Strategy to enhance dendritic cell-mediated DNA vaccination in the lung

Abstract

We here introduce a new paradigm to promote pulmonary DNA vaccination. Specifically, we demonstrate that nanoparticles designed to rapidly penetrate airway mucus (mucus-penetrating particle or MPP) enhance the delivery of inhaled model DNA vaccine (i.e. ovalbumin-expressing plasmids) to pulmonary dendritic cells (DC), leading to robust and durable local and trans-mucosal immunity. In contrast, mucus-impermeable particles were poorly taken up by pulmonary DC following inhalation, despite their superior ability to mediate DC uptake in vitro compared to MPP. In addition to the enhanced immunity achieved in mucosal surfaces, inhaled MPP unexpectedly provided significantly greater systemic immune responses compared to gold-standard approaches applied in the clinic for systemic vaccination, including intradermal injection and intramuscular electroporation. We also showed here that inhaled MPP significantly enhanced the survival of an orthotopic mouse model of aggressive lung cancer compared to the gold-standard approaches. Importantly, we discovered that MPP-mediated pulmonary DNA vaccination induced memory T-cell immunity, particularly the ready-to-act effector memory-biased phenotype, both locally and systemically. The findings here underscore the importance of breaching the airway mucus barrier to facilitate DNA vaccine uptake by pulmonary DC and thus to initiate full-blown immune responses.

Graphical Abstract

We show that nanoparticles designed simply to breach the airway mucus barrier markedly enhance the co-delivery of antigen-encoding plasmids and nucleic acid-based adjuvants to the pulmonary DC, thereby leading to robust and durable immune responses in the lung and beyond to the level unachievable by gold-standard approaches applied in clinic.

1. Introduction

Human lung airways from the trachea to bronchioles are rich in dendritic cells (DC) that persistently sample inhaled foreign matters deposited on the airway lumen to initiate antigen-specific immune responses.^[1] Thus, pulmonary vaccination provides a straightforward means to combat against inhaled pathogens, including influenza and Mycobacterium tuberculosis, or to eradicate lung cancers if immunologically active. Indeed, it is now well appreciated that pulmonary vaccination elicits stronger immune responses in the lung and is more efficient in clearing respiratory pathogens compared to other systemic immunization strategies.^[2] In particular, inhaled DNA vaccination, by promoting both humoral and cellular immunity, can potentially manage numerous lung diseases both in prophylactic and therapeutic manners.^[3] Of note, DNA vaccine holds additional advantageous features in comparison to traditional subunit vaccines, including ease and speed of scale-up, superior stability and potential of global distribution without a need of expensive and cumbersome ‘cold-chain’.^{[4, 5]} Nevertheless, clinical trials of DNA vaccination targeting lung diseases to date have primarily explored conventional systemic approaches (e.g., intramuscular electroporation).^{[4, 6]}

The field of DNA vaccination has been primarily focusing on strategies to enhance DC uptake of DNA vaccines, such as molecular targeting and electroporation.^[7] However, a largely overlooked challenge to inhaled vaccination is the mucus gel layer lining the lung airways, which may hamper the access of inhaled DNA vaccines to pulmonary DC reside in the airway mucosa. The airway mucus is a protective barrier that effectively traps inhaled foreign matters, including large DNA (e.g., plasmid-based DNA vaccines),^[8] via adhesive and/or physical interactions, and rapidly clears them from the lung via the physiological mucociliary clearance (MCC) mechanism.^{[9, 10]} While gene transfer agents (i.e., gene vectors) has been shown to improve diffusion of DNA in mucus to a certain degree,^[11] conventional virus-based and synthetic gene vectors, including those tested in clinical trials, cannot efficiently penetrate human airway mucus.^{[8, 12, 13]} To this end, these gene vectors, following inhaled administration, are unlikely to shuttle DNA vaccine payloads efficiently to pulmonary DC prior to the MCC to induce a robust immune response in the lung.

We have recently developed strategies to engineer gene vectors that precisely address the aforementioned challenge. Specifically, we found that gene vectors designed to possess small particle diameters (~ 50 nm), muco-inert surface coatings (hydrophilic and near-neutrally charged) and excellent colloidal stability in the physiological lung environment were capable of efficiently penetrating human airway mucus.^{[8, 13–15]} Following inhaled administration into the mouse lungs, these mucus-penetrating particles (MPP) provided widespread airway distribution, prolonged lung retention and/or robust transgene expression in the airway epithelium.^{[13, 15]} The findings underscore the ability of the MPP to efficiently penetrate the airway mucus gel layer and avoid the MCC to reach the underlying airway tissue in vivo. We thus hypothesize that the MPP may be utilized to deliver DNA vaccine components, including antigen-expressing plasmids and nucleic acid-based adjuvants, simultaneously to pulmonary DC to mediate strong and durable immunity in the lung and potentially in other mucosal surfaces.

2. Results

2.1. Mucus penetration enhances uptake of DNA-loaded nanoparticles by pulmonary dendritic cells

We engineered nanoparticles capable of efficiently penetrating airway mucus for inhaled delivery of model antigen-expressing plasmids, using a blend method reported in our previous study.^[15] Briefly, a mixture of poly(β amino ester) (PBAE) and polyethylene glycol (PEG)-conjugated PBAE (PEG-PBAE) at an optimized ratio was used to compact ovalbumin (OVA)-expressing plasmids (pOVA) to yield mucus-penetrating particles (pOVA-MPP). In parallel, mucus-impermeable conventional particles carrying pOVA (pOVA-CP) were formulated with PBAE only. The pOVA-MPP exhibited small particle diameters of 55 ± 1 nm and near-neutral surface charges of 1.6 ± 0.3 mV (table S1), physicochemical properties that render particles muco-inert and permeable to airway mucus.^{[10, 16]} In contrast, pOVA-CP possessed larger particle diameters of 120 ± 4 nm and highly positive, muco-adhesive, surface charges of 32 ± 2 mV (Table S1). We then confirmed widespread distribution and deep penetration of pOVA-MPP in the mucus-covered lung airways in vivo following intratracheal administration (Figure 1A). As expected from the particle properties, identically administered pOVA-CP were sparsely distributed as aggregates and primarily localized at mucosal surface lumen away from the airway epithelium (Figure 1B).

