Comparative Study of Acetalated-Dextran Microparticle Fabrication Methods for a Clinically Translatable Subunit-based Influenza Vaccine

Abstract

The most common influenza vaccines are inactivated viruses produced in chicken eggs, which is a time-consuming production method with variable efficacy due to mismatches of the vaccine strains to the dominant circulating strains. Subunit-based vaccines provide faster production times in comparison to the traditional egg-produced vaccines but often require the use of an adjuvant to elicit a highly protective immune response. However, the current FDA approved adjuvant for influenza vaccines (MF59) elicits a primarily helper T-cell type 2 (Th2)-biased humoral immune response. Adjuvants that can stimulate a Th1 cellular response are correlated to have more robust protection against influenza. The cyclic dinucleotide cGAMP has been shown to provide a potent Th1 response but requires the use of a delivery vehicle to best initiate its signalling pathway in the cytosol. Herein, acetalated dextran (Ace-DEX) was used as the polymer to fabricate microparticles (MPs) via double-emulsion, electrospray, and spray drying methods to encapsulate cGAMP. This study compared each fabrication method’s ability to encapsulate and retain the hydrophilic adjuvant cGAMP. We compared their therapeutic efficacy to Addavax, an MF59-like adjuvant, and cGAMP Ace-DEX MPs provided a stronger Th1 response in vaccinated BALB/c mice. Furthermore, we compared Ace-DEX MPs to spray dried MPs composed from a commonly used polymer for drug delivery, poly(lactic-co-glycolic acid) (PLGA). We observed that all Ace-DEX MPs elicited similar humoral and cellular responses to the PLGA MPs. Overall, the results shown here indicate Ace-DEX can perform similarly to PLGA as a polymer for drug delivery and that spray drying can provide an efficient way to produce MPs to encapsulate cGAMP and stimulate the immune system.

Graphical Abstract

1. Introduction

The CDC estimates that in the United States, the influenza virus can cause up to 45 million infections that lead to as many as 810,000 hospitalizations and 61,000 deaths per year (CDC, 2022a). This results in an estimated $11.2 billion economic burden each year (Anupindi, 2022; de Courville et al., 2022). Current vaccine formulations primarily consist of whole, attenuated or inactivated viruses that are manufactured in chicken eggs (CDC, 2022b). This process can take 5 to 6 months to produce a vaccine before distribution (Brauer and Chen, 2015; Organization, 2009; Osterholm et al., 2012; Raymond et al., 2016; Yang, 2013). Furthermore, these vaccines have low and varied efficacy each year which can result from inaccurate predictions of the circulating strain, mutation of the virus midseason, or egg-adapted antigenic changes. This process does not allow for a quick response time for when a new vaccine is needed to counter the circulating strain necessitating the need to explore new approaches to manufacture a vaccine that can be produced more rapidly.

Subunit-based vaccines (such as Flublok^®) offer a much faster production time (< 2 months) compared to egg-based manufacturing methods (Burnett and Burnett, 2020; Margolin et al., 2020). One of the most effective protein antigens for influenza is hemagglutinin (HA). HA is a surface glycoprotein responsible for viral docking to a cell and the primary protein the immune system mounts a defense towards (immunodominant antigen), making it an excellent vaccine antigen (Hendy et al., 2022; Veljkovic et al., 2014). However, one drawback with using protein antigens, like HA, is that they have inherently low immunogenicity when delivered alone. This requires either a high dose of the antigen per vaccination (i.e. FluBlok^®) or the use of an adjuvant to stimulate the immune system. The only FDA approved adjuvant used for an influenza vaccine is MF59, which has a helper T-cell type 2 (Th2)-biased humoral response and fails to induce a strong Th1 cellular response (Ko and Kang, 2018). Adjuvants that can elicit a Th1 response can better provide robust protection against viruses like influenza (Eichinger et al., 2020; Huber et al., 2006).

An adjuvant that can provide a Th1 response is the cyclic dinucleotide 3’3’ cyclic guanosine monophosphate-adenosine monophosphate (cGAMP). cGAMP activates the cytosolic stimulator of interferon genes (STING) pathway to produce pro-inflammatory type I interferon cytokines (Chen et al., 2018a; Gallovic et al., 2022; Junkins et al., 2018). A drawback with cGAMP is that it is a hydrophilic and charged molecule that does not readily permeate through the cellular membrane. To overcome this limitation, cGAMP can be delivered to cells by using a polymeric microparticle (MP) vehicle. Previous studies have shown that polymeric MPs of about 0.5–2 μm allow for the passive targeting of antigen presenting cells (APCs) to deliver cGAMP via phagocytosis, leading to the generation of antigen-specific antibody titers and elevated cellular responses in vivo (Chen et al., 2018a; Gallovic et al., 2022; Junkins et al., 2018).

In this study, cGAMP was encapsulated in a polymeric MPs composed of the biopolymer acetalated dextran (Ace-DEX). With a one-step reaction, water-soluble dextran can be converted into a hydrophobic Ace-DEX polymer by modifying the hydroxyl groups on the sugar backbone into acyclic or cyclic acetal groups (Kauffman et al., 2012). Upon degradation, Ace-DEX hydrolyzes back into dextran along with very low amounts of acetone and ethanol which are pH neutral by-products. Ace-DEX is acid sensitive, allowing MPs to degrade slowly and retain cGAMP at physiological pHs and to rapidly degrade and release the adjuvant when exposed to a phagosome’s acidic environment. While several publications have detailed applications of Ace-DEX particles as a drug delivery platform, more work is needed to study different fabrication methods using this polymer that are continuous and can be easily scaled. Behnke et al. compared different methods of fabrication Ace-DEX nanoparticles (NPs) to allow for feasible translation into the clinic (Behnke et al., 2023). They compared NPs fabricated by microfluidic-assisted nanoprecipitation, batch nanoprecipitation, and emulsion-evaporation and found that Ace-DEX provides a promising material for NP fabrication that can be scaled up and follow GMP standards when encapsulating the leukotriene biosynthesis inhibitor BRP-187.

