Optimizing Delivery in a Multivalent Subunit Influenza Vaccine Using Mixed Polymeric Microparticle Degradation Rates

Abstract

The influenza virus continues to impose a significant yearly burden on society due to the variable efficacy of seasonal vaccines. Further strains like H5N1, that are not included in the seasonal influenza vaccine, may spill over from animal reservoirs and more significantly impact human health. A broadly acting subunit vaccine can offer protection across multiple strains but would have low immunogenicity without an adjuvant, which are currently limited and require delivery systems to mitigate side effects. Further, antigen delivery can be enhanced with carrier systems to provide dose sparing, and thermostability. This study explores acetalated dextran microparticles (Ace-DEX MPs) encapsulating cGAMP and computational optimized broadly reactive antigen (COBRA) hemagglutinin (HA) proteins, to form a multivalent influenza vaccine. Previous research has shown that Ace-DEX cGAMP MPs with varying degradation kinetics can modulate the immune response. Here, we investigate the effects of mixing MPs with different degradation rates to optimize the immune response. Mice vaccinated with slower-degrading cGAMP MPs exhibited higher IgG2a titers and IL-2 producing splenocytes, while those vaccinated with a mix of fast and slow-degrading cGAMP MPs had the highest IFN-γ producing splenocytes. The protection afforded in mice was also shown in ferrets with a H1, H3 and H5 trivalent COBRA formulation adjuvanted by slow degrading cGAMP MPs. Furthermore, using Ace-DEX MPs encapsulating two broadly reactive COBRA H1 and H3 immunogens in particles with fast and slow degradation rates, co-delivered with cGAMP MPs, resulted in less single antigen dominance when the more dominant antigen was encapsulated in the slowest degrading MP. This work underscores the utility of Ace-DEX MPs as a vaccine delivery platform and the impact of MP degradation kinetics on vaccine efficacy.

Graphical abstract

Introduction

Globally each year, influenza virus causes 3-5 million severe acute respiratory infectious.[1] In the United States, the CDC estimates that influenza infects up to 45 million people annually, resulting in 810,000 hospitalizations and 61,000 deaths, with an economic burden of up to $87.1 billion from medical costs and projected loss of earnings.[2, 3] Seasonal vaccination is the primary method of protection against influenza, but it requires yearly reformulation since current vaccines generate strain-specific immunity towards the highly variable globular head region of the hemagglutinin (HA) protein.[4, 5] Additionally, even when the strains are well-matched, current vaccines suffer from low immunogenicity causing the protective response to wane over time.[6]

Strain mismatch between the vaccine and circulating strains is particularly concerning in the consideration of the stockpiled H5N1 vaccines. Current stockpiles are a single clade (H5N1 clade 2.3.4.4b) from the 2000’s.[7, 8] The H5N1 influenza virus poses a significant global health threat due to its high mortality rate and pandemic potential.[9] Unfortunately, H5N1 is not included in the seasonal influenza vaccine because it is very difficult to grow it in eggs which is the most common method to manufacture seasonal influenza vaccines.[10] Additionally, inclusion of a new antigen into a multi-valent influenza vaccine could lead to immunodominance because high-affinity B cells reactive to one antigen may compete B cell responses to other antigens and diminish protection.[11, 12]

A more effective influenza vaccine would incorporate a broadly active or universal influenza antigen. Conserved antigens and epitopes such as M2e (matrix protein 2 ectodomain) and the HA stem region have shown promise but have limitations.[13, 14] M2e vaccines primarily rely on antibody effector functions, while HA stem antibodies can be autoreactive and lead to increased polyreactivity.[15, 16] Both antigens elicit limited neutralizing antibody responses, which are crucial for protection against influenza infection.[17] An alternative and promising approach is COBRA (Computational Optimized Broadly Reactive Antigen) HA. The COBRA methodology generates HAs specific for influenza subtypes (e.g., H1, H3) based on HA sequences from previous years and has been shown to generate neutralizing antibodies against past, present, and future strains. [18]

However, protein vaccines are poorly immunogenic and primarily elicit a T-cell type 2 (Th2) humoral-biased response. Generating a Th1 response has been shown to provide robust protection against influenza.[19-22] Incorporating an adjuvant into the vaccine formulation can increase immunogenicity while providing a Th1-skewed response. The only FDA-approved adjuvant for use in a seasonal influenza vaccine is MF59 (FluAd), an oil-in-water squalene emulsion adjuvant that primarily elicits a Th2-biased response. [23] Therefore, there is a need for an adjuvant system that can pair with a subunit vaccine to provide a Th1 response.

One such Th1-skewing adjuvant is cGAMP that binds to the intracellular stimulator of interferon genes (STING) receptor. However, cGAMP’s hydrophilic nature makes it difficult to pass through the cellular membrane to bind to its cytosolic receptor. Our lab has previously reported encapsulating cGAMP into polymeric acetalated dextran microparticles (Ace-DEX MPs) to enhance the efficacy of a subunit vaccine by passively targeting phagocytes such as antigen-presenting cells (APCs) and rapidly releasing the cGAMP into the phagosome due to its pH sensitive nature.[24-27] Ace-DEX is produced in a one-step reaction where dextran reacts with 2-ethoxypropene to substitute the hydroxyl groups with cyclic and acyclic acetal groups.[28] During synthesis, the coverage of cyclic acetal increases with reaction time. Since cyclic acetals hydrolyze at a slower rate than acyclic acetals, the cyclic acetal coverage (CAC) dictates a wide range of tunable degradation rates. Controlling the CAC of the Ace-DEX used to formulate MPs allows for degradation half-lives (t_1/2) ranging from hours to months at neutral pH.[26, 29] In the phagosome where the pH is closer to 5.0, Ace-DEX MPs degrade approximately two-logs faster than neutral pH, and have a half-time to degradation for fast degrading (20 CAC) MPs of 0.25 hr, medium degrading (40 CAC) MPs at 2.9 hr, and slow degrading (60 CAC) MPs at 21.3 hrs.[29] Those CAC rates were previously selected and characterized due to their unique different release profiles, in addition, they have been tested for in vivo delivery of vaccine adjuvants, showing differences in the specific antibody and cellular immune responses.[30, 31]