Figure 1. In vivo mucus penetration and DC uptake of different DNA-loaded nanoparticles carrying fluorescently labeled pOVA following intratracheal administration.

Representative confocal images demonstrating penetration of pOVA-loaded nanoparticles (magenta), including (A) pOVA-MPP and (B) pOVA-CP, through mouse lung airway mucus 1 hour after the administration (Left; scale bar = 200 𝜇m). DAPI staining represents cell nuclei (blue). The areas enclosed by white boxes are zoomed in (Right; scale bar = 50 𝜇m). (C) Representative flow cytograms demonstrating pOVA-MPP and pOVA-CP uptake by pulmonary DC at 16-hour post-administration. Data shown as mean ± SEM. (D) Percentages of pulmonary DC that took up pOVA-MPP and pOVA-CP at 16-hour post-administration (n = 6). Data shown as mean ± SEM. ***p < 0.0005 by one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison test.

We next went on to investigate our primary hypothesis that efficient mucus penetration would be critical to particle access to, and subsequent uptake by, pulmonary DC following localized administration into the lung airways. Specifically, female inbred C57BL/6 mice intratracheally received fluorescently-labeled pOVA-MPP or pOVA-CP, and uptake of either formulation by pulmonary DC (CD11⁺CD170^-) was quantified 16 hours after the administration. Flow cytometric analysis revealed that DC uptake of pOVA-MPP (25% ± 11%) was significantly greater than that of pOVA-CP (6.8% ± 1.3%) (Figure 1, C and D). We then compared in vitro DC uptake of two formulations to test whether the enhanced DC uptake observed with pOVA-MPP in vivo was attributed to its superior DC-targeting and/or endocytic capacity. However, the trend was reversed in vitro where 3.5% ± 2% and 31% ± 2% of DC were found positive for pOVA-MPP and pOVA-CP uptake, respectively (Figure S1).

2.2. Adjuvant-loaded pOVA-MPP efficiently penetrate human airway mucus ex vivo and activate DC in vitro

To capitalize the critical roles of adjuvants on vaccination,^[17] we next formulated pOVA-MPP with inclusion of short nucleic acid-based adjuvants targeting intracellular toll-like-receptors (TLR); TLR molecules abundant in endocytic compartments of DC promote cross-presentation.^[18] We engineered MPP formulations with a mixture of pOVA and either of p(I:C) (TLR3 agonist) and CpG (TLR9 agonist) and confirmed that the inclusion of adjuvants did not compromise the mucus-penetrating physicochemical properties of pOVA. Specifically, particle sizes (i.e., hydrodynamic diameters), surface charges and colloidal stability (i.e., changes in hydrodynamic diameters and polydispersity index values in phosphate buffered saline (PBS) over time) of pOVA-MPP co-loaded with p(I:C) (p(I:C)/pOVA-MPP) or CpG (CpG/pOVA-MPP) were virtually identical to those of pOVA-MPP formulated without adjuvants (Figure 2, A and B and table S1). We next evaluated the ability of our MPP formulation to protect nucleic acid payloads against extracellular nucleases. Gel electrophoretic analysis revealed that all the payloads, including pOVA, p(I:C) and CpG, remained intact in the formulation following nuclease challenge, unlike carrier-free nucleic acids (Figure 2C). Of note, CpG resisted the DNase-mediated degradation, regardless of the packaging, due to its intrinsically nuclease-resistant, phosphorothioate backbone.^[19]

Fig 2. In vitro characterization of pOVA-MPP formulations with or without adjuvant co-packaging.

(A) Transmission electron micrographs of pOVA-MPP, p(I:C)/pOVA-MPP and CpG/pOVA-MPP. Scale bars = 200 nm. (B) Changes in hydrodynamic diameters (Left) and polydispersity index (PDI) values (Right) of pOVA-MPP, p(I:C)/pOVA-MPP and CpG/pOVA-MPP in PBS over 6 hours (n = 3 – 9). Data shown as mean ± SEM. (C) Electrophoretic analysis showing the ability of p(I:C)/pOVA-MPP (Left) and CpG/pOVA-MPP (Right) to protect respective nucleic acid payloads in presence of DNase and/or RNase. Red and black arrows indicate pOVA and adjuvants, respectively. Leftmost lanes denote the 1 kb plus DNA ladders. (D) Median MSD of pOVA-MPP, p(I:C)/pOVA-MPP and CpG/pOVA-MPP in freshly collected human airway mucus samples (n = 3 – 8) at a timescale of 1 s. The MSD values are directly proportional to particle diffusion rates. Data shown as median ± SEM. p < 0.05, and **p < 0.005 by one-way ANOVA with Kruskal-Wallis multiple comparison test. (E) Percentage of CD11c+ DC co-expressing MHC-II+ and CD86+ following a 6-hour incubation with various adjuvants or particle formulations (n = 4 – 6). Data shown as mean ± SEM. *p < 0.05, **p < 0.005 and ***p < 0.0005 by one-way ANOVA with Dunnett’s multiple comparison test.