Here, we focus on three fabrication methods: double emulsion, electrospray, and spray drying. Double emulsion uses a homogenizer to emulsify the polymer-drug solution to create MPs (Figure 1A). A major limitation to this method is that it is not a single step or continuous process which makes it difficult to scale up. The ideal MP fabrication method would be highly scalable, continuous, and reproducible. Electro-hydrodynamic spraying (E-spray, electrospraying) offers a one-step and continuous process to form MPs by applying a voltage to a polymer-drug solution that is being ejected from a capillary inducing atomization of the liquid to form MPs that are collected on a plate (Figure 1B) (Almería et al., 2011; Batty et al., 2022; Gallovic et al., 2016). While scale up processes are an option commercially for electrospraying, the method is heavily dependent on a slow flow rate of the feed so that the voltage can atomize the solution, thus complicating the scale up process. While electrospray is commonly performed at ambient temperatures, spray drying atomizes a polymer-drug solution with heated air (Figure 1C). This method is a simple, one-step and continuous process that is highly scalable and is already heavily used in the food and pharmaceutical industry, making the translation into the clinic feasible (Piñón-Balderrama et al., 2020). The drawback of spray drying is that it can induce thermal stress, but by tuning the multiple parameters of the process, this limitation can be minimized.

Figure 1.

Schematic representation for the fabrication process of (A) double emulsion, (B) electrospray, and (C) spray drying methods.

Ace-DEX MPs have previously been reported to encapsulate cGAMP by emulsion (Chen et al., 2016; Chen et al., 2018a) and electrospray (Batty et al., 2022; Gallovic et al., 2022). Emulsion Ace-DEX cGAMP MPs have been shown to elicit both humoral and cellular specific responses in mice towards the antigen used in the vaccination (Chen et al., 2018a). Electrospray Ace-DEX cGAMP MPs have been used to vaccinate ferrets and mice and it was found that cGAMP MP group outperformed the soluble cGAMP group for generating neutralizing titers (p<0.01) and it was observed that the viral load in the nasal lavage had significantly decreased (p<0.005) when compared to the emulsion adjuvant Addavax (MF59-like adjuvant) (Gallovic et al., 2022; Junkins et al., 2018). However, different fabrication methods can affect the encapsulation of the cargo and ultimately the therapeutic efficacy when used for a vaccine. For example, Junkins et al. encapsulated cGAMP in Ace-DEX by emulsion and electrospray methods (Junkins et al., 2018). They found that electrospraying could encapsulate cGAMP more efficiently than by emulsion (90% vs 41% encapsulation efficiency (EE), respectively). Further, this difference was polymer dependent since poly(lactic-co-glycolic acid) (PLGA) electrosprayed MPs had lower EE in both electrospray (70%) and emulsion (0%) compared to Ace-DEX counterparts, likely due to the hydrophilic nature of cGAMP and the potential for hydrophilic pockets in Ace-DEX where the sugar backbone has unreacted hydroxyl groups. When Ace-DEX cGAMP MPs were incubated with dendritic cells in vitro, significantly more IFN-β was produced with electrospray MPs than with emulsion MPs (p<0.01). It was also found that dendritic cells incubated with electrospray Ace-DEX MPs produced significantly more IFN-β (p<0.0001) and IL-6 (p<0.05) than electrospray PLGA MPs (Junkins et al., 2018).

The study reported here is focused on encapsulating cGAMP in Ace-DEX and PLGA MPs via spray drying and comparing their therapeutic efficacy to Ace-DEX MPs fabricated from emulsion and electrospray methods. Additionally, the robustness of spray drying was assessed by manufacturing Ace-DEX MPs with different system parameters, and determining how these parameters affect the immunotherapeutic efficacy when used as an influenza subunit-based vaccine by vaccinating with HA and cGAMP MPs. Overall, the aim of the work here was to develop a MP cGAMP delivery vehicle that can be easily manufactured for feasible clinical translatability.

2. Materials and Methods

2.1. Materials

All chemicals were purchased through MilliporeSigma (St. Louis, MO) unless stated otherwise. All assays, biologics, and disposables were purchased through Thermo Fisher Scientific, (Waltham, MA) unless stated otherwise.

2.2. Acetalated Dextran (Ace-DEX) Polymer Synthesis

Ace-DEX was synthesized by first dissolving 71k dextran from Leuconostoc mesenteroides in dimethyl sulfoxide (DMSO) (Kauffman et al., 2012). With the addition of 2-ethoxypropene (Matrix Scientific, Columbia, SC) and the catalyst pyridinium p-toluenesulfonate, the reaction to produce Ace-DEX began in anhydrous conditions. After two hours, the reaction was stopped by quenching the mixture with triethylamine (TEA). To precipitate Ace-DEX, the mixture was added drop-wise to basic water (0.04% v/v TEA in water), isolated by centrifugation, and lyophilized for one day. For further purification, ethanol was used to dissolve the solid, centrifuged to remove ethanol-insoluble impurities, precipitated in basic water, and lyophilized for two days. To characterize the polymer, a ¹H 400 MHz NMR (Inova) was used to determine the cyclic acetal coverage (CAC) of the Ace-DEX (60% CAC).

2.3. Microparticle Fabrication

Emulsion MPs were produced using a double-emulsion water/oil/water (w/o/w) solvent evaporation method (Chen et al., 2018a). Ace-DEX, dissolved in ethyl acetate, was mixed with phosphate buffer saline (PBS) solution containing 3’-3’ cGAMP (Invivogen, San Diego, CA). The mixture was homogenized at 18,000 RPM for 30 seconds (IKA T25 Digital Ultra-Turrax, Cole Parmer, Vernon Hills, IL). A 3% w/v polyvinyl alcohol (PVA) solution in PBS was added and homogenized again for 30 seconds. The mixture was then added to 0.3% PVA solution and stirred for 3 hours. To purify the MPs, the mixture was centrifuged and resuspended in basic water twice and then lyophilized for two days.