Some studies have reported the importance of polymeric particles as controlled release strategies to modulate the immune response locally or systemically, those particles include metal nanomaterials, synthetic and natural polymeric particles, including acetalated dextran as a crucial pH sensitive material for controlled release.[32, 33] Exploring the impact of Ace-DEX MP CAC on antigen and adjuvant release, we have reported that optimal release rates differ for a given antigen or adjuvant, but that encapsulation of antigen and adjuvant in separate particles led to increased correlates of protection and survival after a challenge.[31, 34-37] In exploring the delivery of M2e and cGAMP with Ace-DEX MPs, significant differences were observed across slow, medium, and fast degrading MPs for both antigen and adjuvant. For both M2e and cGAMP, the 60 CAC MPs displayed the highest total titers, but when ovalbumin and murabutide were evaluated, the peak titer responses were observed with the fastest and medium degrading MPs, respectively. Additionally, when cellular responses were evaluated through antigen recall using splenocytes from vaccinated mice, cellular responses peaked at the slowest degrading MPs. To evaluate co-encapsulation versus separate MPs, M2e and cGAMP were combined in one 60 CAC MP at the same loading as separate MPs and given in a prime and boost schedule. Delivery of the antigen and adjuvant in separate MPs lead to significantly greater protection against influenza. Additionally, significant differences in protection from challenge and cross-reactive antibody titers were observed between slow and fast degrading MPs. [34] Our studies concluded that encapsulation of antigens and adjuvants in separate MPs led to better protection and survival against influenza, and that varying MP degradation rates significantly affect immune responses.

Our previous work relied on vaccination with MPs of a single degradation rate. To build on this, we then sought to mix MPs of different CAC to evaluate if this would lead to more sustained and stronger immune responses. cGAMP loaded MPs of 20, 40 and 60 CAC or mixtures thereof, were evaluated for in vitro and in vivo with soluble HA. We then accessed a tri-valent (H1, H3, H5) COBRA and cGAMP MP vaccine in a ferret model of infection. As an assessment of MP degradation on the antigen, we have encapsulated H1 and H3 COBRA antigens in MPs composed of either 20 or 60 CAC Ace-DEX MPs and evaluated individual or mixtures of these MPs in vitro and in vivo. Evaluating a bi-valent vaccine also allows us to explore the effect of antigen immunodominance on degradation rate. Overall, this study further evaluates the effect of MP degradation rate on correlates of protection and protection after influenza challenge.

Materials and Methods

Unless otherwise stated, all chemicals and assays, biologicals, and disposables were purchased from Millipore Sigma (St. Louis, MO) and Thermo Fisher Scientific (Waltham, MA), respectively.

Acetalated Dextran (Ace-DEX) Synthesis

To synthesize Ace-DEX, Dextran (71k kDa) from Leuconostoc mesenteroides was dissolved in dry dimethyl sulfoxide (DMSO).[28] Dextran then reacts with 2-ethoxypropene (Matrix Scientific, Columbia, SC) in the presence of pyridinium p-toluenesulfonate, for either 2, 15, or 120 minutes to produce Ace-DEX of 20, 40, or 60 CAC, respectively. Once the reaction time was met, triethylamine (TEA) was used to quench the reaction. Ace-DEX was then precipitated out when added drop-wise to basic water (0.04% v/v TEA in water), isolated by centrifugation, and lyophilized overnight. Ace-DEX was dissolved in ethanol and centrifuged to remove ethanol-insoluble impurities. The Ace-DEX solution was precipitated again in basic water and lyophilized for two days. To confirm the CAC of the polymer, a ¹H 400 MHz NMR (Inova) was used.

cGAMP Spray-dried Ace-DEX MPs

A Mini Spray Dryer B-290 (Büchi, New Castle, DE) was used to fabricate the cGAMP Ace-DEX MPs of different CAC for evaluation with HA.[27] First, Ace-DEX and cGAMP were dissolved in a 90:10 ethanol:water solvent system at 20 and 0.2 mg/mL, respectively. The solution was sprayed at a flowrate of 10.7 mL/min with nitrogen pumping in at a q-flow of 40mm and heated to 50°C. The aspirator ran at 100%. MPs were then stored in a −20°C freezer. Except where indicated, cGAMP MPs were fabricated using 71k Da dextran. For evaluation of gamma irradiation sterilization, Ace-DEX cGAMP MPs of three different molecular weights (71k, 500k, and 2M kDa) were generated via spray drying.[29] Evaluation of sterilization was performed using gamma irradiation at a total dose of 25 kGy (UNC BRIC), that accumulated over approximately 535 hours at room temperature. A room temperature control cGAMP MPs was also prepared wherein the particle set sat at room temperature for the same duration that irradiation was administered.

