Biomaterials and oxygen join forces to shape the immune response and boost SARS-CoV-2 vaccines

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has led to an unprecedented global health crisis, resulting in a critical need for effective vaccines that generate protective antibodies. Protein subunit vaccines represent a promising approach but often lack the immunogenicity required for strong immune stimulation. To overcome this challenge, we rst demonstrate that advanced biomaterials boost effectiveness of SARS-CoV-2 protein subunit vaccines. Additionally, we report that oxygen is a powerful immunological co-adjuvant, a game-changer in the eld for unlocking the full potential of vaccines. Mice immunized with oxygen-generating cryogel vaccines exhibited a robust and balanced Th1 and Th2 immune response, leading to sustained and high titer production of neutralizing antibodies against SARS-CoV-2. Our data indicate that this platform is a revolutionary technology with the potential to reinforce any vaccine.

One major challenge we anticipated when priming DCs within the subcutaneously injected cryogel is the lack of local vascularization. This environment induces low oxygen tension (hypoxia) due to an imbalance between oxygen supply and consumption (Fig. 1A). Our results indicated that hypoxic conditions of the subcutaneously injected cryogel suppress DC activation, which is consistent with previous reports 15, 16 . Therefore, we engineered oxygen-generating cryogels (O 2 -Cryogels) to mitigate hypoxia-driven immunosuppression 17 . In vivo, the number of hypoxic cells was decreased within subcutaneously injected O 2 -Cryogels compared to standard cryogels (Fig. 1A). In vitro, O 2 -Cryogels restored DC activation by CpG-ODN 1826, increasing the percent of cells positive for activation markers CD86 and CD317 to levels similar to DCs stimulated under normoxic conditions (Fig. 1B). Thus, we proposed that local oxygen supply within cryogel vaccines could potentiate SARS-CoV-2 protein subunit vaccines.