Electrospray MPs were produced using a monoaxial electrospray method (Batty et al., 2023; Watkins-Schulz et al., 2019). Ace-DEX was dissolved in a 90:10 ethanol:water solution with cGAMP. A single 20G blunt stainless-steel needle (Hamilton Company, Reno, NV) and 5 mL syringe containing the Ace-DEX and cGAMP solution was set on a syringe pump (Instech Pump 11 Elite, Harvard Apparatus, Holliston, MA). A stainless-steel plate (McMaster Carr, Elmhurst, IL) was placed 15 cm under the needle to collect the ejected MPs. The positive polarity from a voltage power source (Gamma High Voltage Research, Inc., Ormond Beach, FL) was connected to the corner of the steel plate (2.5 kV) while the negative polarity was attached to the base of the needle (−7 kV). The pump then ejected the Ace-DEX and cGAMP solution at 0.2 mL/hr. The dried MPs were removed from the steel plate and lyophilized for one day.

Spray dried MPs were fabricated using the Mini Spray Dryer B-290 (Büchi, New Castle, DE). Ace-DEX was dissolved in a 90:10 ethanol:water solvent system with cGAMP. Five sets of Ace-DEX MPs were spray dried as listed in Table 1. Baseline MPs were spray dried with parameters that were obtained based on a design of experiments to optimize MP yield. The four other MP batches alter one parameter with respect to the Baseline MPs spraying parameters to determine how each parameter may affect the MPs. Concentration MPs had an increase in polymer feed solution concentration from 2.5 to 25 mg/mL; temperature MPs had an increase in the inlet drying temperature setting from 75 to 100°C; Inlet flowrate MPs had an increase in the feed solutions flowrate from 30 to 55% (% is based on the spray dryer’s internal settings); and Q-flow MPs had a decrease in the amount of nitrogen being pumped into the system to atomize the feed solution from 60 to 40 mm (mm is based on the spray dryers internal settings). PLGA (75:25 lactic:glycolic acid ratio, non-acid terminated) (Polyscience, Niles, IL) was used in this study as a polymer comparative to Ace-DEX. Due to the high hydrophobicity of PLGA, a 90:10 acetone:water solvent system was used instead. The spray drying parameters used for PLGA are listed in Table 1 was based on a design of experiments to optimize MP yield.

Table 1.

Spray dry parameters for the different particle formulations. Bolded values indicate the changed parameter from the Baseline MPs formulation values. EtOH = ethanol; H2O = water.

Particle	Polymer Used	Solvent System	Polymer Concentration (mg/mL)	Temperature (°C)	Inlet Flowrate (%)	Q-Flow (mm)	Aspirator (%)
Baseline	Ace-DEX	90:10 EtOH:H2O	2.5	75	30	60	100
Concentration			25	75	30	60	100
Temperature			2.5	100	30	60	100
Inlet Flowrate			2.5	75	55	60	100
Q-Flow			2.5	75	30	40	100
PLGA	PLGA	90:10 Acetone:H2O	2.5	35	30	60	100

2.4. Microparticle Characterization

To characterize the diameter and polydispersity index (PI), MPs were suspended in basic water and analyzed by dynamic light scattering (Brookhaven NanoBrook 90Plus Zeta Particle Size Analyzer, Holtsville, NY). Scanning electron micrographs (SEM) were generated using a S-4700 Cold Cathode Field Emission Scanning Electron Microscope (Hitachi High-Technologies, Krefeld, Germany) where MPs were suspended in basic water and a droplet was placed onto an SEM stub with carbon tape to dry overnight. To ensure the MPs had low endotoxin levels, all MPs were tested using the Pierce LAL Chromagenic Endotoxin Quantification kit. All MPs were reported to be below 0.125 EU/mg which is below the limit for many preclinical formulations (Brito and Singh, 2011). High-performance liquid chromatography (HPLC) (Agilent 1100 series, Santa Clara, CA) was used to determine the cGAMP content of the MPs. MPs were dissolved in a 75:25 water:acetonitrile with 0.1% v/v trifluoroacetic acid (TFA) solvent system. The sample was pumped into an Aquasil C18 column at 1 mL/min and cGAMP was detected at 210 nm.

2.5. cGAMP Release Kinetics

For cGAMP release kinetics and particle degradation, MPs were suspended in either phosphate buffered saline (PBS, pH 7.4) or 0.3 M sodium acetate buffer (pH 5) at 1 mg/mL. Suspensions were placed in a 5 mL Eppendorf tube and placed on a shaker plate set to 37°C to mimic physiological conditions. An aliquot (300 μL) of the MP suspension was taken at each time. The aliquot was then centrifuged to pellet the MPs; the supernatant was removed, and the pellet was stored in a −20°C freezer. Once all the timepoints were collected, each pellet was dissolved with 300 μL of 75:25 water:acetonitrile with 0.1% v/v TFA solvent and the cGAMP content was measured by HPLC. For images of degraded MPs, the pellet was lyophilized overnight, suspended in basic water, and dried on an SEM stub for imaging the following day.

2.6. Dendritic Cell Stimulation

DC2.4 cells (MilliporeSigma, Burlington, MA) in Roswell Park Memorial Institute (RPMI) 1640 media (supplemented with 10% v/v heat-inactivated fetal bovine serum and 1% v/v penicillin-streptomycin) were seeded in a tissue culture treated 96-well plate at 25,000 cells/well and incubated overnight. MPs were suspended in media and added to cells providing a cGAMP dose of 0.1 μg/mL and incubated for 24 hours. The cells were then centrifuged at 500 × G for 5 minutes, the supernatant was taken to measure for cell cytotoxicity by LDH concentration and cytokines (IFN-β and TNF-α) by sandwich ELISA (Biolegend, San Diego, CA). The percent viability of the cells was measured using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay (Chen et al., 2016).

2.7. In Vivo Vaccination

Mouse studies followed the guidelines set by the National Institutes of Health and were approved by the Institutional Animal Care and Use Committee at UNC. Female BALB/c mice (n=5 per group) that were 6–8 weeks of age from Jackson Laboratories (Bar Harbor, ME) were vaccinated intramuscularly on days 0, 21, and 35. Both the right and left leg received 25 μL for a total of 50 μL per mouse. Mice were vaccinated at a dosage of 1 μg of recombinant protein hemagglutinin (HA) from the A/Puerto Rico/8/1934 (PR8) influenza virus (BEI Resources, Manassas, VA) and 1 μg of cGAMP in MPs suspended in PBS. Groups included PBS only, Addavax + HA, Electrosprayed MPs + HA, and spray dried MPs (Q-flow, Concentration, Temperature, and PLGA) + HA. Vaccinations with Addavax (Invivogen) followed the manufacturers protocol where the Addavax emulsion was mixed with 1 μg HA in an equal volume of PBS, incubated for 30 minutes at room temperature, and mice were injected with a total volume of 50 μL.