Electrosprayed Ace-DEX MPs

H1 (Y2) and H3 (J4) COBRA HA were provided by the Ross Lab from the University of Georgia. Immunogens were produced by HEK293T suspension cells that were transfected with a truncated COBRA J4 and Y2 gene cloned into pcDNA3.1⁺ plasmid.[18] Both J4 and Y2 were encapsulated in Ace-DEX MPs through a coaxial electrospray set developed by our lab.[38] First, 20 or 60 CAC Ace-DEX was dissolved at 25 mg/mL in a ethyl acetate:butanol (70:30 v/v) solvent system. The immunogen was suspended in 5% v/v glycerol water at 8.125 mg/mL. The solutions were loaded into separate glass syringes (Hamilton Company; Reno, NV) where the needles were connected to a coaxial spray needle (Rame-Hart; Succasunna, NJ). The Ace-DEX solution comprised the outer-shell and the antigen was placed in the inner-core portion. The syringes were placed on a syringe pump (Harvard Apparatus, Holliston, MA) where the outer-shell was ejected at 0.85 mL/hr and the inner-core was ejected at 0.02 mL/hr. A negative voltage of −5 kV was set on the needle and a positive voltage of +10 kV was set on the collection plate where the MPs were sprayed onto.

For evaluation of J4 and Y2 encapsulated antigen, cGAMP MPs were also prepared through electrospray. cGAMP (3’-3’ cyclic GMP-AMP; Invivogen; San Diego, CA) was encapsulated in 60 CAC Ace-DEX using a monoaxial electrospray method developed by our lab.[27] Both Ace-DEX and cGAMP were dissolved in an ethanol:water (90:10 v/v) solvent system at 20 and 0.2 mg/mL, respectively. The solution was loaded into a glass syringe and placed on a syringe pump over a collection plate. A negative voltage of −5 kV was placed onto the needle and a positive voltage of +2.5 kV on the collection plate. The pump was set to eject the solution at 0.2 mL/hr.

Ace-DEX MP Characterization

Scanning electron micrograph (SEM) images were taken using a S-4700 Cold Cathode Field Emission Scanning Electron Microscope (Hitachi High-Technologies, Krefeld, Germany; UNC CHANL). To create the SEM stubs used for imaging, MPs were suspended in basic water and a droplet was placed on top of a carbon taped stub to dry overnight. Endotoxin levels were measured using a Pierce LAL Chromogenic Endotoxin Quantification kit where all MPs were found to be below 0.125 EU/mg which is lower than the limit for many preclinical formulations.[39]

For release studies, Ace-DEX MPs were suspended in phosphate buffer saline (PBS, pH 7.4) at 1 mg/mL and kept at 37°C on a heated shaker plate to mimic physiological conditions. Aliquots of 300 μL were taken at each timepoint, centrifuged, and pellet and supernatant were stored at −20°C. cGAMP content in the pellet were measured via HPLC by dissolving the pellet in 300 μL of 75:25 water:acetonitrile with 0.1% v/v TFA and running the solution through an Aquasil C18 column at 1 mL/min. cGAMP was detected at 210 nm. J4 and Y2 loading was measured using a fluorescamine assay.

For cGAMP encapsulated in Ace-DEX by spray drying was found to be 8.4, 7.6, and 8.1 μg/mg of MP for the 20, 40, and 60 CAC Ace-DEX MPs, respectively. The loading for J4 encapsulated in 20 and 60 CAC Ace-DEX MPs were found to be 3.94 and 4.06 μg/mg, respectively. The loading for Y2 encapsulated in 20 and 60 CAC Ace-DEX MPs were found to be 3.44 and 3.19 μg/mg, respectively. The loading of cGAMP encapsulated in 60 CAC Ace-DEX MPs through electrospray was found to be 8.33 μg/mg.

Bone Marrow-Derived Dendritic Cell Stimulation

Bone marrow was taken from C57BL/6 mice and cultured with 10 ng/mL granulocyte-macrophage colony-stimulating factor (GM-CSF) (Peprotech 315-03) and 10 ng/mL IL-4 (Biolegend 574306) to prepare bone marrow-derived dendritic cells (BMDCs).[40] BMDCs were plated at 50,000 cells per well in RPMI media and were allowed to adhere overnight. The following day, BMDCs were treated with 1 μg/mL of cGAMP from MPs for 24 hours. The supernatant was then taken and measured for IFN-β concentration by sandwich ELISA (Biolegend, San Diego, CA). Additionally, BMDCs were treated with MP solutions that were incubated for 48 hours prior to treatment. MPs were suspended in RPMI media and kept on a heated shaker plate at 37°C. After 48 hours, the solution was vortexed and used to treat BMDCs for 24 hours.

Vaccinations

In vivo mouse studies were conducted following the guidelines set by the National Institute of Health which was approved by the Institutional Animal Care and Use Committee at UNC. BALB/c mice were purchased from Jackson Laboratories (Bar Harbor, ME) at 6-8 weeks of age. On days 0, 21, and 35, mice were vaccinated intramuscularly (i.m.) where both legs received 25 μL for a total of 50 μL per mouse. Where indicated, mice received 1 μg of recombinant A/Puerto Rico/8/1934 hemagglutinin (PR8 HA) (BEI Resources, Manassas, VA) and 0.1 μg of cGAMP from spray dried Ace-DEX MPs. Additionally, where indicated, C57BL/6 mice (Jackson Laboratories) at 6-8 weeks old received 1 μg of J4 and/or Y2 in either soluble form or in Ace-DEX MPs and 1 μg of electrosprayed cGAMP in Ace-DEX MPs.

