Biodegradable scaffolds for enhancing vaccine delivery

Abstract Sustained release of vaccine components is a potential method to boost efficacy compared with traditional bolus injection. Here, we show that a biodegradable hyaluronic acid (HA)‐scaffold, termed HA cryogel, mediates sustained antigen and adjuvant release in vivo leading to a durable immune response. Delivery from subcutaneously injected HA cryogels was assessed and a formulation which enhanced the immune response while minimizing the inflammation associated with the foreign body response was identified, termed CpG‐OVA‐HAC2. Dose escalation studies with CpG‐OVA‐HAC2 demonstrated that both the antibody and T cell responses were dose‐dependent and influenced by the competency of neutrophils to perform oxidative burst. In immunodeficient post‐hematopoietic stem cell transplanted mice, immunization with CpG‐OVA‐HAC2 elicited a strong antibody response, three orders of magnitude higher than dose‐matched bolus injection. In a melanoma model, CpG‐OVA‐HAC2 induced dose‐responsive prophylactic protection, slowing the tumor growth rate and enhancing overall survival. Upon rechallenge, none of the mice developed new tumors suggesting the development of robust immunological memory and long‐lasting protection against repeat infections. CpG‐OVA‐HAC2 also enhanced survival in mice with established tumors. The results from this work support the potential for CpG‐OVA‐HAC2 to enhance vaccine delivery.


| INTRODUCTION
Biomaterial-based vaccines have been used to address challenges associated with conventional bolus vaccination strategies.2][3][4][5][6][7][8] These studies have strongly supported a key role of sustained release of vaccine components in enhancing the immune response.While effective at inducing an immune response, the formulation itself could include components that could result in a persistent foreign body response (FBR), such as those that use long-lasting polymers, 2,9,10 that could be limiting in clinical settings. 9,11On the other hand, a degradable biomaterial-based vaccine which maximizes activation of the adaptive immune responses while avoiding a long-lasting FBR could be a potential alternative.Hyaluronic acid (HA) is a polysaccharide-based polymer that is abundant in tissues including skin, cartilage, and synovial fluid.It has been widely studied in biomedical applications including drug delivery as HA can be modified to form a matrix to encapsulate drugs, such as growth factors and chemotherapeutic agents, and release them in a sustained manner via controlled degradation. 12We have previously demonstrated that the degradation of HA can be immune-responsive, mediated by oxidative burst of neutrophils, 12 and is also processed by endogenous hyaluronidases 13,14 (HYAL).[6] Here, we hypothesized that a HA-based vaccine formulation would enhance immune cell activation and sustain release of encapsulated vaccine components to generate a durable protective immune response.We had previously developed a macroporous injectable HA hydrogel, termed HA cryogel, made with bioorthogonal crosslinking chemistry. 12HA cryogels mediated sustained release of protein therapeutics to enhance innate immune cell regeneration. 12In this work, we extend our previous findings to further develop HA cryogels as depots for sustained release of vaccine components.After evaluating cryogel formulations generated from commercially sourced HA by immunophenotypic and histological assessments, we selected low endotoxin HA, termed HAC2 for evaluation with encapsulated model antigen OVA, adjuvants CpG-ODN 1826 (CpG) and granulocyte macrophage colony stimulating factor (GM-CSF).Both CpG and GM-CSF comparably enhanced the response to the vaccine, but GM-CSF increased inflammation and the FBR at the injection site.We selected CpG-loaded OVA-HAC2 (CpG-OVA-HAC2) for dose ranging studies.
The intensity of the adaptive immune response was dose dependent on CpG-OVA-HAC2 and significantly delayed in settings of neutrophil dysfunction.In post-hematopoietic stem cell transplant (HSCT) mice, CpG-OVA-HAC2 enhanced antibody induction three orders of magnitude greater than dose-matched bolus vaccination.In a B16-OVA melanoma mouse model, prophylactic and therapeutic administration of CpG-OVA-HAC2 slowed the tumor growth rate and enhanced overall survival.