2.8. Humoral Response Characterization

Blood serum was collected from mice on days 14, 28, and 42 after vaccination via submandibular cheek bleeds. To measure for anti-HA antibodies, flat-bottom high-binding polystyrene 384-well plates (Greiner, Kremsmünster, Austria) were coated with 25 μL of HA in PBS at 0.1 μg/mL and stored at 4°C overnight. Wash buffer was 0.05% Tween 20 in PBS and blocking buffer was 50 μL of was buffer with 3% non-fat instant milk (Food Lion, Raleigh, NC). Secondary antibodies were Goat Anti-Mouse IgG Fc-HRP, Goat Anti-Mouse IgG2a-HRP, or Goat Anti-Mouse IgG1-HRP (Southern Biotech, Birmhingham, AL). Tetramethylbenzidine (TMB) one component substrate (Southern Biotech) was used as the developing solution. The absorbance was read at 450 and 570 nm. To analyze the data, the absorbance from 450 nm was subtracted from the background absorbance from 570 nm. Values were curve fitted using the built-in model analysis “log(inhibitor) vs. response – Variable slope (four parameters)” in Graphpad Prism 9. Titer values were interpolated by determining which dilution value intercepts the endpoint value (Frey et al., 1998).

For the antibody-dependent cellular phagocytosis (ADCP) assay, carboxylate-modified fluorescent microspheres (beads) were coated with HA using sulfo-n-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) (Butler et al., 2019; Earnest et al., 2019). The beads were then incubated with diluted sera for two hours. The sample solution was then added to J774a.1 murine macrophages (ATCC, Manassas, VA) that were plated overnight at 50,000 cells/well in a non-tissue culture treated 96-well plate for 1 hour. The cells were centrifuged then washed with 5 mM EDTA in PBS twice. Cells were then fixed with 3% paraformaldehyde and analyzed for microsphere uptake using an Attune NxT Flow Cytometer. The data was analyzed and reported as the phagocytic score which Is the percent of microsphere positive cells divided by ten thousand times the geometric mean fluorescence intensity in the particle channel.

For the antibody-dependent complement deposition (ADCD) assay, day 42 sera samples were heat inactivated at 56°C for 30 minutes then diluted (Boudreau et al., 2020). HA-conjugated microspheres were generated as detailed above and added to the sera solution and incubated at room temperature for two hours. Reconstituted lyophilized guinea pig complement (Cedarlane Labs, Burlington, ON, Canada) in barbitone buffer was added to the bead/sera samples and incubated at 37°C for 20 minutes. Samples were centrifuged and washed with 15 mM EDTA in PBS twice. Beads were then stained with FITC anti-guinea pig C3 in PBS for 30 minutes. Beads were washed three times with 3% v/v FBS in PBS. Beads were analyzed using the Attune NxT Flow Cytometer. The data was analyzed and reported as the complement deposition score which is the percent of complement positive beads divided by ten thousand times the geometric mean fluorescence intensity in the complement channel.

2.9. Antigen Recall Experiments

On day 45, spleens and popliteal lymph nodes were harvested from vaccinated mice and placed in a 15 mL falcon tube with 3 mL of RPMI 1640 (supplemented with 10% v/v heat-inactivated fetal bovine serum and 1% v/v penicillin-streptomycin) (Batty et al., 2023). Spleens and lymph nodes were then processed into single-cell suspensions and seeded at 1,000,000 cells/well in a pre-coated 96-well PVDF filter plate for ELISpot or a tissue-cultured flat-bottom 96-well plate for ELISA. Cells were either stimulated with 10 μg/mL of HA PR8 peptide array (BEI Resources, Manassas, VA) in media or non-stimulated by dosing with media, respectively. ELISpots were developed following the manufacturer’s protocol (BD, Franklin Lakes, NJ). For the ELISA, the cells were pelleted and the supernatant was taken to measure for soluble cytokines following the manufacturer’s protocol (Invitrogen, San Diego, CA).

2.10. Flow Cytometry Characterization

Popliteal lymph nodes and spleens were isolated from vaccinated mice at day 45. With or without stimulation, single-cell suspensions were acquired and stained with fluorochrome-conjugated antibodies anti-mouse CD4, CD8, CD19, B220, CD138, IgM, GL7, CD95, PD-1, and CXCR5 antibodies (Biolegend). Briefly, cells were incubated with antibodies for 30 minutes at room temperature prior to washing and fixing with 4% formaldehyde for 10 minutes. For intracellular cytokine staining, isolated draining lymphocytes and splenocytes were stimulated with peptide for 36 hours. Phorbol myristate acetate (PMA), ionomycin, and Golgi inhibitor monesin were added for the last two hours of stimulation. After that, cells were stained with surface markers and a fixable viability dye (Tonbo Biosciences, San Diego, CA). Permeabilization buffer (Invitrogen) was used to wash, fix, and permeabilize samples, followed by staining with anti-mouse IFN-γ and TNF-α antibodies (Biolegend) for one hour. Stained samples were then run on a Thermo Fisher Attune NxT machine at UNC Flow Cytometry Core Facility and data was analyzed on Flowjo software.

2.11. Statistical Analysis

GraphPad Prism 10 was used to generate and format figures. The statistical analysis tools provided in the software were used to analyze data by an ANOVA analysis followed by a Tukey’s pairwise comparison.