Serum Antibody Concentration

On days 14, 28, and 42, anti-HA antibodies (PR8 HA, Y2 or J4, as appropriate) were measured from serum taken via submandibular cheek bleeds. First, 25 μL of HA in PBS at 0.1 μg/mL was added to each well of a 384-well plate (Greiner, Kremsmünster, Austria) and stored at 4°C overnight. Plates were washed with wash buffer composed of PBS with 0.05% Tween 20. Wells were blocked using 3% non-fat instant milk (Food Lion, Raleigh, NC) in PBS. Secondary antibodies used were Goat Anti-Mouse IgG Fc-HRP, IgG1 Fc-HRP, and IgG2a-HRP (Southern Biotech, Birmhingham, AL). Samples were developed using tetramethylbenzidine one component substrate (Southern Biotech, Birmhingham, AL) with 2N sulfuric acid as the quenching buffer. Absorbance was read at 450 and 570 nm. Antibody titers were calculated by curve fitting the values against the background-corrected absorbance using the “log(inhibitor) vs. response – Variable slope (four parameters)” built-in model in Graphpad Prism 9 and interpolating which dilution intercepts the endpoint value.[41]

To measure for antibody effector functions, PR8-HA conjugated microspheres were used.[42] Carboxylate-modified fluorescent microspheres were mixed with sulfo-n-hydroxysuccinimide and 1-ethyl-3(3-dimethylaminopropyl)carbodiimide hydrochloride. PR8 HA was then incubated with the microspheres for 2 hours to allow for conjugation to the surface. Afterwards, microspheres were then washed and stored at 4°C until use.

Antibody-dependent complement deposition (ADCD) was assessed on day 42 serum.[43] First, sera was heat inactivated at 56°C for 30 minutes then diluted. HA-conjugated microspheres were added to the sera solution and kept at room temperature for 2 hours. Reconstituted lyophilized guinea pig complement (Cedarlane Labs, Burlington, ON, Canada) suspended in barbitone buffer was added to samples and incubated at 37°C for 20 minutes. Samples were washed with 15 mM EDTA in PBS and stained with FITC anti-guinea pig C3 in PBS for 30 minutes. Microspheres were then washed with 3% v/v FBS in PBS and analyzed using the Attune NxT Flow Cytometer. ADCD score was then calculated using the percent of complement positive beads divided by 10,000 times the geometric mean fluorescence intensity from the complement channel.

Antibody-dependent cellular phagocytosis (ADCP) was assessed on day 42 serum.[44] First, J774a.1 murine macrophages (ATCC, Manassas, VA) in RPMI media were plated at 50,000 cells/well in non-tissue culture treated 96-well plates and incubated overnight. HA-conjugated microspheres were added to diluted sera for 2 hours. Samples were then added to J774a.1 cells and incubated at 37°C for 1 hour. Cells were centrifuged and washed with 5 mM EDTA in PBS. Cells were then fixed with 3% paraformaldehyde in PBS. Cells were analyzed with an Attune NxT Flow Cytometer for microsphere uptake. ADCP score was then calculated using the percent of microsphere positive cells divided by 10,000 times the geometric mean fluorescence intensity in the microsphere channel.

Antigen Recall

Spleens were harvested from vaccinated mice and processed into single-cell suspensions in RPMI media. Splenocytes were seeded at 1,000,000 cells/well on a pre-coated 96-well PVDF filter ELISpot plate. Cells were either stimulated with 10 μg/mL of PR8 HA peptide array (BEI Resources, Manassas, VA) in media or non-stimulated with media. Following the manufacturers protocol (BD, Franklin Lakes, NJ), plates were developed and counted for spots of cells that produced IFN-γ and IL-2 in an immunospot counter (C.T.L., Cleveland, OH).

Flow Cytometry of Splenocytes and Lymphocytes

Spleens, popliteal and inguinal lymph nodes from vaccinated mice were harvested and processed into single-cell suspensions. Splenocytes and pooled lymphocytes were plated at 1,000,000 cells/well on a non-tissue culture treated 96-well plate. Cells were incubated with either media (unstimulated), or 1 μg of J4 or Y2 for 36 hours. Cells were then stained with a flow staining panel of: fixable viability dye (FVD; eBioscience Fixable Viability Dye eFluor 506), anti-CD19 (APC, clone: 6D5, concentration: 0.5 μg/mL), anti-GL7 (PE, clone: GL7, concentration: 0.125 μg/mL), and anti-CD38 (PE/Cyanine7, clone: 90, concentration: 0.25 μg/mL). Cells were then fixed with 1% paraformaldehyde in PBS. Cells were analyzed using a Thermo Fisher Attune NxT provided by the UNC Flow Cytometry Core Facility.

Ferret study

Y1 (H1), J4 (H3), and/or IAN5 (H5) COBRA HA were provided by the Ross Lab from the University of Georgia. COBRA consensus amino acid sequences were obtained from the screening of public online databases; Y1 was derived from H1 sequences isolated between 2013 and 2015, J4 from H3 sequences between 2013 and 2016, and IAN5 from H5 sequences from 2011 to 2015.[45] Ferrets from Triple F Farms (Gillett, PA, USA), 6 months of age were vaccinated i.m. on days 0, 21, and 42. For each vaccination, they received 15 μg of cGAMP encapsulated in spray dried 60 CAC Ace-DEX MPs with 15 μg of each Y1, J4, and/or IAN5 (45 μg total protein for trivalent formulation). Ferrets were bled 14 days after each vaccination for HAI titers. On day 70 they were challenged with H1N1 A/California/07/2009 (2x10⁶ pfu) or H3N2 A/Kansas/14/2017 (1x10⁷ pfu). At day 5 postinfection, nasal washes (NW) were performed on them to evaluate viral titers. Their weight was checked daily for 13 days after infection to monitor weight loss. Viral titers were evaluated through plaque assays as previously described, NW were frozen at −80°C until analysis, when the samples were thawed and plated over a confluent monolayer of MDCK cells to determine number of plaques per sample.[46]