Vaccine components were solubilized with the polymer solution prior to mixing and overnight incubation at À20 C to form HA cryogel vaccines (Figure 1a).HA was initially sourced from two commercial suppliers and resulting HA cryogels are referred to as HAC1 and HAC2 for HA sourced from supplier 1 and 2, respectively.For confocal imaging and in vivo degradation tracking, HA-Nb was reacted with Tz-Cy5 to form Cy5-labeled HA-Nb (Cy5-HA-Nb) which was mixed with HA-Tz to form Cy5-labeled HA cryogels, referred to as HAC1:Cy5 and HAC2: Cy5 (Figure S1a).To visualize pores and assess pore interconnectedness, HAC1:Cy5 and HAC2:Cy5 were incubated with fluorescein isothiocyanate (FITC)-labeled 10 μm diameter melamine resin particles and imaged using confocal microscopy.FITC-particles were co-localized with Cy5-labeled HA polymer to the confocal depth limit for HAC1:Cy5 and HAC2:Cy5 (Figure 1b).Next, we lyophilized HA cryogels and measured surface porosity using scanning electron microscopy (SEM; Figure 1c).
Average pore diameter for HAC1 and HAC2 were found to be 59.5 We next sought to quantify the effect of degradation on OVA release.In vitro release assays for OVA-HAC1 and OVA-HAC2 were conducted with or without HYAL2 solution in phosphate buffered saline (PBS).In HYAL2 solution, 88.8% ± 6.6% and 86.3% ± 6.8% of OVA released from OVA-HAC1 and OVA-HAC2, respectively, over the first day with the remaining OVA released throughout the course of the study until the gels were fully degraded (Figure 1g).In PBS, 22.2% ± 2.8% and 23.0% ± 0.7% of OVA burst release from OVA-HAC1 and OVA-HAC2, respectively, within the first day with minimal release over the rest of the study (Figure 1g).The encapsulation efficiency of OVA, based on in vitro release, was 77.8% for OVA-HAC1 and 77.0% for OVA-HAC2.To determine in vivo OVA release kinetics, OVA was functionalized with Cy5 (OVA:Cy5) prior to encapsulation within HAC1 (OVA:Cy5-HAC1) and HAC2 (OVA: Cy5-HAC2) and measured using IVIS Release was quantified by measuring the attenuation of the fluorescence signal relative to the initial measurement.The release of OVA:Cy5 from HAC1 and HAC2 was sustained over multiple weeks with accelerated OVA:Cy5 release from HAC1 as compared with HAC2 (Figure 1h,i).
Next, we sought to determine whether the aforementioned differences in innate immune cell infiltration and anti-OVA antibody titers might be due to differences in endotoxin content.We added LPS, the major constituent of endotoxin, in HA-Tz and Cy5-HA-Nb from to generate OVA-HAC2 with endotoxin content of 5. ) and OVA:Cy5-HAC2.(i) Measuring OVA:Cy5 release from HAC1 and HAC2 by quantification of total radiant efficiency normalized to initial 6-h timepoint.Data in d, g represents mean ± SD of n = 4 cryogels.Data in f represents mean ± SEM of n = 5 mice.Data in i represents mean ± SEM of n = 4 mice.Data in (d,f,g,i) compared using two-way ANOVA with Bonferroni multiple comparison test.In g comparison of PBS and HYAL2 release was conducted by pooling measurements for OVA-HAC1 and OVA-HAC2.Figure 1a  and 800% of OVA-HAC1 endotoxin content (Table S1).Innate immune cell infiltration into OVA-HAC2 with LPS was assessed 7-days postinjection and compared with OVA-HAC2 without added LPS.Inclusion of LPS had no effect on total myeloid, neutrophil, monocyte, macrophage, or DC infiltration in OVA-HAC2 (Figure S2c-h).Mice were administered OVA-HAC2, low-LPS OVA-HAC2, and high-LPS OVA-HAC2 in a prime and boost setting 11-days apart and were bled at pre-determined timepoints.No significant differences were quantified with anti-OVA IgG1 antibody titers between any of the test groups (Figure S2i).Based on these results, we selected HAC2 for further assessment.