3. Results and Discussion

3.1. Spray Dried Ace-DEX MPs Can Encapsulate cGAMP With Desirable Release Kinetics

To first characterize the MPs fabricated via double emulsion, electrospray, or spray drying, we quantified the EE of cGAMP loaded in each MP. MPs made by emulsion had significantly reduced EE compared to electrospraying and spray drying (Table 2) where emulsion MPs had a 63.9% EE whereas electrosprayed and spray dried MPs had over 100% EE (p<0.005). This reduced encapsulation with emulsion MPs is expected because the hydrophilic cGAMP can partition into the aqueous phase during fabrication and we have previously reported similar differences in encapsulation (Chen et al., 2018a; Junkins et al., 2018). In electrospray, satellite or small droplets can often be expelled through repulsion forces from the Taylor cone, resulting in some loss of yield (Sanders et al., 2023). We hypothesized that these satellite particles have reduced cGAMP loading, compared to the larger particles that deposit on the collection plate, thereby resulting in greater than 100% EE for the molecule. Similar to the electrosprayed MPs, spray dried MPs also resulted in over 100% EE. In spray drying, smaller particles are often recirculated more easily rather than depositing into the collection vessel from the cyclone, resulting in them likely being disproportionately trapped in the device filter and lost in the overall yield (Sosnik and Seremeta, 2015). We hypothesize that cGAMP-poor satellite particles are lost in the electrospray and spray drying process and we calculated EE based on actual loading determined analytically, an EE of >100% can be achieved which has been previously seen (Batty et al., 2022). While further investigation is needed to confirm the cause of the high EE, fabricating MPs via spraying methods results in more efficient encapsulation of cGAMP compared to an emulsion-based method.

Table 2.

MP characterization of each formulation showing the yield percentage when formulated by spray drying, diameter (nm), polydispersity index (PI), EE of cGAMP, and the cGAMP weight loading.

Particle	Fabrication Method	Polymer	Diameter (nm)	PI	Encapsulation Efficiency (%)	cGAMP Weight Loading (μg/mg particle)
Baseline	Spray dry	Ace-DEX	674.3 ± 16.9	0.290	112.0 ± 16.6	11.2 ± 1.7
Concentration			1,121.8 ± 57.4	0.328	110.5 ± 8.9	11.1 ± 0.9
Temperature			686.7 ± 15.7	0.258	105.8 ± 13.0	10.6 ± 1.3
Inlet Flowrate			640.0 ± 21.9	0.242	103.5 ± 2.6	10.3 ± 0.3
Q-Flow			731.4 ± 14.9	0.275	102.8 ± 10.9	10.3 ± 1.1
Electrospray	Electrospray		712.8 ± 107.7	0.291	116.6 ± 7.2	11.7 ± 0.7
Emulsion	Double Emulsion		308.0 ± 26.1	0.190	63.9 ± 1.9	6.4 ± 0.2
PLGA	Spray dry	PLGA	1,151.4 ± 178.5	0.275	114.4 ± 4.4	11.4 ± 0.4

Next, we sought to analyze and compare the morphology and size of each MP. Emulsion particles tended to have a lower polydispersity index and a smaller diameter due to the high and consistent shear stress being experienced on each particle (Figure 2A). Electrosprayed Ace-DEX MPs exhibited a more porous surface morphology (Figure 2B) which is consistent with previous findings (Batty et al., 2022). Spray drying Ace-DEX MPs and altering the parameters mostly showed consistent results in terms of their morphology and sizing (Figure 2C–G). Spray dried MPs exhibited a collapsed morphology with surface buckling. When the polymer concentration in solution was increased (concentration MPs) an increase in diameter (1,122 nm) and polydispersity (0.328) (Figure 2D) was observed. This increase in size is logical as increasing the concentration of the polymer in solution leads to more polymer occupying the volume of a single droplet when drying, resulting in larger particles. When the polymer solution feed flowrate was increased (Inlet flow MPs) (Figure 2F) a slight difference in surface morphology was seen, where each particle appeared more rounded rather than fully collapsed. Collapsed MPs indicate surface buckling due to formation of an outer-shell in the initial phases of drying followed by evaporation of the remaining solvent inside of the shell which leads to a collapsed surface (Vehring, 2008). We hypothesize that increasing the polymer feed solution flowrate can lead to less heat being transferred to each individual droplet, slowing down the drying process and allowing the Ace-DEX shell to solidify more fully, rather than collapsing upon itself. Lastly, PLGA spray dried MPs exhibited a different surface morphology (Figure 2H) compared to Ace-DEX spray dried MPs. This is attributable to the difference in polymer used along with the different parameters used to fabricate the MPs. Overall, changes in spray drying parameters resulted in comparable Ace-DEX MP morphology and size. This indicates that spray drying Ace-DEX MPs can tolerate a certain degree of change in parameters while still producing comparable MPs which is beneficial in scaling up the process to aid in translatability into the clinic.

Figure 2.

Scanning electron microscopy images of cGAMP encapsulated MPs. Ace-DEX MPs are shown for (A) Emulsion, (B) Electrospray, (C) Baseline, (D) Concentration, (E) Temperature, (F) Inlet flow, and (G) Q-flow. Lastly, (H) PLGA spray dried MPs are shown. Scale bar drawn is 5 μm.

To assess how each MP will retain cGAMP, release profiles were generated by suspending MPs in PBS and kept at 37°C to mimic physiological conditions. Emulsion MPs exhibited a burst release of 63% in the first 24 hours with minimal cGAMP being further released for 35 days (Figure 3A). This burst release has been previously seen with Ace-DEX emulsion MPs and is because the hydrophilic cGAMP partitions toward the aqueous interface of the continuous phase at the surface of the particle during the emulsion process, maintaining a high concentration of adjuvant towards the surface of the MPs (Chen et al., 2018a). The remaining cGAMP would not be released until the MPs degrade through surface erosion (Stiepel et al., 2022). Electrosprayed MPs have reduced burst compared to the emulsion MPs, with a 39% burst release followed by near first order release of the drug up to 14 days. In contrast to both these particle formulations, spray dried baseline MPs have a triphasic release profile where 10% of cGAMP releases in an early burst followed by a period of slow release of cGAMP until the release rate increases around day 21. With both spray dried and electrosprayed MP fabrication, the continuous phase is a gas (air) and not a liquid-based phase like emulsion fabrication. It is expected for these methods that the dispersity of cGAMP in the MP is more dependent on the drying process of the drug/polymer droplet, making it less likely to accumulate at the particle surface as with emulsion-based MPs (Tanhaei et al., 2021). The difference observed in release kinetics between spray dried versus electrosprayed MPs is likely due to the porous morphology of electrosprayed MPs resulting in an increased surface area. We hypothesize that the increase in surface area leads to faster degradation of the electrosprayed MP through greater surface interaction with the aqueous environment and therefore more hydrolytic degradation. The spray dried PLGA MP release profile showed minimal amounts of cGAMP released over the course of 35 days. This is due to the high molecular weight of PLGA used in this study, non-acid terminated ends, and the ratio of lactic acid to glycolic acid (75:25) making hydrolysis of the polymer slower (Makadia and Siegel, 2011). Based on the release profiles, spray dried MPs offer a decrease in burst release and increased cGAMP retention over time which is highly desirable as a higher portion of cGAMP can be delivered to target cells during vaccination.