Results and Discussion

Effect of Mix CAC MPs on cGAMP delivery

Ace-DEX MPs were fabricated of 20, 40 and 60 CAC and displayed consistent morphologies with similar weight loadings (Figure 1A-C). We have previously reported the morphology characterization of cGAMP and COBRA encapsulated in 20, 40 and 60 CAC MPs, where the average diameter in all cases is around 650 μm (with a polydispersity index of around 0.400), [31, 47, 48] consistent with the observed in Figure 1A-C. Importantly, for both cGAMP and COBRA MPs, we have also reported that the MP morphology is stable and does not significantly change after a 6-month period.[39] The release of cGAMP from the MPs displayed unique durations with half-lives ranging from half-a-day to one month (Figure 1 D-F). These results are consistent with previous findings where release kinetics of a drug can be tuned based on the CAC of Ace-DEX used for the polymeric matrix[29, 49] and unique from previous studies with emulsion particles.[26, 27] When cGAMP MPs of different CAC were cultured with BMDCs, the fast degrading 20 CAC had the highest production of IFN-β in the supernatant compared to 40 and 60 CAC MPs at 24 hours (Figure 1G; Supplemental Table 1) and a CAC dependent cytokine response was observed for the single CAC MPs. This would correspond to our previous report[50] that 20 CAC Ace-DEX MPs are taken up quickly, degraded in phagosome, and permit release of phagosomal contents into the cytosol where the STING receptor is located. When MPs of different CACs were combined, CAC dependent differences were observed, but not as significant as single CAC MPs, which would indicate CAC and CAC mixtures affect IFN-β production.

Figure 1.

Scanning electron micrographs of cGAMP encapsulated in MPs composed of (A) 20, (B) 40, or (C) 60 CAC Ace-DEX. BMDCs were treated with 1 μg/mL of cGAMP in their respective MPs. Scale bar drawn is 10 μm. (D, E) cGAMP release profiles of the 20, 40, and 60 CAC cGAMP MPs when suspended in PBS at 1 mg/mL and kept on a heated shaker (37°C) up to 42 days. (F) Calculated release half-lives for data presented in E. (G) IFN-β production by bone marrow derived dendritic cells. Statistics for all groups are provided in Table S1.

To evaluate the effect of combined MP degradation rate in vivo, mice were vaccinated in a prime-boost-boost schedule (days 0, 21 and 35). Serum isolated on day 42 was evaluated for antibody effector function (Figure 2A; Supplementary Figure 1). ADCD (antibody-dependent compliment deposition) had a non-significant inverse correlation with degradation rate, wherein the slowest degrading MPs (60 CAC) had the highest score. When CACs were mixed, the ADCD varied, but not significantly so across all mixed groups. In comparing 60 CAC MPs to mixtures with 60 CAC, higher complement deposition functionality was observed in the 60 CAC MPs compared to mice vaccinated with the 20/60 and 40/60 cGAMP MPs. The trends in ADCD largely correlated with day 28 IgG2a titers (Supplementary Figure 2). Differences between the CACs were not as readily observed with evaluation of ADCP (antibody-dependent cellular phagocytosis; Supplementary Figure 1). By day 42, nearly all vaccinated groups had equally high generation of IgG1 titers (Supplementary Figure 2) which most likely is contributing to the lack of differences in ADCP functionality. Overall, for antibody functionality the slowest degrading MPs (60 CAC) had the highest ADCD scores, while differences in ADCP functionality were minimal across all groups due to uniformly high IgG1 titers by day 42.

Figure 2.

BALB/c mice were vaccinated in a prime-boost schedule. Day 42 serum was assessed for (A) antibody-dependent complement deposition. Antibody titers were analyzed till day 112 to assess for long-term humoral immunity. The area under the curve (AUC) of (B) Total IgG, (C) IgG1, and (D) IgG2a is shown. Titer over time for IgG2a is also presented in (E). 45 days after the first prime vaccination, spleens were harvested and evaluated for their cellular responses. ELISpots measuring (F) IFN-γ spots per a one-million cells with representative spot images provided over their corresponding groups. Data is shown as mean±SD. * p≤0.05, ** p≤0.01.

To understand the role of particle degradation and CAC mixtures on humoral responses, serum was isolated in vaccinated mice out to day 112, post prime. Consistently all total IgG, and IgG1 antibody titers were similarly high, compared to soluble HA alone, across all time points regardless of MP CAC (Figure 2 BC; Supplementary Figure 3). Differences in the area under the curve (AUC) for titers were noted for total IgG and IgG1 across the CACs, with the most pronounced differences observed in Th1-skewing IgG2a. The 60 CAC MPs elicited significantly higher titers compared to the 20 CAC group, and the slowest degrading CAC mixture (40/60) also showed significant differences compared to the fast-degrading 20 CAC group (Figure 2 DE). IgG2a antibodies are known to require longer stimulation to be generated than Th2 skewing IgG1, which may underscore the importance of MP degradation in skewing immune response.[51] These trends align somewhat with what was observed previously with M2e and cGAMP, however differences in antibody titers were observed across all isotypes previously. This difference may be due to the use of a peptide antigen over a whole protein antigen that would have more epitopes.[34]