Using multiple particle tracking (MPT) analysis,^[20] we then investigated whether the inclusion of adjuvants impact on the mucus-penetrating property of pOVA-MPP in airway mucus freshly expectorated from patients visiting the Johns Hopkins Adult Cystic Fibrosis Center. The pathological airway mucus is highly viscoelastic due to mucus build-up and/or chronic infection/inflammation, which is a hallmark of numerous obstructive lung diseases and reinforces the airway mucus as a delivery barrier.^{[9, 10]} The MPT measures various transport parameters, such as mean square displacement (MSD); MSD is a measure of the distances traveled by individual particles at a given time interval (i.e., timescale) and thus is directly proportional to particle diffusion rates.^{[20, 21]} We confirmed that all MPP formulations, including pOVA-MPP, p(I:C)/pOVA-MPP and CpG/pOVA-MPP, exhibited comparably high MSD values (Figure 2D), in agreement with our prior observation with MPP formulations carrying reporter-expressing plasmids.^[15] In contrast, pOVA-CP were unable to efficiently diffuse in the human airway mucus, displaying significantly lower MSD values compared to all MPP formulations (Figure 2D). We next sought to investigate the abilities of different formulations to induce DC maturation in vitro. We found that both p(I:C)/pOVA-MPP and CpG/pOVA-MPP significantly increased the percentage of DC positive for maturation markers (CD86⁺MHC-II⁺;^[22]) compared to untreated control, carrier-free adjuvants and the adjuvant-free counterpart (i.e., pOVA-MPP) (Figure 2E). Out of two different adjuvant-loaded MPP formulations, CpG/pOVA-MPP showed the greatest level of matured DC (Figure 2E), and thus further investigation was conducted with CpG/pOVA-MPP.

2.3. Intratracheally administered CpG/pOVA-MPP traffic to the local lymph node via DC and enhance effector T-cell responses

Efficient delivery of DNA vaccines to DC and subsequent migration to lymph node (LN) is critical steps in inducing robust cytotoxic T-cell (CTL) immune response.^[23] As shown above, pOVA-MPP were efficiently taken up by pulmonary DC in vivo (Figure 1, C and D) and adjuvant-loaded pOVA-MPP were capable of enhancing DC maturation in vitro (Figure 2E). Thus, we next sought to mechanistically determine the trafficking of CpG/pOVA-MPP to mediastinal LN using immunohistochemical analysis. We found that two days after a single intratracheal administration (Figure 3A), fluorescently-labeled CpG/pOVA-MPP were found co-localized with pulmonary DCs (CD11c⁺CD170^-) in the lung airway interstitium (Figure 3B) and subsequently trafficked to mediastinal LN (Figure 3C).

Figure 3. Pulmonary DC uptake, LN trafficking and antigen-specific CD8+ T-cell response following intratracheal administration of CpG/pOVA-MPP.

(A) Experimental immunization schedule. (B) Representative confocal image showing uptake (white arrows) of CpG/pOVA-MPP (magenta) by pulmonary DC (CD11c+; yellow). Scale bar = 20 𝜇m. (C) Representative confocal image showing CpG/pOVA-MPP (magenta) localization in mediastinal LN. Scale bar = 200 𝜇m. Cell nuclei are stained with DAPI (blue). (D) Percentage of CD8+ T-cells (CD3ε+ CD8+) expressing OVA-specific MHC-I peptide (SIINFEKL) in lung, respective draining LN and spleen following pOVA-mediated DNA vaccination via different administration routes (n = 5 – 12). ID: intradermal; IM-EP: intramuscular electroporation; IT; intratracheal. Data shown as mean ± SEM. *p < 0.05 and ***p < 0.0005 by one-way ANOVA with Dunnett’s multiple comparison test.

We then sought to evaluate induction of OVA-specific CTL response in vivo following a cycle of immunization and boost with the identical formulation (Figure 3A). The level of OVA-specific CTL response was quantitatively determined by flow cytometric analysis of effector T-cells (CD3ε⁺CD8⁺) positive for SIINFEKL-MHC I pentamer where SIINFEKL is an antigenic epitope of OVA protein.^[24] Intratracheally (IT) administered CpG/pOVA-MPP showed significantly greater OVA-specific CTL responses in the lung, mediastinal LN and spleen compared to identically administered carrier-free CpG/pOVA that exhibited the responses comparable to those of the naïve untreated control (Figure 3D). We also compared IT CpG/pOVA-MPP with carrier-free CpG/pOVA administered via other conventional delivery routes, including intradermal (ID) injection or intramuscular injection followed by electroporation (IM-EP). We found that IT CpG/pOVA-MPP exhibited significantly greater OVA-specific CTL responses compared to both ID CpG/pOVA and IM-EP CpG/pOVA in all three different tissues, including the lung, respective LN (i.e., mediastinal and inguinal LN for IT and ID/IM-EP, respectively) and spleen (Figure 3D). In particular, IT CpG/pOVA-MPP resulted in ~40% of OVA-specific CTL in the lung, unlike all other control groups that exhibited negligible levels of pulmonary CTL responses.

We further analyzed immune responses by evaluating the responsiveness of OVA-specific helper (CD4⁺) and effector (CD8⁺) T-cells to ex vivo re-stimulation. Specifically, CD4⁺ and CD8⁺ T-cells in the lung (Figure S2A), respective LN (Figure S2B) and spleen (Figure S2C) were harvested 7 days after the boost, re-stimulated ex vivo with phorbol myristate acetate (PMA)/ionomycin or SIINFEKL peptide,^[24] and T-cells producing interferon-γ (IFN-γ) were quantified using flow cytometry. Similar to our observation with the in vivo CTL response study (Figure 3D), IT CpG/pOVA-MPP group exhibited the greatest frequency of activated CD8⁺ T-cells (IFN-γ⁺CD8⁺) uniformly in all three different tissues (Figure S2). We also confirmed that IT CpG/pOVA-MPP yielded significantly greater frequencies of pulmonary (Figure S2A) and splenic (Figure S2C) CD4⁺ T-cell activation (i.e., percentage of IFN-γ⁺CD4⁺ T-cells) upon ex vivo re-stimulation, compared to all other test conditions, including IT CpG/pOVA, ID CpG/pOVA and IM-EP CpG/pOVA groups.