Figure 3.

cGAMP release profiles from all MPs were assessed for each of the particle formulations by suspending each in PBS at 1 mg/mL and kept at 37°C on a heated shaker plate to mimic physiological conditions. (A) depicts the three different formulation methods for Ace-DEX MPs (electrospray, emulsion, and baseline spray dried MPs) along with PLGA spray dried MPs. (B) cGAMP release profile of spray dried MPs with each parameter changed. Data is shown as mean±SD. ANOVA statistical analysis was performed.

Since different fabrication methods lead to changes in release kinetics of cGAMP, we then analyzed the release profiles of the other batches of Ace-DEX spray dried MPs. The release profiles of the Ace-DEX spray dried MPs generated through a single change in a fabrication variable, exhibited very similar trends to one another (Figure 3B). The only significant difference seen in the release profile was from the concentration MPs, where a much slower release was observed. Due to the increase in size of the concentration MPs, there is effectively less surface area per volume for hydrolysis to occur therefore likely leading to a slower release of cGAMP. This reproducibility in cGAMP release kinetics demonstrates spray drying’s robustness as a MP fabrication method, which is important when scaling up to produce large amounts of MPs at good-manufacturing practice (GMP) quality. The slow release of cGAMP from the MPs is ideal as the cGAMP MPs will reside at the injection site until they are taken up by APCs through phagocytosis (Chen et al., 2018b; Nguyen et al., 2022; Schwendeman et al., 2014). By retaining a high fraction of cGAMP in the MPs, they can then release the adjuvant intracellularly once taken up, due to the low pH of the phagosome and acid-liability of Ace-DEX.

3.2. Spray Dried Ace-DEX MPs Stimulate More Cytokine Production Compared to Spray Dried PLGA MPs

To confirm that cGAMP was being efficiently delivered to cells to initiate STING signalling in the cytosol, DC2.4s were incubated with each MP formulation, and the supernatant was taken and measured for cytokines that are produced from the activation of STING. Firstly, DC2.4s showed comparable percent viability when cultured with the formulations (Figure 4A), indicating that the MPs did not significantly alter cell growth and viability. Emulsion MPs showed no cytotoxicity, whereas all other treatment groups had 10–15% cytotoxicity (Figure 4B). To confirm bioactivity through cGAMP signalling, IFN-β and TNF-α concentrations were measured in the supernatant of the treated cells. Additionally, minimal IFN-β and TNF-α concentrations were measured from the supernatant of cells cultured with the emulsion MPs (Figure 4C,D). This perhaps could be due to the high burst release of cGAMP from the emulsion MPs, thereby having a decreased amount of cGAMP retained within the MPs. Cells treated with the electrosprayed MPs produced considerable amounts of IFN-β, whereas all Ace-DEX spray dried MPs treated cells had significantly lower amounts of cytokine production (p<0.001) (Collier et al., 2018; Junkins et al., 2018). This could be attributed to the difference in cGAMP release profiles where electrospray MPs has faster release kinetics of cGAMP, likely due to its morphology. This then can lead to higher amounts of IFN-β being produced from cells within the 24 hour incubation timeframe. DC2.4s also produced comparable amounts of TNF-α when treated with either electrosprayed or spray dried Ace-DEX MPs (Figure 4D). Cells treated with the PLGA spray dried MPs produced significantly lower amounts of IFN-β and TNF-α compared to Ace-DEX MPs indicating an advantage of using Ace-DEX over PLGA to deliver cGAMP in vitro. This decrease in cytokine production is, again, perhaps due to the slow-release profile of cGAMP seen with the PLGA MPs and the acid-sensitivity of Ace-DEX. Overall, all Ace-DEX spray dried MPs stimulated cytokine production in cells at a similar level, regardless of any changes in fabrication parameter, indicating spray drying’s robustness to produce viable MPs as vaccine carriers.

Figure 4.

In vitro characterization of MPs with DC2.4s. DC2.4s were incubated with each cGAMP MP for 24 hours. After the 24 hours (A) cell viability was measured by MTT assay, (B) cytotoxicity was measured by LDH assay, and (C) IFN-β and (D) TNF-α concentration in the culture supernatant was measured by ELISA. Data is shown as mean±SD. ** p≤0.01, *** p≤0.001, **** p≤0.001 with respect to PLGA. #### p≤0.001 with respect to Electrospray.

3.3. cGAMP Loaded Spray Dried Ace-DEX MPs Provide a Th1 Humoral Response

We next evaluated our MPs in vivo by vaccinating mice with down-selected MP groups to assess their efficacy as adjuvants. Emulsion MPs were not included due to the significantly lower encapsulation efficiency, and lack of cytokine production when incubated with DC2.4s. Electrospray MPs were included as a control since previous studies have shown that Ace-DEX electrospray MPs can elicit a robust immune response (Batty et al., 2022; Gallovic et al., 2022). Spray dried Ace-DEX MPs that were fabricated with an increase in nitrogen drying gas pressure (Q-flow MPs), an increase in polymer concentration in the feed solution (concentration MPs), and an increase in drying gas temperatures (temperature MPs) were chosen. They were chosen because Q-flow MPs exhibited the highest burst release of cGAMP, concentration MPs had significantly slower release overall, and temperature MPs had the lowest burst release. An MF59-like squalene emulsion (Addavax) group was included as a control adjuvant that is clinically approved for influenza vaccines (Chen et al., 2018a; Ko and Kang, 2018). Soluble cGAMP was not included due to previous findings where Ace-DEX cGAMP MPs significantly outperformed unencapsulated cGAMP in enhancing the immune response (Chen et al., 2018a; Gallovic et al., 2022; Junkins et al., 2018). The antigen used in this study was the recombinant protein HA from the A/Puerto Rico/8/1934 (PR8) influenza virus. The adjuvanticity of the spray dried Ace-DEX cGAMP MPs was then assessed by measuring for HA-specific immune responses in vaccinated BALB/c female mice.