Cellular responses are also needed to provide protection against influenza and studies have shown that duration of innate signaling can affect the generation of antigen-specific memory T cells.[19-22, 52] Splenocytes isolated on day 45 from vaccinated mice and stimulated with HA showed that only the mice vaccinated with the 20/60 CAC MP mixture had significantly higher IFN-γ production compared to the PBS-vaccinated mice (Figure 2F). However, for IL-2, mice vaccinated with higher CAC cGAMP MPs trended to have more IL-2 producing cells with the 20/60 group producing in between the 20 and 60 CAC MPs (Supplemental Figure 4). This was not in alignment with previous studies with murabutide Ace-DEX MPs of varying CACs, where the faster degrading MPs had the highest cytokine production, underscoring that each adjuvant has different optimized kinetics, likely because they signal differently in the cell and this also is reflected in the differential expression of cytokines.[53, 54] Overall, the cGAMP MP data indicates that the most pronounced differences are observed between the slowest and fastest degrading MPs. That the slowest degrading Ace-DEX has the biggest impact on innate as well as humoral immunity. Cellular responses seemed the most affected by the combination of MPs of different CAC, albeit not statistically.

Protection in a Ferret Challenge Model

As previously showed, the slow degrading MPs elicited a higher immune response overall. For that reason, we evaluated the efficacy of slow degrading cGAMP MPs as a multivalent COBRA vaccine in a ferret model of influenza infection. The ferret model is the ideal animal model for influenza because of their relatively small size and they possess various clinical features associated with human influenza infection.[55] Ferrets were immunized at a prime-boost-boost schedule with a trivalent formulation containing soluble proteins H1 COBRA (Y1), H3 COBRA (J4), and H5 COBRA (IAN5), co-delivered with cGAMP MPs. This trivalent cGAMP MP group was compared to a trivalent Addavax (MF59 mimic) group as well as a monovalent Y1 with cGAMP MPs group. The HAI titers of the ferrets immunized with trivalent COBRAs and cGAMP MPs were similar to the Addavax control. In comparing the monovalent H1 to the trivalent formulation, the HAI against H1 viruses (CA/09, MI/15, Bris/19 and Guangdong/19) was on average higher and more rapid than the trivalent formulations (Figure 3 AB, Supplemental Figure 5), indicating that although there is the same concentration of each antigen between groups, the additional antigens in the formulation are inhibiting the neutralization of the H1 antigen.

Figure 3.

Ferrets were injected intramuscularly at day 0, 21 and 42 and challenged with A/California/07/2009 (CA/09 - H1N1) or A/Kansas/14/2017 (KS/17 - H3N2). HAI titers were evaluated against (A) A/California/07/2009 or (B) A/Hong Kong/4801/2014. Viral titers of (C) CA/09 or (D) KS/17 in nasal washes (NW) were evaluated at day 5 post-infection. Weight loss percentage was monitored during 13 days after infection of (E) CA/09 or (F) KS/17. Data is shown as mean±SD. Dotted lines on A and D indicates seroconversion and seroprotection levels (lower and upper, respectively). Dotted line on B and E indicated the endpoint of weight loss. * p≤0.05, **** p≤0.0001.

Interestingly, for both challenge strains (H1N1 CA/09 and H3N2 KS/17), the viral titers from nasal washes at the peak of infection (day 5) showed minimal virus in vaccinated ferrets for both H1N1 and H3N2 infections, demonstrating the protection induced by this multivalent vaccine (Figure 3 CD). The weight loss of the trivalent vaccinated group was slightly decreased when compared with the monovalent vaccinated group. None of vaccinated groups reached 20% weight loss and showed protection against both influenza strains (Figure 3 EF).

With both the HAI and challenge data, the monovalent vaccine generated on average, and in some cases a significantly, greater response than the trivalent formulation, regardless of adjuvant. This would indicate that the immunogenicity of the H1 antigen is inhibited by the inclusion of the H3 and H5 antigens in the trivalent formulation, comparable to what has been observed in similar ferret studies.[56] On the phylogenetic tree of influenza A, H1 and H5 are nearby in the Group 1 viruses whereas H3 viruses are more genetically diverse in the Group 2 viruses.[57] Likely the overlap is between H1 and H3 since they are more closely related. In a mouse study of a trivalent H1, H3, H5 COBRA formulations adjuvanted with cGAMP, a similar reduction was seen for HAI and challenge in comparison to the monovalent formulation with H1 COBRA Y1.[58] Therefore we evaluated the effect of particle degradation on antigen dominance using COBRA antigens for H1 and H3.