2.4. Intratracheal immunization with CpG/pOVA-MPP mediates OVA-specific effector CD8⁺ T-cell dissemination to distal mucosal tissues

To assess the capacity of IT CpG/pOVA-MPP to induce OVA-specific CTL responses in distal mucosal tissues, we analyzed CD8⁺ T-cells harvested from gastrointestinal (GI) and vaginal tract after intratracheal immunization. Similar to the earlier studies, we compared IT CpG/pOVA-MPP with other carrier-free and/or conventional route controls, including IT CpG/pOVA, ID CpG/pOVA and IM-EP CpG/pOVA groups. Mice were immunized as scheduled in Figure 3A and the OVA-specific effector CD8⁺ T-cells positive for SIINFEKL-MHC I pentamer in mesenteric LN and Peyer’s patches in GI as well as the whole vagina were quantified using flow cytometry 7 days after the boost. We found that IT CpG/pOVA-MPP group exhibited significantly greater frequencies of OVA-specific CD8⁺ T-cells both in GI and vaginal tracts compared to all the control groups (Figure 4A). We also confirmed that IT CpG/pOVA-MPP significantly increased the percentage of CD8⁺ T-cells (CD3ε⁺CD8⁺) harvested from the mediastinal LN that express a gut-homing integrin α₄β₇ compared to the carrier-free IT CpG/pOVA group 7 days after the boost immunization (Figure S3).

Figure 4. Trans-mucosal CD8+T-cell responses following pulmonary immunization with CpG/pOVA-MPP.

Percentage of CD8+ T-cells (CD3ε+CD8+) expressing OVA-specific MHC-I peptide (SIINFEKL) in mesenteric LN, Payer’s patch and vagina (A) with (n =4 – 5) or (B) without (n = 3 – 5) adoptive T-cell transfer. CD8+ T cells from OT-I mice were adoptively transferred into C57BL/6 mice one day prior to immunization with CpG/pOVA-MPP and OVA-specific T-cell response was quantified 3 days after the immunization. ID: intradermal; IM-EP: intramuscular electroporation; IT; intratracheal. Data shown as mean ± SEM. *p < 0.05, **p < 0.005, ***p < 0.0005 and ****p < 0.00005 by (A) one-way ANOVA with Dunnett’s multiple comparison test or (B) two-sided Student’s t-test.

We next sought to further confirm dissemination of OVA-specific CD8⁺ T-cells to distal mucosal tissues by IT CpG/pOVA-MPP using an adoptive T-cell transfer study. Specifically, C57BL/6 mice received an intravenous injection of CD8⁺ T-cells harvested from OT-I mice and subsequently immunized with IT CpG/pOVA-MPP or left untreated on the following day. Three days after the immunization, significantly greater frequencies of OVA-specific CD8⁺ T-cell response were detected via flow cytometry in both GI and vaginal tracts compared to the untreated control (Figure 4B).

2.5. Intratracheal immunization with CpG/pOVA-MPP establishes OVA-specific long-term CD8⁺ T-cell responses and effector memory–biased immunity in the lung

It is critical to establish antigen-specific memory to achieve a long-term protective or therapeutic immunity. We thus first assessed the long-term OVA-specific CTL response mediated by IT CpG/pOVA-MPP in comparison to the carrier-free IT pOVA-MPP control 70 days after the immunization. Similar to our observation at 21-day post-immunization (i.e. 7 days after the boost) (Figure 3D), the MPP formulation significantly increased the percentage of OVA-specific CD8⁺ T-cells in the lung, mediastinal LN and spleen (Figure 5A). To confirm the establishment of OVA-specific CD8⁺ T-cell memory, we further analyzed the cells for the expression of memory-associated surface markers, including CD44 and CD62L.^[25] Of note, central (T_CM) and effector (T_EM) memory T-cells are distinguished by relative expression of CD62L where T_CM and T_EM exhibit CD44^hiCD62L^hi and CD44^hiCD62L^lo, respectively.^[25] We found at 70 days post-immunization that IT CpG/pOVA-MPP markedly increased both memory phenotypes compared to carrier-free IT CpG/pOVA control in the lung, mediastinal LN and spleen (Figure 5B). In addition, IT CpG/pOVA-MPP established OVA-specific T-cell memory strongly biased towards the effector memory phenotype both at the site of administration (i.e., lung) and systemically in the spleen (Figure 5B). In particular, the bias was most prominent in the lung with over 4-fold greater frequency of T_EM compared to that of T_CM.

Figure 5. Memory T-cell responses following pulmonary immunization with CpG/pOVA-MPP.

(A) Percentage of CD8+ T-cells (CD3ε+CD8+) expressing OVA-specific MHC-I peptide (SIINFEKL) in lung (Left), mediastinal LN (Middle) and spleen (Right) 70 days after the immunization (n = 4 – 5). ***p < 0.0005 by one-way ANOVA with Dunnett’s multiple comparison test. Data shown as mean ± SEM. (B) Percentage of OVA-specific central memory T cells (T_CM) co-expressing CD44^hi and CD62L^hi and effector memory T cells (T_EM) co-expressing CD44^hi and CD62L^lo in lung (Left), mediastinal LN (Middle) and spleen (Right) 70 days after the immunization (n = 4 – 8). Data shown as mean ± SEM. *p < 0.05, **p < 0.005, ***p < 0.0005 by two-sided Student’s t-test with Holm-Sidak multiple comparison test.