| Encapsulation of adjuvants enhances HA cryogel-based vaccine efficacy and alters foreign body response
3][24] GM-CSF and CpG were encapsulated in OVA-HAC2 (GMCSF-CpG-OVA-HAC2) and in vitro release studies were conducted with or without HYAL2 solution in PBS.In HYAL2 solution, 88.7% ± 4.5% and 97.1% ± 0.4% of CpG and GM-CSF, respectively, released from OVA-HAC2 over the first day with the remaining release throughout the course of the study with cryogel degradation (Figure S3a,b).In PBS, 53.8% ± 7.1% and 15.8% ± 3.6% of CpG and GM-CSF, respectively, burst release from the cryogels over the course of the first day with minimal release over the rest of the study (Figure S3a,b).From the release in PBS, encapsulation efficiency was calculated to be 46.2% and 84.2% for CpG and GM-CSF, respectively.
Monocytes were lower in all adjuvanted OVA-HAC2 formulations compared with OVA-HAC2.However, GMCSF-CpG-OVA-HAC2 had the lowest number of monocytes (Figure 4e).Macrophage infiltration was modestly lower in GMCSF-CpG-OVA-HAC2, but similar in the other formulations (Figure 4f).DC infiltration was similar between the test groups and constituted <3% of total cellular infiltrates in all formulations (Figure 4c,g).In the cryogel formulations removed 21-days after injection, live cell percentages were >75% on average in all groups.Myeloid cell infiltration and the constituent cells of the infiltrates were similar between groups (Figure S4b-h).
HAC2, OVA-HAC2, and adjuvanted OVA-HAC2 formulations were administered in mice and removed 10-days post injection for histomorphometry-based H&E staining.All formulations were infiltrated with cells and encased in a fibrotic capsule (Figure 4h).The cellularity was greater the OVA-HAC2 vaccine formulations which included GM-CSF as a constituent (Figure 4i).However, only mice which received GMCSF-CpG-OVA-HAC2 had increased capsule thickness (Figure 4j).
Initial immunization studies were conducted without HA in which mice received a single prime and boost bolus subcutaneous injection of vaccine components 11-days apart.Mice received either (1) OVA only, (2) OVA + GM-CSF, (3) OVA + CpG, or (4) OVA + GM-CSF + CpG and periodically bled post-prime to assess anti-OVA IgG1 antibody concentrations (Figure S5a).OVA + GM-CSF and OVA + GM-CSF + CpG enhanced antibody titers compared with OVA only (Figure S5b).OVA + CpG generated a minimal antibody response, comparable to mice which received OVA only (Figure S5b).
Next, the aforementioned OVA-HAC2 vaccine formulations were administered to mice in a prime and boost 11-days apart.Assessments of anti-OVA IgG1 antibody titers showed a significant difference between OVA-HAC2 and all other adjuvanted formulations (Figure 4k).
There was no difference in antibody titers between any adjuvanted OVA-HAC2 formulations (Figure 4k).The same cohorts of mice were used to measure degradation of the OVA-HAC2:Cy5 formulations using IVIS.In a separate study, mice were administered prime and boost adjuvanted OVA-HAC2 and OVA-HAC2 only and sacrificed 3-weeks post prime to assess CD45 + B220 À CD8 + SIINFEKL + cells (OVA-specific cytotoxic T cell) in both the draining axillary lymph nodes (LNs) and spleen (Figure S5c,d).The test groups and unvaccinated mice had comparable antigen-specific cytotoxic T cells in the lymph node and spleens (Figure S5e,f).
We concluded that all adjuvanted OVA-HAC2 vaccine formulations comparably enhanced antigen-specific adaptive immune response.However, formulations that included GM-CSF were associated with an exacerbated the foreign body response.Therefore, we selected CpG-OVA-HAC2 for further studies.

| CpG-OVA-HAC2 induces a dose dependent adaptive immune response
We evaluated the release of CpG from CpG-OVA-HAC2 in vivo.