Effect of Mix CAC MPs on HA delivery

To understand if MP degradation can play a role in overcoming antigen dominance, MPs of different degradation rates were used to encapsulate H1 and H3 COBRAs (Supplemental Figure 6). COBRA antigens that are broadly protective against H1 viruses (Y2) and H3 viruses (J4) were encapsulated in fast (20 CAC) or slow (60 CAC) MPs and were evaluated individually and in mixtures (Figure 4). The serum antibody titers from mice vaccinated in a prime-boost-boost schedule, displayed changes in antibody titers when Y2 and J4 were given separately or together in soluble form, with cGAMP MPs. J4 specific titers were reduced for total IgG and IgG1 when antigens were combined, compared to when they were given at the same amount separately (Figure 4; Supplemental Tables 2-4; Supplemental Figures 7-9). However, when J4 was encapsulated into Ace-DEX MPs, J4 antibody levels were increased closer to what was observed with soluble J4 alone. Similarly, Y2 specific total IgG was increased when encapsulated. Further, when both antigens were encapsulated in 60 CAC MPs, their titers were trending higher than when encapsulated in 20 CAC MPs. The mix of 20 and 60 CAC MPs faired similarly except for IgG2c. Encapsulating Y2 in 20 CAC MPs and co-vaccinating with J4 in 60 CAC MPs led to baseline levels of Y2 specific titers. However, when Y2 was encapsulated in the more slowly degrading 60 CAC MPs and co-vaccinated with J4 in 20 CAC, IgG2c titers increased while J4 specific titers stayed high. This influence of CAC was notable for titers, across isotypes, and at all time points measured (Supplemental Figures 7-9). In general, Y2 appeared to influence the titer development more for J4 than vice versa. Encapsulating Y2 in a slower degrading polymer seemed to diminish the antigen’s immunodominance when co-delivered with J4, such can be seen with the decrease in the immune response when the antigens are encapsulated in slow-release MPs, as seen with IFN-γ response in splenocytes when Y2 is encapsulated in 60 CAC MPs (Supplemental Figure 11). A decrease of immunodominance when antigens are delivered in slow-release platforms have been reported as a result of a continuous presence of antigen in the draining lymph node, stimulating a higher local follicular helper T cell and germinal center responses and deriving in a similar systemic antibody response.[59, 60] This would indicate that changing the controlled release of one antigen can skew immune responses towards one or both antigens.

Figure 4.

Mice were vaccinated in a prime-boost-boost schedule (day 0, 21, and 35) with PBS, and soluble or encapsulated J4 and/or Y2 in 20 or 60 CAC MPs. All mice received cGAMP MPs as an adjuvant except for the PBS group. Day 41 sera were analyzed for (A) Total IgG, (B) IgG1, and (C) IgG2c titers specific for J4 and Y2. Full statistics are presented in Supplemental Table 2-4. Antibody titers at additional time points (Day 14 and 28) are given in Supplemental Figures 7-9. Representative HAI against (D) A/California/07/2009 (H1N1)pdm09 (CA/09) and (E) A/Texas/50/2012 (H3N2) (TX/12) for day 41 sera. Additional HAI are presented in Supplemental Figure 10. * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001, ns – not significant.

Hemagglutination inhibition (HAI) is used as a metric by the FDA to determine seasonal influenza vaccine efficacy.[61] Whereas seasonal influenza vaccines will primarily only inhibit the strain included that year in the vaccine, COBRA antigens will display broad inhibitions across multiple strains. Figure 4 presents two representative HAI against A/California/07/2009 (H1N1)pdm09 (CA/09) and A/Texas/50/2012 (H3N2) (TX/12), with other strains for both H1 (A/Michigan/45/2015, A/Brisbane/02/2018 and A/Guangdong-Maonan/SWL1536/2019) and H3 (A/Switzerland/9715293/2013, A/Hong Kong/4801/2014, and A/Hong Kong/2671/2019) viruses presented in Supplemental Figure 10. For H1 HAIs, the Y2 Ace-DEX MP encapsulated groups across all strains except GD/19 outperformed the soluble groups, with titers that on average are much higher than observed with soluble groups where most mice do not seroconvert. Across the H3 strains, the soluble J4 groups performed on average the same as all encapsulated MP groups. This would trend with the antibody results that J4 elicited higher antibody responses than the Y2 antigen (Figure 4), and it would appear those titers are also significantly inhibitory and often more so with slower degrading polymer.

In addition to increased humoral responses, J4 also seems to promote higher cellular responses in vivo than Y2. When splenocytes isolated from vaccinated mice on day 45 were restimulated with Y2 or J4, significant increases in IFN-γ and IL-2 were observed via ELISpot for the J4 antigen, and only moderate increase was noted for the Y2 antigen (Supplemental Figure 11). Notably, encapsulated antigen, when combined, had significantly greater IFN-γ production compared to soluble antigens when combined, and was similar to either antigen alone in soluble form. This would indicate that the inhibition of antigen recall observed with delivery of two antigens can be overcome with encapsulation of antigen, as measured through cytokine responses. It was observed that the J4 20 CAC and Y2 60 CAC had the greatest cytokine production of all the groups, again underscoring the fine tuning that can be achieved with Ace-DEX MPs. Similarly, the generation of germinal center B cells in spleen and lymph node illustrate that on average, J4 generates a more significant response, and that response is maximized when it is encapsulated in fast degrading 20 CAC MPs and Y2 is encapsulated in slow degrading 60 CAC MPs (Supplemental Figure 11). On the other hand, the generation of memory T cells is generally higher when both Y2 and J4 are encapsulated in slow degrading 60 CAC MPs (Supplemental Figures 12 and 13). These results are similar with our findings in ferrets (Figure 3), where the antibody response (HAI titers) given by the H1 antigen was decreased when H1 and H3 COBRA MPs are co-immunized, however, the delivery of both antigens in 60 CAC were able to decrease the immunodominant effect. There is a lack of effective influenza challenge models in mice but given the full protection and the lack of viral titers in the ferrets immunized with the trivalent formulation, it is expected that the mice immunized with the bivalent 60 CAC MPs are also protected due to the similarity in the antibody response. Overall, encapsulating COBRA HA into Ace-DEX MPs with different degradation rates can modulate immune responses, enhancing antibody and cellular responses, and potentially improving vaccine efficacy against multiple influenza strains.