2.6. Intratracheally immunization with CpG/pOVA-MPP enhances anti-cancer effect in an orthotopic mouse model of aggressive lung cancer

Encouraged by the aforementioned observation, we next sought to determine whether the pulmonary vaccination with CpG/pOVA-MPP provided an enhanced anti-cancer effect in an orthotopic syngeneic mouse model of lung cancer in comparison to conventional route (i.e. ID CpG/pOVA and IM-EP CpG/pOVA) controls. To do this, we first immunized C57BL/6 mice following the schedule in Figure 6A, inoculated mice intratracheally with highly aggressive and poorly immunogenic.^{[26, 27]} Lewis lung carcinoma (LLC) cells expressing SIINFEKL, and monitored their survival over time (Figure 6A). While the survival of mice received ID CpG/pOVA were comparable to that of untreated naïve mice, IM-EP CpG/pOVA was able to significantly enhance the survival compared to these two groups (Figure 6B). However, IT CpG/pOVA-MPP group exhibited by far greatest survival compared to all other groups; the median survival days of mice in the IT CpG/pOVA-MPP, IM-EP CpG/pOVA and ID CpG/pOVA groups and of untreated mice were 42, 26, 15 and 13 days, respectively (Figure 6B). We confirmed that sham administration via intradermal and intratracheal routes did not affect the survival of the animal model (Figure S4).

Figure 6. Survival of an orthotopic mouse model of OVA-expressing lung cancer following pulmonary immunization with CpG/pOVA-MPP.

(A) Experimental schedule. C57BL/6 mice were immunized as described in Figure 3A and OVA-LLC cells were intratracheally inoculated into the lung 7 days after the boost. Mice received CpG/pOVA or CpG/OVA-MPP via different administration routes. (B) Kaplan-Meier survival curve (n = 5 – 10). ID: intradermal; IM-EP: intramuscular electroporation; IT; intratracheal. *p < 0.05 and ***p < 0.0005 by log rank test with Gehan-Breslow-Wilcoxon test.

3. Discussion

In this study, we demonstrate that DNA-loaded mucus-penetrating particles (i.e. MPP) readily reach and internalize into pulmonary DC following intratracheal administration to a markedly greater extent compared to otherwise identical mucus-impermeable conventional particles (i.e. CP). Based on this finding, we engineered MPP carrying both antigen-expressing plasmids and nucleic acid-based adjuvants (i.e., CpG/pOVA-MPP) and confirmed that the formulation mediated significantly stronger and more durable adaptive immunity in the lung and other remote mucosal surfaces compared to conventional systemic DNA vaccinations. In addition, inhaled MPP-mediated vaccination unexpectedly provided greater systemic immunity than dose-matched antigen-expressing plasmids and adjuvants co-administered via routes commonly applied for systemic vaccination. The findings here underscore the critical role of improving the access to DC, in addition to more widely explored strategies to augment DC uptake, on achieving robust mucosal and potentially systemic immunity.

Our experimental data suggest that the uniquely high mobility in mucus facilitates cross-sectional penetration and lateral spread of MPP through the airway mucus gel layer, thereby enhancing the probability of particle encounter and uptake by pulmonary DC. While inhaled CP were found clumped up sparsely and superficially at the very lumen of the gel layer, identically administered MPP exhibited uniform airway distribution near to the epithelial surface. We have previously demonstrated that particles as large as 200 nm in diameters (i.e. larger than the particle diameters of both CP and MPP in this study) possessing muco-inert surfaces are capable of efficiently penetrating human and mouse airway mucus.^{[28, 29]} Thus, the favorable in vivo behavior of MPP is most likely attributed to the ability to resist adhesive interactions with mucus mesh and aggregation in the physiological lung environment. Of note, it has been previously reported that a similar surface shielding improves colloidal stability of nanoparticles, thereby facilitating their penetration through rat nasal mucosa.^[30] The enhanced MPP uptake by pulmonary DC is also likely due to their ability to circumvent the MCC by rapid and timely mucus penetration.^{[9, 16]} We previously demonstrated that at least 80% of the initial dose of MPP designed for inhaled delivery of small molecule drugs^[28] and nucleic acids^[13] were retained in the lung 2 hours after the administration, in sharp contrast to at most of 30% for otherwise identical mucus-impermeable counterparts. Accordingly, inhaled MPP carrying DNA vaccine components were found efficiently taken up by pulmonary DC resided in the lung interstitium, trafficked to the local lymph node and subsequently elicited strong pulmonary immunity in the present study. The intrinsic ability to mediate DC uptake (i.e. in vitro DC uptake) alone appeared insufficient in promoting particle uptake by pulmonary DC in vivo, as evidenced by the negligible DC uptake observed with inhaled CP despite its superior in vitro DC uptake capacity compared to MPP.

We show here that our pulmonary vaccination approach mediates long-lasting CTL and memory T-cell responses in an antigen-specific manner both locally in the lung and systemically. Importantly, T_EM-biased response mediated by inhaled MPP is particularly pronounced in the lung, whereas T_CM-biased response is generally observed with conventional vaccination (i.e. electroporation).^[31] T_CM must undergo multiple steps upon encountering pathogens, including activation, expansion, differentiation and trafficking, to initiate significantly-delayed effector responses.^[32] In contrast, T_EM in the lung, readily available by the MPP-mediated DNA vaccination, immediately acts on and rapidly removes respiratory pathogens,^[33] thereby efficiently preventing their replication at the early stage of infection.^[34] The observation is likely due to the ability of MPP to enhance the DNA vaccine uptake by pulmonary DC, which leads to more profound and durable antigen presentation to T-cells.^[35] The finding here is also in line with an evidence provided by a previous study demonstrating a positive correlation between the ratio of antigen-activated DC to T-cell and the magnitude of T_EM-biased response.^[36] Albeit to a lesser extent, inhaled MPP induced T_EM-biased response in the spleen as well, potentially providing a means to elicit fast-acting systemic immunity.