To first assess the immune response that can be elicited by spray dried MPs, the humoral response was analyzed by measuring for anti-HA IgG antibodies. By day 42, mice vaccinated with the temperature and PLGA spray dried MPs had significantly lower total IgG titers compared to Addavax vaccinated mice (p<0.01) (Figure 5A), whereas the electrospray, Q-flow, and concentration MPs performed similarly. Addavax outperformed all the spray dried MPs in the generation of serum IgG1 titers on day 42 (p<0.05) (Figure 5B). However, all cGAMP MPs had an increasing trend of IgG2a titers compared to Addavax (Figure 5C). This was to be expected as Addavax alone is a potent stimulator of the Th2 response (indicated by IgG1) but falls short for stimulating Th1 responses (indicated by IgG2a for BALB/c mice) (Ko and Kang, 2018). In contrast to Addavax, cGAMP MPs can provide a Th1 response which is desirable in an influenza vaccine as it has been shown that having both antibody types rather than the skewing of one or the other better correlated to vaccine efficacy and protection (Huber et al., 2006). Mice vaccinated with the temperature and PLGA spray dried MPs had significantly lower IgG2a titers measured by day 42 compared to the electrospray MP vaccinated group (p<0.05). This may suggest that elevated temperatures during fabrication affected the potency of cGAMP in the temperature MPs and that the morphology or release mechanisms of the PLGA MPs may be hindering their pro-inflammatory response, however more work is needed to confirm this hypothesis. Together these findings suggest that cGAMP encapsulated by spray drying in Ace-DEX Q-flow and Concentration MPs can offer a strong Th1 response compared to Addavax, which is essential in providing a robust influenza vaccine, while performing similarly to the electrospray MPs.

Figure 5.

BALB/c (n=5) mice were vaccinated (I.M.) on days 0, 21, and 35. Mice were bled on days 14, 28, and 42 via submandibular cheek bleeds and antibody specific PR8 HA (A) total, (B) IgG1, and (C) IgG2a titers were measured. Data is shown as mean±SD. * p≤0.05 with respect to Electrospray at corresponding timepoint. # p≤0.05, ## p≤0.01 with respect to Addavax at corresponding timepoint.

3.4. Spray Dried cGAMP MPs Provides Comparable Antibody Effector Functionality to Addavax and Electrospray MPs

Th1 type antibodies (e.g. murine BALB/c IgG2a) can have additional functionalities due to interactions with their Fc receptor to induce effector functions (Stewart et al., 2014; Vafa et al., 2014). Antibody effector functions include antibody dependent cellular phagocytosis (ADCP), where a macrophage can phagocytose the cell, and antibody dependent complement deposition (ADCD), where complement component 1q (C1q) binds to the antibody and begins the cascade of complement activation to opsonize cells (Boudreau and Alter, 2019). Antigen-specific Th1 type antibodies can use antibody effector functions to clear both the virus and infected cells that are expressing viral proteins on their surface before virions are formed, halting the spread of infection. Here, the capacity of the vaccine to provide antibodies with effector functions were measured.

Mice vaccinated with either the concentration or PLGA spray dried MPs had significantly higher ADCP scores compared to mice vaccinated with Addavax (p<0.05), electrospray MPs (p<0.05), and Q-flow MPs (p<0.01) (Figure 6A). When measuring for ADCD, all adjuvant groups performed similarly, which was unexpected (Figure 6B) as the cGAMP MPs had increased trends of IgG2a, which is known to have effector functions for BALB/c mice, compared to Addavax. The increased ADCP activity in the Addavax group could be explained by the amount of IgG1 antibodies measured in these mice. Previous studies have shown that murine IgG1 antibodies have light-affinity binding for effector functions (Stewart et al., 2014). Since the Addavax vaccinated mice had the most IgG1 titers measured, these antibodies may be contributing to both ADCP and ADCD activity (Huber et al., 2006; Leatherbarrow and Dwek, 1984; Nimmerjahn et al., 2005; Nimmerjahn and Ravetch, 2005; Stewart et al., 2014). Overall, spray dried Ace-DEX MPs provide comparable humoral responses to Addavax and electrospray MP while also exhibiting robustness in manufacturing by having comparable antibody effector functionality between spray dried batches.

Figure 6.

Day 42 sera from vaccinated mice were used to measure for antibody effector functions: (A) antibody dependent cellular phagocytosis (ADCP) and (B) antibody dependent complement deposition (ADCD). Data is shown as mean±SD. * = p≤0.05, ** = p≤0.01 with respect to PLGA. # = p≤0.05, ## = p≤0.01 with respect to Concentration.

3.5. Ace-DEX cGAMP Spray Dried MPs Enhances Cellular Response Compared to Addavax

Next, to further characterize the cellular response, we sought to assess for HA-specific responses from vaccinated mice. HA-specific production of IFN-γ and IL-2 were measured by ELISpot (Figure 7A, B) and the supernatant was taken to measure for soluble cytokines by ELISA (Supplementary Figure 1A–D). Splenocytes from Addavax vaccinated mice produced miniaml IFN-γ after antigen recall (Figure 7A, Supplementary Figure 1A). This was to be expected since it is not known for Addavax to stimulate type-I interferon cytokines and it was observed that Addavax vaccinated mice had a decrease trend in IgG2a antibodies (Figure 5C) (Ko and Kang, 2018). IFN-γ promotes Th1-type antibodies (murine BALB/c IgG2a) but can also decrease total and Th2-type antibodies (murine IgG1) which explains the results seen in Figure 5 (Vazquez et al., 2015). Furthermore, all mouse groups had comparable IL-2 production from splenocytes except of mice vaccinated with the temperature MPs (Figure 7B, Figure S1B). Mice vaccinated with the temperature MPs had lower amounts of IFN-γ (p<0.05 to electrospray MPs) and IL-2 (p<0.05 to electrospray and concentration MPs) when measured by ELISpot. This further suggests that the elevated temperatures when spray drying could be affecting the cGAMP potency and thereby both the humoral and cellular responses. Further investigation is warranted to confirm this hypothesis.