Toxicology and Post Processing Sterilization

To evaluate the potential toxicity of cGAMP MPs in vivo, given its higher immunogenicity compared to HA protein, a toxicity study was conducted using mice. The mice were divided into groups and injected intramuscularly with either PBS, soluble cGAMP, a 20/40/60 CAC mixture of blank MPs, or single CAC encapsulating cGAMP. After a period of 7 days, the mice were subjected to toxicological assessments. Given that activation of the cGAS-STING pathway can potentially lead to liver disease, we specifically focused on evaluating liver inflammation and indicators of liver distress.[62] These indicators included ALT (alanine aminotransferase) and ALB (albumin), which are critical markers for liver function and health[41]. Immunohistological analysis of liver tissue, stained with hematoxylin and eosin (H&E), was performed to detect any signs of inflammation (Figure 5A-F). The results revealed no significant differences in liver inflammation among the various experimental groups. This suggests that the presence of cGAMP MPs did not induce any noticeable inflammatory response in the liver. Furthermore, blood chemistry analysis was conducted to assess the levels of ALT and ALB, which are indicative of liver distress. The concentrations of both ALT and ALB in the blood were found to be within the normal range and showed no statistically significant differences when compared to the PBS control group (Figure 5GH). These findings collectively indicate that both blank MPs and MPs encapsulating cGAMP are non-toxic and do not adversely affect liver function or cause liver inflammation, which aligns with our previous findings.[24] This conclusion is supported by the absence of significant changes in liver histology and blood chemistry parameters.[63, 64]

Figure 5.

Mice were injected intramuscularly and assessed for toxicity effects 7 days later. Immunohistological images were taken for (A) PBS, (B) blank 20/40/60 MPs, (C) soluble cGAMP, and (D) 20, (E) 40, (F) and 60 CAC cGAMP MPs. Seven days after injection, albumin (ALB) and alanine transferase (ALT) concentrations were analyzed in the sera. Scale bar is drawn to 50 μm. Data is shown as mean±SD.

Scaling up the production of polymeric MPs necessitates a sterilization step to comply with Good Manufacturing Practice (GMP) standards. Common sterilization methods, such as 0.2 μm filtration, ethylene oxide gas, and electron beam (e-beam) treatments, often present challenges, including toxic residues or inadequate material penetration. To address these issues, gamma irradiation was employed, a method proven not to affect the immunogenicity of cGAMP MPs.[24] Spray-dried Ace-DEX cGAMP MPs of three different molecular weights (71k, 500k, and 2M kDa) were sterilized using gamma irradiation at a dose of 25 kGy. Post-irradiation, scanning electron micrographs revealed no structural deterioration of the MPs (Figure 6 AB). However, gamma irradiation did alter the release kinetics of cGAMP, particularly for the 71k and 500k MPs, resulting in an accelerated release of cGAMP (Supplemental Figure 14).

Figure 6.

SEM images (A) 500k, and (B) 2M Ace-DEX MPs after a total of 25 kGy gamma irradiation. (C) BMDCs were treated with 1 μg/mL of cGAMP from each MP for 24 hours and IFN-β concentration was measured in the supernatant via ELISA. Scale bar drawn is 10 μm. Data is shown as mean±SD. * p≤0.05, *** p≤0.001.

This accelerated release led to increased production of IFN-β from BMDCs treated with the irradiated MPs (Figure 6 C). Despite these variations in release profiles and IFN-β production, the study demonstrated that Ace-DEX MPs could be effectively sterilized using gamma irradiation without compromising their adjuvanticity. These findings underscore the viability of gamma irradiation as a sterilization method for Ace-DEX cGAMP MPs, ensuring both safety and functionality in large-scale production settings.

Conclusions

Ace-DEX MPs efficiently encapsulate the STING agonist cGAMP and protein COBRAs. The tunable degradation (CAC) of these polymers significantly impacts humoral, cellular, and protective immune responses in mouse and ferret models of influenza. Changes in the degradation of cGAMP MPs allows modulation of immune responses, with higher CAC Ace-DEX MPs showing longer cGAMP retention and enhanced IgG2a and IL-2 responses in vaccinated mice. The cGAMP MP data reveals that the most significant differences are between the slowest and fastest degrading MPs, with the slowest degrading Ace-DEX having the greatest impact on both innate and humoral immunity, while cellular responses were most influenced by the combination of MPs with different CAC, though not statistically significant. When slow degrading cGAMP MPs are used to adjuvant a trivalent (H1, H3, H5 COBRAs) vaccine in ferrets, the soluble H1 antigen has a marked reduction in HAI compared to the monovalent H1 vaccine. We then hypothesized that adding COBRAs into Ace-DEX MPs of different CACs could affect antigen dominance. Therefore, we encapsulated COBRAs in Ace-DEX MPs with different release rates and evaluated them in vivo in a mouse model. When H3 COBRA (J4) was encapsulated into fast-releasing polymer and COBRA H1 (Y2) was encapsulated in slow degrading polymer, we observed antibody production and cellular responses at or above the antigen alone, thus overcoming the antigen dominance. This approach could improve the immunogenicity and protection of multivalent vaccines.

In addition to illustrating efficacy, our MPs alone or loaded with cGAMP did not illustrate toxic effects on the liver, highlighting their safety in vivo. To facilitate scale-up, post processing sterilization was evaluated. Gamma irradiation sterilization was found to be effective for higher molecular weight Ace-DEX MPs, maintaining their ability to stimulate innate immunity. In conclusion, Ace-DEX MPs encapsulating cGAMP and COBRA HA demonstrate a promising and scalable approach for enhancing immune responses and vaccine efficacy, with effective sterilization methods ensuring safety and stability for large-scale production.

Highlights.

• Ace-DEX MPs deliver COBRA antigens and cGAMP for multistrain coverage

• MP degradation rate impacts antibody and T cell immune responses

• Mixed MPs reduce antigen dominance and improve protection in mice/ferrets