SARS-CoV-2 vaccine fabrication and characterization
Hyaluronic acid-based cryogel vaccines were fabricated by cryogelation, as previously described (Fig. 1C, steps 1-3) 13,18,19 . This process results in an elastic biomaterial with a highly interconnected macroporous network, allowing immune cells to tra c in and out of the cryogel. To fabricate cryogelbased SARS-CoV-2 vaccines, we incorporated both the nucleocapsid protein (N), which encapsulates the virus RNA and the receptor-binding domain (RBD) of the spike protein (S) that is responsible for virus entry into cells. The cryogels also contained granulocyte macrophage colony-stimulating factor (GM-CSF), a molecule that stimulates and promotes recruitment of various immune cells, including DCs 18,20 , and the adjuvant CpG-ODN 1826, a ligand for TLR9 (toll-like receptor 9) that recruits and activates DCs, speci cally plasmacytoid DCs (pDCs) [21][22][23] . These cryogel-based vaccines were formulated to induce a robust humoral immune response 23,24 . To enhance vaccine immunogenicity, oxygen was considered as an immunological co-adjuvant that would eliminate local hypoxia at the site of vaccine administration 25 . Thus, oxygen-producing calcium peroxide (CaO 2 ) particles and acrylate-PEG-catalase (APC) were incorporated within the cryogel vaccine formulations before freezing, as previously described 17 . The resulting O 2 -Cryogel vaccine was designed to generate oxygen upon the reaction of CaO 2 with water and to eliminate hydrogen peroxide byproducts through a catalase-mediated breakdown.
Following subcutaneous immunization (Fig. 1C), we hypothesized that cryogel-based protein subunit vaccines would induce DC-mediated humoral immunity (Fig. 1D) 18 . Sustained release of immunomodulatory factors (GM-CSF and CpG-ODN 1826) promotes DC in ltration into the macroporous network of cryogels, wherein DCs would uptake protein antigens (N and RBD proteins) and simultaneously become activated by CpG-ODN 1826 and the increased local oxygen tension. Activated, antigen-loaded DCs would migrate to draining LNs to initiate the activation of antigen-speci c T cells and B cells. Furthermore, protein antigens and adjuvants released from the cryogel would also drain to the draining LNs, directly enhancing B cell and DC activation. A subset of activated B cells would differentiate into plasma cells with the primary role of producing large quantities of SARS-CoV-2-binding antibodies. A fraction of these antibodies would be neutralizing antibodies and exert their inhibitory activity by abrogating binding of the virus RBD to the human receptor angiotensin-converting enzyme 2 (ACE2).
The encapsulation of RBD and N proteins within O 2 -Cryogel VAX polymer walls was characterized by confocal microscopy and release from the cryogel by ELISA (Fig. S1A-B). Both proteins were effectively entrapped and colocalized within the polymer network. Both Cryogel VAX and O 2 -Cryogel VAX exhibited an initial burst release of their payload, followed by a sustained release of the immunomodulatory factors GM-CSF and CpG-ODN 1826 and the antigen RBD (Fig. S1B). Notably, there were no differences among the encapsulation (Fig. S1C)  To test the vaccines, eight-week-old female BALB/c mice were immunized by subcutaneous injection of two O 2 -Cryogel VAX or Cryogel VAX (one on each ank) at day 0 (prime) and day 21 (boost) ( Fig. 2A).
Control groups were injected with either PBS (sham-negative control), cryogel-free vaccine (Bolus VAX ), or Freund's-based vaccine (Freund VAX -positive control) ( Table 1). Blood serum analysis revealed that, although low titers of immunoglobulin M (IgM) antibodies were found across all groups (Fig. S2A), Cryogel VAX and O 2 -Cryogel VAX induced high titers of RBD-speci c binding immunoglobulin G (IgG) antibodies after only 21 days ( Fig. 2B and Fig. S2B). These titers increased substantially following boost immunization, peaking at 1.4 × 10 6 at day 42 for animals immunized with Cryogel VAX and 3.1 × 10 6 at day 56 for animals immunized with O 2 -Cryogel VAX , amounts two orders of magnitude greater than those in control groups. Interestingly, O 2 -Cryogel VAX induced higher production of RBD-speci c binding IgG antibodies than Cryogel VAX did, and these titers were sustained for nearly 2 months (study endpoint).
Similarly, O 2 -Cryogel VAX immunization resulted in high titers of N-speci c binding IgG antibodies comparable to those induced by Freund VAX , and 3-and 5-times higher than those generated by Cryogel VAX and Bolus VAX, respectively.  (Fig. 2C, upper). Importantly, 1.7% of O 2 -Cryogel VAX -induced anti-RBD IgG antibodies were neutralizing from day 21 onward (Fig. 3C, lower). We also assessed the neutralization potency of antibodies by plaque reduction neutralization test (PRNT) using VeroE6 cells infected with authentic SARS-CoV-2 ( Fig. 2D) 26 . As expected, O 2 -Cryogel VAX immunization led to high neutralizing titers, which intensi ed from day 21, reaching a reciprocal IC 50 value of nearly 10,000 at day 56 (study endpoint).
Collectively, these data demonstrated that the cryogel platform potentiates vaccine e cacy. Furthermore, additive oxygen as a co-adjuvant strongly boosted the humoral response, as shown by the production of antibodies with high neutralizing activity. O 2 -Cryogel VAX promotes local immune cell recruitment and B cell production in LNs To understand how the vaccines work, we characterized the immune response following prime and primeboost immunizations in mice. At day 21 and 56, draining LNs, spleens, and cryogels were explanted ( Fig. 2A). In comparison to the injection sites of Cryogel VAX , sites of both prime and boost O 2 -Cryogel VAX injections were markedly enlarged, indicating increased in ammation and immune cell in ltration (Fig. 3A). Overall, unlike blank cryogels, large numbers of in ltrated immune cells were retrieved from both types of cryogel vaccines ( Fig. 3B and Fig. S3A, S3B). Most explanted cryogels exhibited low and comparable numbers of CD4 + and CD8 + T cells, whereas high numbers of CD11b-positive myeloid cells, but no DCs, were present (Fig. 3B). Furthermore, the total number of cells positive for the B cell marker CD19 was 2-fold higher in O 2 -Cryogel VAX compared to those in blank cryogels (Fig. S3B). However, the exact identity of these cells is unclear, as they were also CD11b-positive and did not have other B cell markers such as MHCII (Fig. S3A). Additionally, only a small population of MHCII-positive CD11b + cells was observed. Interestingly, evidence for an ongoing adaptive immune response was found in a small number of cryogel vaccines. These cryogel vaccines contained a lower fraction of CD11b + myeloid cells but relatively greater proportions of T cells and MHCII + B cells (see the outliers in Fig. 3B and Fig. S3B). Compared to day 21, immune cell numbers in prime cryogel vaccines decreased at day 56, indicating that both types of cryogel vaccines do not generate chronic and potentially dangerous in ammatory responses. However, cell counts in boost O 2 -Cryogel VAX were slightly, but signi cantly, increased compared to other cryogel vaccines at day 56.
Analysis of LNs in mice immunized with Cryogel VAX and O 2 -Cryogel VAX con rmed that a robust immune response was induced. This resulted in at least a 4-fold greater increase in total immune cell numbers than that observed among mice receiving sham injections at both time points ( Fig. 3C and Fig. S4A, S4B).
In particular, the frequency of MHCII + B cells within LNs was greatly increased in mice immunized with both cryogel-based vaccines. Although the frequency of CD4 + T cells was reduced at day 21 in LNs from mice receiving cryogel vaccines (Fig. 3C), overall CD4 + and CD8 + T cell numbers increased after vaccination (Fig. S4B). These data showed that cryogel vaccines induce a strong B cell-mediated immune response in LNs and display restrained adaptive immune responses within the cryogels following initial priming.
O 2 -Cryogel VAX induces balanced Th1/Th2-associated immune responses Previous reports with experimental SARS-CoV vaccines [27][28][29] indicated that an unbalanced T helper 1 (Th1) and T helper type 2 (Th2) immune responses, especially when biased towards Th2, is associated with poor clinical outcomes in infected patients 30,31 due to vaccine-associated enhanced respiratory disease (VAERD). Therefore, we analyzed the balance between Th1 and Th2 immune responses. Production of antibody subclass IgG1 is indicative of Th2 responses, and IgG2a/b/c and IgG3 are indicative of Th1 responses 32 . In our study, vaccines across all groups elicited IgG2 and IgG1 subclass RBD-binding antibodies, indicating induction of both Th1 and Th2 immune responses (Fig. 4A). Both cryogel-based vaccines promoted the production of IgG2b. However, O 2 -Cryogel VAX enhanced Th2-biased antibody production as high titers of IgG1 were found in immunized mice resulting in lower IgG2a/IgG1 and IgG2b/IgG1 ratios (Fig. 4B). Interestingly, O 2 -Cryogel VAX was the only vaccine that induced IgG3 production, which is associated with Th1 responses (Fig. 4A-B).
Higher concentrations of Th1 cytokines IFNγ, TNFα, and IL-2, as well as IL-6, were quanti ed in mice immunized with O 2 -Cryogel VAX , compared to those immunized with Cryogel VAX . Furthermore, we noted low concentrations of the Th2 cytokine IL-13 in O 2 -Cryogel VAX . As expected, blank cryogels were associated with low or negligible amounts of these cytokines.
To more directly assess the Th1/Th2 balance, we investigated the cytokine pro le of antigen-speci c T cells generated with both cryogel vaccines. The intracellular production of cytokines by splenocytes from immunized mice was examined following stimulation with peptides derived from viral S or N proteins. Cells were isolated at day 21 after prime immunization. Splenocytes from O 2 -Cryogel VAX -immunized mice stimulated with N-derived peptides showed increased fractions of IL-5-producing CD4 + and CD8 + T cells and IL-13-producing CD4 + T cells ( Fig. 4E and g. S5). These results indicated the presence of N proteinspeci c Th2 cells. However, no differences were noted following stimulation with S-derived peptides, and the proportions of IFNγ-, IL-4-, or IL-17-producing T cells were also comparably low. Collectively, these data suggested that both types of cryogel vaccines elicited a balanced Th1/Th2 immune response, even though it was more prominent for O 2 -Cryogel VAX .