CpG:Cy5 release was sustained over a period of 3 weeks with 85.0% ± 3.1% of release occurring within the first week (Figure 5a).The effect of vaccine dose escalation on the adaptive immune response was assessed by increasing the number of CpG-OVA-HAC2 administered.Mice received either a single dose prime, two-dose prime, single dose prime and boost, or two-dose prime and boost.Assessment of anti-OVA IgG1 antibody titers showed that a single prime vaccination induced significantly lower antibody titers than mice which received higher doses (Figure 5b).Notably, there was no difference in anti-OVA IgG1 antibody titers in mice which received any of the other higher vaccine doses (Figure 5b).CD45 + B220 À CD8 + SIINFEKL + cells (OVA-specific cytotoxic T cells) were assessed 3-weeks post-prime in the draining axillary lymph nodes (LNs) (Figure 5c).Only mice which received two CpG-OVA-HAC2 as both a prime and boost had a significantly elevated percentage of OVA-specific cytotoxic T cells in axillary LN compared with other groups (Figure 5d).There was no difference in OVA-specific cytotoxic T cell percentage in any of the groups in the spleen 3-weeks post prime (Figure S6).To better study the effect of polymer degradation and the development of an adaptive immune response, we utilized gp91 phoxÀ mice which have neutrophils that lack the ability to perform oxidative burst. 12,25,26We have also previously demonstrated that HA cryogel degradation in this model is impaired. 12Two CpG-OVA-HAC2:Cy5 were administered in gp91 phoxÀ or B6 mice as a prime and boost 11-days apart.IVIS measurements on mice showed minimal degradation of CpG-OVA-HAC2:Cy5 in gp91 phoxÀ mice and complete degradation in B6 mice (Figure 5e).The same cohorts of mice were bled at predetermined timepoints for assessment of anti-OVA IgG1 antibody titers.There was a significant delay in the development of anti-OVA IgG1 antibodies in gp91 phoxÀ mice with none having detectable anti-OVA IgG1 antibody titers two-weeks post prime (Figure 5f).2-days later (Figure 5g).Two CpG-OVA-HAC2:Cy5 or bolus CpG + OVA were administered to mice in a prime and boost 11-days apart and bled at predetermined timepoints (Figure 5g).Mice that received CpG-OVA-HAC2 had measurable anti-OVA IgG1 antibody titers whereas there were no detectable anti-OVA IgG1 antibodies in mice which received bolus vaccination (Figure 5h).The immunization study was repeated in the context of post-allogenic HSCT, with BALB/cJ donor mice and B6 recipients.All the mice in this study succumbed to graft-versus-host disease-like pathology within 5-weeks post-HSCT.
Blood was collected once at 2-weeks post prime, 4-weeks post-HSCT.There were no detectable anti-OVA IgG1 antibodies in either group.

| CpG-OVA-HAC2 mediates prophylactic protection in a B16-OVA melanoma mouse model
We sought to assess whether CpG-OVA-HAC2 enhanced prophylactic protection in a mouse melanoma model.Mice were administered either one or two CpG-OVA-HAC2 or two bolus CpG + OVA injections in a prime and boost setting.Mice received 100,000 B16-OVA cells administered subcutaneously 3-weeks post prime or in unvaccinated mice and tumor growth rate and survival were compared (Figure 6a).In all unvaccinated mice, tumors were visible 2-weeks post tumor inoculation with rapid tumor growth thereafter (Figure 6b,c).In mice that received two bolus CpG + OVA injections, tumor growth was delayed and were visible in all mice 3-weeks post tumor inoculation (Figure 6b).In contrast, only 60% of the mice that received a single CpG-OVA-HAC2 as prime and boost had visible tumors.A total of 40% of the mice that received two CpG-OVA-HAC2 as prime and boost each had visible tumors 4-weeks post tumor inoculation (Figure 6b).In all vaccinated mice, tumor growth rate was significantly reduced and increased the mean survival time (Figure 6c).Overall survival was dose responsive to CpG-OVA-HAC2.Mice that received two CpG-OVA-HAC2 as a prime and boost had the highest overall survival of 60% (Figure 6d).Mice that received a single CpG-OVA-HAC2 had an overall survival of 40% whereas mice receiving bolus vaccination and unvaccinated mice all succumb to B16-OVA melanoma (Figure 6d).Vaccinated mice were bled 6-weeks post prime, 3-weeks post tumor inoculation, and mice which received bolus CpG-OVA-HAC2 had lower anti-OVA IgG1 antibody titers than mice which received CpG-OVA-HAC2 (Figure 6e).Mice which survived B16-OVA melanoma challenge were re-challenged with 100K B16-OVA cells three-months post initial inoculation.All mice survived rechallenge without visible tumors for at least 40 days post re-challenge.Next, we assessed whether CpG-OVA-HAC2 would provide therapeutic protection against murine melanoma.Mice were administered 100,000 B16-OVA cells administered subcutaneously 3-days prior to one or two CpG-OVA-HAC2 in a prime and boost setting (Figure 6f).CpG-OVA-HAC2 administered therapeutically slowed tumor growth and prolonged survival as compared with unvaccinated mice (Figure 6g,h).