Figure 7.

On day 45 post prime vaccination, spleens were isolated and made into a single-cell suspension to evaluate their cellular response. ELISpots for (A) IFN-γ and (B) IL-2 expression were used. Representative images of an ELISpot well are shown under their corresponding vaccination group. Data is shown as mean±SD. * p≤0.05, ** p≤0.01 with respect to Addavax. # p≤0.05 with respect to Electrospray. + p≤0.05 with respect to Concentration.

In addition to antigen recall, intracellular cytokine staining was also used to measure for IFN-γ producing T cells in spleens and popliteal lymph nodes. From the spleen, it was again observed that mice vaccinated with the temperature MPs had a decreased trend in IFN-γ positive CD4 T cells and significantly lower IFN-γ positive CD8 T cells compared to electrospray (p<0.005), q-flow (p<0.05), concentration (p<0.05), and PLGA (p<0.01) MP vaccinated groups (Figure 8A, B). Additionally, temperature MP vaccinated mice had significantly lower percentages of IFN-γ positive CD8 T cells in the lymph node compared to electrospray (p<0.05) (Figure 8D), while the other vaccination groups resulted in comparable IFN-γ positive T cells (Figure 8C, D, Supplemental Figure 1C, D). While cGAMP MPs provide a way to elicit a Th1 response by inducing the production of type-I interferons, the results indicate that elevating the temperature may hinder the adjuvanticity of cGAMP (Bot et al., 1998). Therefore, spray dry parameters should be optimized in a way to allow for complete drying with the least amount of thermal stress exhibited on the MPs during formulation to allow the adjuvant to remain intact and potent.

Figure 8.

On day 45 after the first vaccination, (A, B) spleens and (C, D) lymph nodes were isolated from mice and made into single-cell suspensions (n=5). Intracellular cytokine staining was used to measure for IFNγ⁺ cells for (A, C) CD4⁺ and (B, D) CD8⁺ T cells. Data is shown as mean±SD. * p≤0.05, ** p≤0.01, *** p≤0.005 with respect to Temperature.

To further investigate the cellular response and its potential to stimulate plasma cell proliferation, T follicular helper (Tfh) cells were measured via flow cytometry. Mice vaccinated with an adjuvant had comparable amounts of Tfh cells in the popliteal lymph nodes (Figure 9A). Tfh cells are present in the lymph node after being stimulated by APCs and help to activate and continuously stimulate B cells to differentiate into plasmablasts and germinal center B cells (Kräutler et al., 2017). All adjuvanted vaccinated mice had comparable amounts of germinal center B cells in the popliteal lymph node (Figure 9B). Germinal center B cells then differentiate into plasma cells which are important to produce high-affinity antibodies. In the popliteal lymph node, Addavax vaccinated mice had the highest percentage of plasma cells but was not statistically significant against the cGAMP MP vaccinated mice (Figure 9C). This was to be expected as there was an increasing trend of total IgG titers from the Addavax group seen in Figure 5A. Plasma cells from the spleen showed that the mice vaccinated with temperature MPs showed a decrease trend in plasma cell percentage (Figure 9D) which is unsurprising given the decreased cellular responses from splenocytes seen in Figure 7A,B and Figure 8A,B. Overall, these results suggest that cGAMP encapsulated MPs in both Ace-DEX and PLGA from either electrospray and spray dry can provide for the generation of Tfh cells, germinal center B cells, and plasma cells comparable to that of Addavax (Gallovic et al., 2022). With proper optimization, spray drying has demonstrated promise as a scalable and translatable fabrication method for the design of vaccines eliciting a strong Th1 response essential to protect against influenza (Kanojia et al., 2016).

Figure 9.

On day 45 after the first vaccination, (A–C) popliteal lymph nodes and (D) spleens were isolated from mice and made into single-cell suspensions (n=5). Cells were stained to measure for (A) T Follicular Helper Cells (CD4+PD-1^hiCXCR5+) and (B) Germinal Center B cells (CD19+IgM+GL7+CD95+). Cells were then stained to measure for Plasma Cells (CD19+CD138+) from the (C) popliteal lymph node and (D) spleen. Data is shown as mean±SD.

4. Conclusions

To efficiently translate polymeric MPs into the clinic, an inexpensive, continuous fabrication process is needed. Here, we demonstrate that Ace-DEX MPs fabricated by spray drying can act as an adjuvant carrier to deliver cGAMP to cells with enhanced responses in vitro compared to emulsion MPs and comparable humoral and cellular responses in vivo to electrospray MPs and Addavax. We have also demonstrated the robustness of this fabrication method by evaluating the therapeutic responses of MPs that were spray dried with altered polymer solution concentrations, feed solution inlet flowrate, drying gas temperature and pressure (Q-flow). Most spray dried Ace-DEX MP batches performed similarly in vivo which is desirable when scaling up this process. We observed a decreased trend in adjuvant efficacy from mice vaccinated with Ace-DEX MPs generated at elevated temperatures. This may indicate that the thermal stress imposed on the MPs during fabrication can affect the potency of the encapsulate but more work is required to confirm this finding. Both Ace-DEX and the PLGA spray dried MPs elicited comparable therapeutic responses, with some outcomes lessened for PLGA MPs (e.g. in vitro cytokine production). This comparison further validates the use of having Ace-DEX as a drug delivery biopolymer. Overall, our results show that spray drying can offer a robust, one-step, continuous process that can encapsulate cGAMP in Ace-DEX MPs with high reproducibility to be used as an adjuvant system for a subunit influenza vaccine. In the future, we can work towards evaluating this platform for storage outside the cold chain to better facilitate administration in resource limited settings.