Discussion
Nearly every decade for the past 30 years, a novel coronavirus pandemic emerges, pushing the healthcare system to its limit 33 . Although the current outbreak had long been predicted, SARS-CoV-2 has created the most severe crisis in recent history 34,35 . The rapid development of an effective and safe vaccine against this virus is the most effective strategy to end this pandemic. Among them, protein subunit vaccines have been widely investigated against SARS-CoV-2 due to their performance and safety record, and such vaccines have already shown promising early results in phase 1/2 clinical trials 11 . Yet, subunit vaccines still have to overcome their poor immunogenicity 36 . In addition, an ideal antiviral vaccine should be versatile and rapid to design, enabling rapid response to the public health emergency. To overcome these challenges, our team leveraged a cryogel-based vaccine platform to strengthen protein subunit vaccines and induce a strong and sustained anti-SARS-CoV-2 immune response 18 . In addition, we showed that oxygen is a powerful immunological co-adjuvant that shapes and reinforces the immune response. This work demonstrated how robust and modular the cryogel-based vaccine technology is, which was successfully and quickly adapted from cancer to an infectious disease at breakneck speed (< 3 months).
We found that Cryogel VAX induces a balanced Th1/Th2 immune response while enhancing the e cacy of a conventional protein subunit vaccine by 100-fold (Bolus VAX ). This is most likely due to the ability of cryogels to control the release of immunomodulatory factors while activating high numbers of resident immune cells. Following prime-boost immunization, Cryogel VAX elicited a strong humoral immune response for nearly 2 months (study endpoint) and was associated with high levels of anti-RBD IgG antibodies and strong neutralizing activity to SARS-CoV-2. In addition, Cryogel VAX induced CD4 + and In summary, our study unveils the magnitude of an advanced biomaterial-based technology to harness the power of protein subunit vaccines, leading to a rapid and protective anti-SARS-CoV-2 immune response. Additionally, we report the synergistic effect of vaccines engineered to provide oxygen as a powerful immunological co-adjuvant. Lastly, although our efforts focused on protein subunits, this platform is compatible with other strategies, such as live attenuated or inactivated pathogens and nucleic acid vaccines, and may boost the e ciency of existing vaccines or those under development. Because vaccinologists believe the rst generation of the vaccines currently in clinical trials might not be optimal, our study opens new possibilities to leverage the COVID-19 vaccines that are in clinical trials and develop improved versions.