Mice that received a single CpG-OVA-HAC2 as a prime and boost survived 45.9 ± 13.8 days and mice that received two CpG-OVA-HAC2 as a prime and boost survived 45.5 ± 12.0 days post-tumor challenge.

| DISCUSSION
Here, we demonstrate that a hyaluronic acid cryogel-based biodegradable vaccine mediates sustained release of vaccine components, enhances antigen-specific adaptive immunity in healthy and immunodeficient mice, and provides protection against B16-OVA melanoma in a vaccine dose-dependent manner.The source of material and selection of adjuvants had a significant impact on the infiltrating innate immune cell subsets that constitute the FBR.We demonstrate that CpG-OVA-HAC2 formulation enhances anti-OVA adaptive immune response without significantly altering the FBR and that the adaptive immune response was enhanced with escalating vaccine doses.The effect of high-dose Cpg-OVA-HAC2 was studied in immune deficient models.In gp91 phoxÀ mice, the degradation of CpG-OVA-HAC2 was delayed which resulted in slowed activation of the adaptive immunity.
In post-HSCT mice, adaptive immune activation with CpG-OVA-HAC2 was greatly enhanced as compared with bolus CpG + OVA vaccination, highlighting the enhancement in adaptive immune response even in immunodeficient contexts.Lastly, enhancement in immune response was assessed in both prophylactic and therapeutic vaccine studies involving B16-OVA melanoma, in which the CpG-OVA-HAC2 slowed tumor growth and enhanced survival relative to unvaccinated mice and mice receiving bolus vaccination.
HA was selected as the polymer as it is ubiquitous in the ECM and has a long history of use as a biodegradable material to effectively facilitate sustained drug delivery.As an ECM component, HA provides cues to regulate inflammation and tissue repair. 27In the context of tissue damage or injury, HA undergoes degradation resulting in the production of low molecular weight fragments. 28,290][31][32] The downstream signaling of TLR activation results in secretion of TNF-α, IL-12, and IL-1β which promote inflammation. 33,34Additionally, HA fragments have been shown to induce expression of matrix metalloproteinases, which degrade ECM components and facilitate tissue remodeling. 35,36Thus, the use of HA in biomaterials-based vaccine not only provides sustained release of vaccine components but also potentially offers additional cues to regulate inflammation.5][6] Consistent with prior work, our results showing that HA cryogel encapsulating OVA alone elicits anti-OVA IgG1 antibody titers without an additional adjuvant. 4eking to assess the effect of HA from different sources, we tested the effect of formulating HA cryogels from two commercial vendors.HA cryogels from both vendors had comparable pore size and morphology, in vitro degradation kinetics, and in vitro OVA release kinetics.There were differences in endotoxin levels, as determined by measuring LPS content, with HAC1 having 10.8-fold higher endotoxin levels than HAC2.However, the endotoxin levels of HAC1 and HAC2 were well below FDA guidelines for implantable devices. 37,38When cryogels were administered, OVA-HAC1 induced higher anti-OVA IgG1 titers than OVA-HAC2 but was associated with a neutrophil dominated FBR with enhanced capsule thickness.
However, when we added external LPS into HAC2 such that the resulting endotoxin levels were $80% and 800% of those in HAC1, there were no significant differences in anti-OVA IgG1 antibody titers and FBR when compared with OVA-HAC2 without added LPS.These results suggest that increased LPS content alone in OVA-HAC1 cannot explain the differences in anti-OVA IgG1 antibody titers and FBR.