As a proof-of-concept, we demonstrated here that inhaled MPP provided a significantly prolonged survival of mice bearing an orthotopic lung cancer compared to gold-standard immunization approaches applied in the clinic, presumably due to the markedly enhanced CTL response. Nevertheless, we could not entirely prevent the death of tumor-bearing mice despite the strong antigen-specific effector activity and long-term memory T-cell response established in the lung by the inhaled MPP. One possible explanation may be the particularly aggressive and poorly immunogenic nature of LLC cells. It has been reported that immunostimulatory gene expression is significantly suppressed in an LLC-based mouse model compared to those established with other widely used cancer cell lines.^[27] In parallel, we have previously demonstrated that downregulation of immunosuppressive genes correlates with improved overall survival of mice bearing LLC-based tumors.^[37] Thus, it is conceivable that MPP-based inhaled DNA vaccination, in conjunction with other immunomodulatory strategy, may further enhance the therapeutic anti-cancer efficacy.

We demonstrated that pulmonary DNA vaccination by MPP led to robust trans-mucosal antigen-specific CTL responses both in GI and vaginal tracts, consistent with prior observation with pulmonary DNA or subunit vaccination.^{[38, 39]} This is most likely attributed to the ability of our vaccination approach to promote crosstalk between the lung and distal mucosal surfaces, as evidenced by the elevation of CD8⁺ T-cells expressing a gut-homing integrin in the local LN of the lung (i.e. mediastinal LN).^[38] We note that carrier-free DNA vaccination given intratracheally (i.e. IT CpG/OVA) was unable to induce antigen-specific CTL responses in remote mucosal surfaces, highlighting the key contribution of MPP formulation on establishing trans-mucosal immunity.

Efficient pulmonary vaccination by locally-administered MPP was somewhat alluded by previous reports demonstrating greater antigen-specific immunity by inhaled over standard systemic vaccination approaches.^{[39, 40]} However, enhanced systemic immunity by inhaled MPP was not expected a priori. It has been shown that inhaled nanoparticle-based DNA vaccination is capable of inducing a robust systemic immunity but to a level on par to that achieved by intramuscular immunization.^[39] Although it is difficult to make a direct comparison as pulmonary DC uptake is not quantified in this prior study, we speculate that the high in vivo DC uptake of MPP (i.e. ~25% of overall pulmonary DC) observed in our study may partially account for the discrepancy. We also note that number of DC accessible to DNA vaccine components administered via conventional intradermal and intramuscular routes would be limited, potentially leading to suboptimal systemic immunity.

4. Materials and Methods

4.1. Polymer synthesis

We synthesized PBAE polymers using the method that we previously detailed.^[15] Briefly, a two-step Michael addition reaction was used to synthesize non-PEGylated PBAE polymers. First, 4-amino-1 butanol and 1,4-butanediol diacrylate were reacted at a molar ratio of 1.1:1 molar at 90 °C for 16 hours in tetrahydrofuran (THF; 500 mg mL⁻¹) to yield acrylate-terminated PBAE polymers possessing molecular weight (MW) of 6.0 ± 0.2 kDa. The synthesized polymers were purified by washing three times with an excess of cold ether and dried under vacuum without exposure to light for 14 days to remove residual ether. The acrylate-terminated PBAE polymers were then reacted with 30-molar equivalents of 2-(3-aminopropylamino) ethanol in THF (100 mg mL⁻¹) at room temperature for 6 hours, followed by the purification and solvent-removal steps.

For preparing PEG-PBAE polymers, non-PEGylated PBAE polymers were first synthesized using the aforementioned method with some modifications. Briefly, 4-amino-1 butanol and 1,4-butanediol diacrylate were reacted at a molar ratio of 1.2:1 in THF (500 mg mL⁻¹) to yield acrylate-terminated PBAE polymers possessing MW of 4.0 ± 0.2 kDa and subsequently reacted with 30 molar equivalents of 1,3-diaminopropane. These intermediate polymers were extensive washed and dried as described above after each reaction step was completed. The polymers were then reacted with 2.05 molar equivalents of methoxy-PEG-succinimidyl succinate (JenKem) in THF (350 mg mL⁻¹) overnight at room temperature, followed by purification and solvent-removal steps, to yield the final product of PEG-PBAE polymers. Both non-PEGylated PBAE and PEG-PBAE polymers were dissolve in dimethyl sulfoxide (100 mg mL⁻¹) and stored at −20 °C for future use.

4.2. ¹H nuclear magnetic resonance (NMR) spectroscopy

Polymers were characterized by NMR as previously described.^[15] Briefly, ¹H NMR spectra of non-PEGylated PBAE and PEG-PBAE dissolved in deuterated methanol (MeOH-d4; Cambridge Isotope Laboratories) were recorded on a Bruker spectrometer (500 MHz). ¹H chemical shifts were reported in ppm (δ) and the MeOH peak was used as an internal standard. Data were processed using iNMR software.

4.3. DNA-loaded nanoparticle formulation & characterization

The OVA-expressing plasmid used in this study (pCI-neo-sOVA) was a gift from Dr. Maria Castro (Addgene plasmid # 25098;^[41]). The nucleic acid-based adjuvants, including CpG and poly(I:C), were purchased from InvivoGen. For microscopic and flow cytometric analysis, plasmid DNA was fluorescently labeled with the either Cy5 or MFP488 fluorophores using the Mirus Label IT tracker intracellular nucleic acid localization kit (Mirus Bio) according to the manufacturer’s instruction.