BMDC generation for DC activation studies
Dendritic cells (DC) activation studies were performed using bone marrow-derived dendritic cells (BMDCs) generated from 6-8-week-old female C57BL/6 mice (Charles River) as previously described 18 . Brie y, femurs of mice were explanted, disinfected in 70% ethanol for 5 min, washed in DPBS, and then bone ends were removed, and the marrow ushed with DPBS (2 mL, 27G needle). Next, cells were mechanically dissociated by pipetting, centrifuged (5 min, 300 g), and resuspended (10 6  were seeded in non-treated p6 well plates (2 × 10 6 cells per well) in 5 mL of complete RPMI medium supplemented with 20 ng/mL mGM-CSF. At day 3, another 5 mL of RPMI medium containing 20 ng/mL mGM-CSF was added to each well. At days 6 and 8, half of the media was sampled from each well, centrifuged, and the cell pellet was resuspended in 5 mL of fresh RPMI media supplemented with only 10 ng/mL mGM-CSF before re-plating. BMDCs were collected at day 10 (non-adherent cells) and used to evaluate DC activation in normoxia or hypoxia.
Imaging of encapsulated N and RBD proteins within the cryogel network RBD or N protein was dissolved in sodium bicarbonate buffer (pH 8.5) at 0.5 mg/mL and reacted with Alexa Fluor 488-NHS ester or Alexa Fluor 647 NHS ester (Click Chemistry Tools), respectively, for 2 h at 4 °C. Fluorochrome-modi ed proteins were puri ed via spin ltration over 10 kDa Amicon Spin Filters (Sigma Aldrich) and washed 5 times with DPBS. Concentration of puri ed proteins was determined by UV-Vis absorbance measurements at 280 nm, after correcting for uorophore absorbance, using the Nanodrop One (ThermoFisher). O 2 -Cryogels containing the uorescently labeled RBD and N proteins were fabricated as described above. After thawing, cryogels were washed four times with 1 mL of DPBS and imaged by confocal microscopy (Zeiss 800).

