated with 10 μM of MHC class I epitope SIINFEKL peptide (OVA_257–264; Invivogen). Instead of treatment with peptide, positive controls were treated with PMA (50 ng/ml; Invivogen) and ionomycin (2 μg/ml; Sigma) and negative controls were treated with cRPMI. Cells were incubated at 37 °C and 5% CO2 for 1 h 30 min. BD GolgiPlug™ protein transport inhibitor (BD Biosciences) was then added to each well and cells were incubated for an additional 5 h 30 min.³⁷ After incubation, cells were washed with PBS and stained with eFluor® 450 fixable viability dye (eBioscience) for 30 min at 4 °C. Cells were next washed with FACS buffer (PBS+2% FBS) and stained with anti-CD8α-APC/Cy7 (clone 53–6.7; Tonbo) and anti-CD3ε-PerCP/Cy5.5 (clone 145–2C11; Tonbo), as well as Fc-block (anti-CD16/CD32, clone 2.4G2; Tonbo), for 1 h at 4 °C. Cells were washed 2× in FACS buffer, then fixed and permeabilized by incubating for 10 min at 4 °C with BD Cytofix/Cytoperm (BD Biosciences), according to manufacturer instructions. Cells were then washed 2× with 1X BD Perm/Wash Buffer (BD Biosciences) and incubated for 1 h at 4 °C with antibodies against intracellular cytokines: anti-IFNγ-APC (clone XMG1.2; BD Biosciences) and anti-TNFα-PE (clone MP6-XT22; BD Biosciences). Finally, cells were washed once with 1X Perm/Wash buffer, resuspended in FACS buffer supplemented with 50 nM dasatinib, and analyzed by flow cytometry using a 3-laser LSR-II flow cytometer (BD) and FlowJo software (v.10.4.2). Data are reported as the percentage of CD8α⁺CD3ε⁺ cells that are IFNγ⁺ and/or TNFα⁺ after subtraction of background values from negative (unstimulated) controls. Representative gating for lungs and spleens can found in Figure S3D.

Histology

Lungs from mice immunized with OVA-NP/CpG or OVA+CpG, or from untreated mice, were harvested on d1 and d12 post-immunization. Tissue was fixed in 10% neutral buffered formalin, processed routinely, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E). Sections were evaluated by an experienced veterinary pathologist blinded to the composition of the groups. Representative images are provided in Figure S4B.

Fluorescence Microscopy

Mice were immunized i.n. with antigen-NP conjugates containing OVA labeled with Alexa Fluor 647 (OVA₆₄₇-NP). After 24 h, mice were injected i.v. with anti-CD45-Brilliant Violet 421 antibody (clone 30-F11; BD Biosciences) and Alexa Fluor 488-labeled tomato lectin (Vector Laboratories) to visualize intravascular lung leukocytes and vasculature, respectively.

Mice were euthanized and lungs were harvested and fixed by inflation with 1 mL of 4% paraformaldehyde followed by 15% sucrose administered through the trachea. Lungs were frozen in OCT (Fisher Scientific). Ten-micron tissue sections were evaluated by fluorescence microscopy using an Axioplan widefield microscope (Zeiss) equipped with a 20x objective, 405-, 488-, 532-, and 633-nm laser lines, and a Hamamatsu ORCA-ER monochrome digital camera.

Uptake and Activation in Pulmonary Innate Immune Cells

To identify the effects of the NP vaccine on innate immune cell uptake and activation in lungs, mice were immunized with fluorescently labeled OVA₆₄₇-NP/CpG₄₈₈ or OVA₆₄₇+CpG₄₈₈, or with PBS (control). In the fluorescent formulations, OVA was labeled with Alexa Fluor 647 as described above (OVA₆₄₇). Alexa Fluor 488-labeled CpG (CpG₄₈₈) was purchased from Integrated DNA Technologies (IDT; Skokie, IL). After 24, 48, or 72 h, mice were euthanized and lungs were harvested, as well as spleens (to assess systemic biodistribution). Organs were imaged using an IVIS Lumina III Imaging System (PerkinElmer, Waltham, MA) to visualize and quantify tissue-level OVA₆₄₇ fluorescence after immunization. IVIS image files were analyzed using Living Image® software (version 4.5.5, PerkinElmer).