We formulated DNA-loaded nanoparticles using our previous established method^[15] with some modifications. The polymer solution was prepared with PBAE only or a mixture of PBAE and PBAE-PEG (at a wt/wt ratio of 2:3 based on PBAE mass) for CP or MPP, respectively. To engineer DNA-loaded nanoparticles, five volumes of nucleic acids, including labeled or unlabeled plasmid DNA with or without either CpG or poly(I:C), at 0.1 mg mL⁻¹, were added dropwise to one volume of a polymer solution at a PBAE-to-nucleic acid wt/wt ratio of 60:1 while vortexing. DNA-loaded nanoparticles were then washed with five volumes of ultrapure distilled water at 950 ×g for 8 minutes each time and concentrated to 0.5 mg mL⁻¹ using Amicon Ultra Centrifugal Filters (100,000 molecular-weight cutoff; Millipore).

For the nanoparticle characterization, hydrodynamic diameters and polydispersity index were measured in ultrapure water by dynamic light scattering and ζ-potential was measured in 10 mM NaCl at pH 7.0 by laser doppler anemometry, using a Zetasizer Nano ZS90 (Malvern Instruments). Transmission electron microscopy (H7600; Hitachi High Technologies America) was conducted to determine the morphology and geometric dimension of DNA-loaded nanoparticles.

To evaluate the ability of MPP to protect nucleic acid payloads, MPP containing 1 μg of pOVA and 0.25 μg of either CpG or poly I:C were treated with 2.5 unit of DNase I (Thermo Fisher Scientific) at 37°C for 15 minutes. Same amounts of carrier-free pOVA with either CpG or poly I:C were used as controls. Samples were then treated with 365 μg EDTA (Sigma) and further incubated at 65°C for 10 minutes. To induce de-compaction of the MPP, samples were incubated with heparin (Sigma Aldrich) at a 3:1 (w/w) ratio of heparin to DNA at room temperature for 10 minutes. Samples and 1 kb plus DNA ladder (Thermo Fisher Scientific) were then loaded into a 0.9% agarose gels containing SYBR Safe (Thermo Fisher Scientific), and electrophoresis was conducted sequentially at 50 and 100 V for 10 and 35 minutes, respectively. Finally, Gels were imaged using a ChemiDoc imaging system (Bio-RAD).

4.4. Single cell suspension preparation & flow cytometry

Pulmonary immune cells were collected by finely chopping harvested lung tissues, followed by digestion in a media containing 5 mg collagenase D (Worthington) and 1.25 mg of DNase I (Worthington) at 225 rpm in a shaker at 37°C for 40 minutes. Vaginal immune cells were isolated in a similar manner, but digested in a media containing 4 mg collagenase (Sigma) and 1.25 mg of DNase I (Worthington) for 2 hours. In parallel, immune cells from spleen, Peyer’s patches and different LN, including inguinal, mediastinal and mesenteric LN, were isolated by mechanically disrupted respective tissues. Cells were then passed through 70 and 40 μm cell strainers sequentially. Red blood cells from lung and spleen were removed using ammonium-chloride-potassium lysing buffer (Thermo Fisher Scientific) according to the manufacturer’s instruction. Tissue incubation and washing were done with RPMI 1640 medium (Thermo Fisher Scientific) supplemented with 10% heat-inactivated fetal bovine serum (HI-FBS; Thermo Fisher Scientific).

For flow cytometric experiments, cells were first stained with the LIVE/DEAD fixable dead cell stains kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. Cells were then incubated with purified anti-mouse CD16/CD32 antibody (Biolegend) on ice for 10 minutes to block Fc receptors. For the surface staining, cells were incubated with the antibody combination in eBioscience flow cytometry staining buffer (Thermo Fisher Scientific) at 4°C for 30 minutes. To stain intracellular markers, intracellular fixation & permeabilization buffer set (Thermo Fisher Scientific) was used according to manufacturer’s instruction. Antibodies against mouse CD45, CD8, CD3, CD4, CD11c, CD86, IFN-γ, CD11b, CD170, MHC I, MHC II and integrin α₄β₇ were purchased from Thermo Fisher Scientific. The R-phycoerythrin labeled SIINFEKL-MHC I pentamer was purchased from ProImmmune. All the flow cytometric experiments were done using SONY SH800S Cell Sorter and analyzed with the FlowJo Software (FlowJo).

4.5. Cell culture

Bone marrow-derived dendritic cells (JAWSII) cells were purchased from ATCC and SIINFEKL-expressing LLC cells were kindly provided by Dr. Amer A. Beg (Moffitt Cancer Center). JAWSII cells were maintained in Alpha Minimum Essential Medium (MEM) supplemented with ribonucleosides, deoxyribonucleosides, 4 mM L-glutamine (Thermo Fisher Scientific), 1 mM sodium pyruvate, 5 ng mL⁻¹murine granulocyte-macrophage colony-stimulating factor (Thermo Fisher Scientific), 20% HI-FBS (Thermo Fisher Scientific) and 1% penicillin/streptomycin. SIINFEKL-expressing LLC cells were cultured in MEM supplemented with 10% HI-FBS and 1% penicillin/streptomycin.

4.6. In vitro dendritic cell uptake and activation studies

JAWSII cells were plated on a 12-well plate at 200,000 cells/well for both studies. To evaluate DC uptake, the cells were treated with either MPP or CP carrying 0.5 μg of Cy5-labeled pOVA. To examine in vitro activation of DC, cells were treated with MPP carrying pOVA with or without CpG or poly I:C at a plasmid-to-adjuvant wt/wt ratio of 5:1. Cells in controls groups were treated with PBS or carrier-free nucleic acids at a same dose. For both studies, cells were incubated at 37°C for 6 hours prior to flow cytometric analysis.