Release of immunomodulatory factors and antigens from cryogels
To determine the in vitro release kinetics of GM-CSF, CpG-ODN, and RBD protein from Cryogel VAX and O 2 -Cryogel VAX , gels were brie y washed in 70% ethanol followed by 2 DPBS washes. Each washed gel was incubated in sterile DPBS with 2% BSA in a microcentrifuge tube under orbital shaking at RT. The entire supernatant was removed periodically and replaced with the same amount of fresh buffer. GM-CSF, CpG-ODN, and RBD protein released in the supernatant were detected by either ELISA (GM-CSF: BioLegend ELISA MAX™ Deluxe, RBD: Elabscience SARS-CoV-2 Spike Protein S1 RBD ELISA Kit) or iQuant™ ssDNA quanti cation assay (GeneCopoeia). The N protein release kinetics was not determined due to the instability of the protein at high concentration, buffer, and study duration.
Antibody titration by enzyme-linked immunosorbent assay (ELISA) Anti-RBD IgG and IgM antibody titers were determined using a SARS-CoV-2 Spike S1-RBD IgG&IgM ELISA detection kit (Genscript). Anti-N IgG and IgM antibody titers were determined using a SARS-CoV-2 Nucleocapsid Protein IgG ELISA Kit (Lifeome). Both kits were optimized by replacing the HRP-conjugated IgG or IgM anti-human antibody with an HRP-conjugated IgG (H + L) goat anti-mouse antibody (FisherScienti c) or an HRP-conjugated IgM (Heavy chain) goat anti-mouse antibody (FisherScienti c), respectively. Immunoglobulin isotyping was evaluated using Ig Isotyping Mouse Uncoated ELISA Kit (ThermoFisher) following the manufacturer's recommendation by measuring absorbance at 450 nm on a plate reader (Synergy HT). All ELISAs were performed on mouse sera that were heat-inactivated (30 min at 56 °C). Endpoint titers were determined as the maximum dilution that emitted an optical density exceeding 4 times the background (sera of mice vaccinated with Sham vaccine).

SARS-CoV-2 surrogate virus neutralization test (sVNT)
The detection of neutralizing antibodies against SARS-CoV-2 that block the interaction between RBD and the human ACE2 (hACE2) cell surface receptor was determined using an sVNT according to the manufacturer's protocol (Genscript). Brie y, heat-inactivated mouse sera were pre-incubated with HRP-RBD (30 min at 37 °C) to allow the speci c binding of neutralizing antibodies. Then, the mixture was transferred into a plate coated with hACE2 and incubated for 15 min at 37 °C. The unbound HRP-RBD, as well as HRP-RBD bound to non-neutralizing antibody, will interact with the hACE2, while neutralizing antibody-HRP-RBD complexes will remain in suspension and will be removed during washing. TMB substrate was used to detect the non-neutralized HRP-RBD. Therefore, the absorbance was inversely proportional to the titer of anti-SARS-CoV-2 neutralizing antibodies. For this experiment, 10-fold dilutions of mouse sera (10 − 1 to 10 − 8 ) were used.

Cytokine quanti cation
Cytokine levels in mouse sera and cryogels were quanti ed using LEGENDplex™ mouse Th cytokine panel

Statistical analysis
Flow cytometry data were analyzed using FlowJo software. Gating was done as depicted in Fig. S3A

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.

Data Availability
All data are available in the manuscript or the supplementary materials. Overview of the process for fabrication and evaluation of cubiform cryogel and O2-Cryogel vaccines.
Step 1 involves freezing vaccine components, enabling crosslinking of solutes around ice crystals (step 2).
Thawing results in an interconnected macroporous network with vaccine components encapsulated within the polymer network (step 3). Addition of calcium peroxide and catalase to the vaccine components before cryogelation produces O2-CryogelVAX capable of sustained production of oxygen. In step 4, cryogels are subcutaneously injected into mice for preclinical vaccine studies. (D) Illustration describing a model for DC-mediated cryogel-induced immunity. HAGM: hyaluronic acid glycidyl methacrylate; initiator: ammonium persulfate and tetramethylethylenediamine. Overview of the process for fabrication and evaluation of cubiform cryogel and O2-Cryogel vaccines.
Step 1 involves freezing vaccine components, enabling crosslinking of solutes around ice crystals (step 2).
Thawing results in an interconnected macroporous network with vaccine components encapsulated within the polymer network (step 3). Addition of calcium peroxide and catalase to the vaccine components before cryogelation produces O2-CryogelVAX capable of sustained production of oxygen. In step 4, cryogels are subcutaneously injected into mice for preclinical vaccine studies. (D) Illustration describing a model for DC-mediated cryogel-induced immunity. HAGM: hyaluronic acid glycidyl methacrylate; initiator: ammonium persulfate and tetramethylethylenediamine. CryogelVAX-treated mice were tested after prime (day 21: D21) and prime-boost (day 56: D56) immunizations. Data points show individual serum sample, and data is represented as mean ± SEM. *P < 0.05, **P < 0.01 and ***P < 0.001.     were assessed by ELISA. Th1 and Th2 cytokine levels were measured in mouse serum at day 24 (C) and