After imaging, lungs were processed as described above to obtain single-cell suspensions. Lung samples were stained for flow cytometric analysis of pulmonary immune cells using a modified version of the panel described by Misharin et al.⁵⁸ This panel was used to distinguish seven different cell types: (1) alveolar macrophages (AMφ); (2) interstitial macrophages (IMφ); (3) CD103⁺ dendritic cells (CD103⁺ DC); (4) CD11b⁺ dendritic cells (CD11b⁺ DC); (5) monocytes/macrophages (Mono/Mφ); (6) granulocytes (Gran); and (7) cells not included in other populations (Other). It was also used to quantify the amount of cargo co-localization (OVA⁺CpG⁺ cells) and expression of activation marker CD86 in each cell subset and at each timepoint. The following antibodies were used: anti-CD64-PE (BioLegend; X54–5/7.1), anti-CD24-PE/Cy7 (BioLegend; M1/69), anti-CD11b-PerCP/Cy5.5 (BioLegend; M1/70), anti-CD11c-APC/Cy7 (Tonbo; N418), anti-I-A/I-E-Brilliant Violet 605 (BD; M5/114.15.2), antiCD45.2-Brilliant Violet 650 (BioLegend; 104), and anti-CD86-PE/Dazzle™ 594 (BioLegend; GL-1). Ghost Dye™ Violet 510 (Tonbo) was used to discriminate live vs. dead cells and AccuCheck counting beads (Thermo Fisher Scientific) were included in samples to allow for calculation of absolute cell counts. Samples were stained with viability dye for 30 min at 4 °C, washed with FACS buffer (PBS+2% FBS, 50 nM dasatinib), incubated with Fc-block (anti-CD16/CD32, clone 2.4G2; Tonbo) for 15 min at 4 °C, and then stained for 1 h at 4 °C with the antibody panel listed above. Finally, cells were washed once, resuspended in FACS buffer, and analyzed by flow cytometry. Data were collected using a 3-laser Fortessa (BD) and analyzed with FlowJo software (v.10.4.2). Representative gating for OVA and CpG uptake in lung cell subsets can be found in Figure S5.

Measurement of Cytokines

Cytokines were measured in serum, BAL, or lung homogenates using a LEGENDplex™ bead-based immunoassay (BioLegend). Mice were immunized i.n. and blood, BAL, and lungs were harvested 6 h, 24 h, 48 h, and 7 d after immunization. Blood was obtained by cardiac puncture and BAL was collected by lavage with 1 ml sterile PBS containing a cocktail of protease inhibitors (Roche cOmplete™ Mini EDTA-free Protease Inhibitor Cocktail, Sigma). The right sides (4 lobes) of lungs were collected in 1 ml M-PER™ Mammalian Protein Extraction Reagent (Thermo Fisher Scientific) with protease inhibitors and homogenized using a gentleMACS™ Octo Dissociator and M tubes (Miltenyi Biotec), according to manufacturer instructions. Lung homogenates were then centrifuged for 10 min at 4200 rpm and supernatants were collected and frozen at −80 °C until analysis. BAL samples were frozen at −80 °C without further processing. Blood was centrifuged for 10 min at 14,000 rpm (2×), and sera were collected and frozen at −80 °C until analysis. Prior to use with LEGENDplex™ kits, samples were thawed and centrifuged for 10 min at 10,000–12,500 rpm to remove debris. Lung samples were filtered through 40 μm cell strainers for additional debris removal. The following cytokines were measured: IFNγ, TNFα, IFNα, IFNβ, IL-6, IL-33, IL-12p70, and IL-1β. LEGENDplex™ kits were used according to manufacturer instructions. Flow cytometric data was collected with a 3-laser LSR II (BD) and analyzed with LEGENDplex Data Analysis Software (v.8.0).

Vaccinia Virus Propagation, Intranasal Virus Challenge, and Lung Burden

Recombinant vaccinia virus expressing influenza virus nucleoprotein, ovalbumin SIINFEKL peptide, and enhanced green fluorescent protein (VV.NP-S-EGFP) was obtained through the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH (NR-624; BEI Resources, Manassas, VA). The virus was grown in HeLa cells and titrated using BSC-40 cells. For titration, crystal violet stain (Eng Scientific, Clifton, NJ) was used to visualize plaques 48 h after applying serial 10-fold dilutions of the virus in HBSS+0.5% (w/v) BSA to confluent monolayers of BSC-40 cells.

For respiratory virus challenge, 10- to 16-week-old immunized male mice were anesthetized i.v. with ketamine/xylazine as described above and inoculated i.n. with a sublethal dose (1×10⁷ pfu) of virus in 80 μl sterile PBS. Mice were monitored daily for morbidity and weight loss. On day 6 post-infection, lungs from individual mice were harvested into 2 ml HBSS (supplemented with 0.5% (w/v) BSA and 1X pen/strep, sterilized by vacuum filtration) and frozen at −80 °C.

To determine viral burden using a plaque assay, previously frozen lungs were thawed, homogenized in HBSS using a Tissue Tearor (BioSpec Products, Bartlesville, OK), and subjected to one additional freeze-thaw cycle. Serial 10-fold dilutions of lung homogenates were plated on confluent monolayers of BSC-40 cells. After 48 h, plaques were visualized by crystal violet staining.

Influenza A Virus Challenge

Influenza A virus (strain A/Puerto Rico/8/1934, subtype H1N1) (PR8) was obtained through BEI Resources, NIAID, NIH (NR-348; Manassas, VA). For respiratory virus challenge, 14- to 18-week-old immunized female mice were anesthetized i.v. with ketamine/xylazine as described above and inoculated i.n. with a lethal dose (200 FFU) of virus in 80 μl sterile PBS. Mice were monitored daily through d14 post-infection for morbidity, weight loss, and survival. After infection, mice were euthanized when weight loss exceeded 30% of initial body weight, in accordance with IACUC guidelines.

Statistical Analysis

Statistical analyses were performed as indicated in figure legends. All analyses were done using GraphPad Prism software, version 6.07. Results are expressed as mean ± SEM with ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 being considered statistically significant.