4.7. Multiple particle tracking

Airway mucus samples were collected from patients visiting the Adult Cystic Fibrosis Center at Johns Hopkins University via spontaneous expectoration under a written informed consent in accordance with the Johns Hopkins Institutional Review Board. The motions of DNA-loaded nanoparticles carrying Cy5-labled plasmids in the freshly collected mucus samples were captured by high-resolution fluorescent video microscopy and quantified by MPT analysis using a software custom-written in MATLAB (MathWorks), as we have previously reported.^[20]

4.8. Animal Studies

All animals were handled in accordance with the policies and guidelines of the Johns Hopkins University Animal Care and Use Committee. Female C57BL/6 mice (6 – 8 weeks old; Charles River) were anesthetized with an intraperitoneal injection of 2,2,2-tribromoethanol (Sigma-Aldrich) or isoflurane. To evaluate in vivo performances of MPP (e.g. IT CpG/pOVA-MPP group), we treated mice with a single intratracheal dose of MPP carrying 20 μg of either fluorescently labeled or unlabeled pOVA with or without 4 μg of CpG at a volume of 50 μL via a microsprayer (MicroSprayer Aerosolizer Model IA-1C; Penn-Century). For comparison, mice in IT CpG/pOVA group were identically treated with carrier-free nucleic acids at the same dose and volume. In parallel, mice in ID and IM-EP CpG/pOVA groups received a mixture of 20 μg pOVA and 4 μg CpG at a volume of 24 μL in footpad and quadriceps, respectively, using a 30G needle attached to a gas-tight syringe (Hamilton). For the IM-EP CpG/pOVA group, an additional procedure of electroporation (Harvard Apparatus ECM830) was conducted at 2 × 60 ms pulses and 200 V cm⁻¹.

To evaluate the uptake of DNA-loaded nanoparticles by pulmonary DC in vivo, mice were intratracheally treated with either of CP or MPP carrying MFP488-labeled pOVA and sacrificed 16 hours after the administration to determine the uptake using flow cytometric analysis. For microscopic observation of DNA-loaded nanoparticles, lung tissues were harvested different time points after the intratracheal administration, embedded in optimum cutting temperature compound (Finetek) solution, cryosectioned using a CM1950 cryostat (Leica Biosystems) and imaged using a confocal LSM 710 microscope (Carl Zeiss). Specifically, we determine the distribution of DNA-loaded nanoparticles in tracheal lumen and lung airway interstitium 1 hour after mice received either of CP or MPP carrying Cy5-labeled pOVA with and without CpG, respectively. In parallel, trafficking of DNA-loaded nanoparticles to the mediastinal LN and lung was evaluated 48 hours after treatment with MPP carrying Cy5-labeled pOVA and CpG. Cells were labeled or stained using antibodies against CD11c (abcam) and/or CD170 (abcam), Alexa Fluor 568 Goat Anti-Armenian hamster IgG secondary antibody (abcam), Alexa Fluor 488 Goat Anti-Rabbit IgG secondary antibody (abcam) and 4′,6-diamidino-2-phenylindole (DAPI). Of note, microscopic settings were carefully adjusted to avoid introduction of any background fluorescence using lung tissues sections from untreated mice.

To evaluate the immune response and memory establishment, mice were immunized and boosted at day 0 and 14, respectively. Mice were sacrificed at day 21 and 70 for assessing immune response and memory establishment, respectively, and relevant tissues were harvested and processed for flow cytometric analysis. For the adoptive transfer study, we harvested spleen from OT-I mice (6–8 weeks; Jackson laboratory) for CD8⁺ T-cell isolation and subsequently enrichment using a negative selection EasySep mouse t-cell isolation kit (STEMCELL Technologies). We then treated C57BL/6 mice with 1 × 10⁶ splenic CD8⁺ T cells from OT-I mice via a tail vein injection one day prior to the immunization and sacrificed them three days after the immunization for subsequent flow cytometric analysis.

For the proof-of-concept anti-cancer efficacy study, we established an orthotopic mouse model of lung cancer by intratracheal inoculation of C57BL/6 mice with 1 × 10⁶ LLC cells expressing SIINFEKL 7 days after the boost immunization and monitored survival of mice over time.

4.9. Ex vivo re-stimulation study

We evaluated the immunological responsiveness of T-cells from different organs by quantifying the intracellular IFN-γ production in the CD8⁺ and CD4⁺ T-cells isolated from the immunized mice using a previously reported method.^[24] Briefly, we prepared single cell suspensions from the lung, spleen and LN harvested 7 days after the boost immunization, using the procedure described above. Cells were cultured in Iscove modified Dulbecco medium media (Thermo Fisher Scientific) supplemented with 10% HI-FBS, 2 mM L-glutamine and 1% penicillin/ streptomycin at 37 °C. Cells from spleen and LN were plated on a 6-well plate at a 7 × 10⁶ cells/well and cultured in presence of monensin (Sigma) and 1 μg mL⁻¹ of SIINFEKL (InvivoGen) or 100 μg mL⁻¹ OVA (InvivoGen) for 6 hours to re-stimulate OVA-specific CD8⁺ or CD4⁺ T-cells, respectively. In parallel, cells from the lung were plated in the same manner and re-stimulated with 1 μg mL⁻¹ ionomycin and 50 ng mL⁻¹ PMA (InvivoGen) for 4 hours. Cells from all three organs were additionally treated with 5 μg mL⁻¹ brefeldin A (Thermo Fisher Scientific) for 2 and 3 hours to inhibit secretion of newly produced IFN-γ from CD8⁺ and CD4⁺ T-cells, respectively, prior to the cell collection. Cells were then stained for flow cytometric analysis.

4.10. Statistical Analysis

All the data presented did not go through pre-processing. Statistical analyses of two or multiple comparisons were conducted using two-sided Student’s t-test or one-way analysis of variance (ANOVA) with appropriate post-hoc analyses, respectively, and survival of animals was compared using the log rank test in Graphpad Prism 7. Differences were considered to be statistically significant at a level of p < 0.05.