[41] Inspired by previously reported biomaterial-based vaccine formulations, we tested the effect of encapsulating the adjuvants CpG and GM-CSF. 1,2,9,42CpG is a clinically approved TLR9 agonist and has also been extensively studied for use in other vaccine formulations. 43,44GM-CSF is a clinically approved agent for bone marrow recovery and has also been studied as a potential adjuvant in vaccine formulations. 45In contrast to other common TLR agonists such as Monophosphoryl-Lipid A (MPLA) or squalene, CpG and GM-CSF are water soluble and amenable for encapsulation in HA cryogels.Avoiding the need for organic solvents, solubilizers and stabilizers reduces interference with the crosslinking process.If needed, encapsulating multiple TLR agonists, for example CpG and poly(I:C), may represent a strategy for improving immunogenicity of the HAC2-based vaccine formulation. 46,47 our studies, inclusion of either GM-CSF or CpG into OVA-HAC2 enhanced anti-OVA IgG1 antibody titers comparably and inclusion of both did not further enhance anti-OVA IgG1 antibody titers.
When we assessed innate immune cell infiltration in the various vaccine formulations, inclusion of GM-CSF lead to enhanced neutrophil infiltration, consistent with other reports.Supporting these results, analysis of H&E stained HAC2 vaccine formulations showed an increase in cellularity with formulations including GM-CSF.As the inclusion of GM-CSF in the vaccine increased inflammation at the injection site and did not further improve the immune response, CpG-OVA-HAC2 was chosen for further evaluation.
Our previous finding of the importance of degradation in mediating release of encapsulated components 12 was further validated in the present work.In vitro release in PBS and HYAL2 solutions showed an initial burst release followed by a period of sustained release.Burst release in PBS is expected and is observed in nearly all cryogel-based drug delivery systems.The burst release in HYAL2 solution can be attributed to rapid initial degradation of the HA cryogels, during which $50% of gel degraded after first 9-h for HAC1:Cy5 and HAC2:Cy5, respectively.In vivo, the slower rate of degradation of HA cryogels corresponded to more sustained release of encapsulated vaccine components.
CpG-OVA-HAC2 was administered in immunodeficient mice to assess the mechanism and potential enhancement of vaccine efficacy.
Consistent with our previous work, the degradation of HA cryogels in gp91 phoxÀ mice was impaired. 12Moreover, the induction of anti-OVA IgG1 antibody titers in the gp91 phoxÀ were significantly delayed, supporting the need for release of vaccine components.Post-HSCT Mice which received CpG-OVA-HAC2 injections developed robust anti-OVA IgG1 antibody titers starting 2-weeks post-prime whereas mice which received bolus vaccination only developed failed to develop a detectable adaptive immune response until 6-weeks post-prime.
These results support the utility of CpG-OVA-HAC2 in generating a robust adaptive immune response in immunodeficient mice.
Although two dose prime and boost was more effective at inducing a greater cellular and humoral immune response than a single dose prime and boost, the enhancement did not significantly change the efficacy of prophylactic prevention and therapeutic treatment of tumors between the two groups, as measured by overall survival.
9][50] To potentially improve cancer vaccine efficacy, particularly in established tumors, combining standard-of-care therapy, for example, chemotherapy or checkpoint inhibitors, with cancer vaccines has proven to be an effective strategy to prolong survival compared with either therapy alone. 2,51,52The distinct mechanisms by which these agents operate might leverage both pre-existing tumor-specific immune cells and generate new anti-tumor immunity for a potentially stronger and more durable response.
The results from this study show that by sustaining release of vaccine components, CpG-OVA-HAC2 induces a more potent adaptive immune response compared with conventional bolus vaccination.
These results are consistent with those from other biodegradable injectable scaffold systems.This includes an in situ crosslinked polymernanoparticle hydrogel-based vaccine platform which shows sustained OVA release and enhanced anti-OVA adaptive immune response relative to bolus control. 1 A potential advantage of HA cryogel-based vaccines is that the gels are pre-formed and injected individually, allowing for more control of dosing and less inter-operator variability.Our results are also consistent with another approach using nanocellulose hydrogels, which shows sustained release of OVA provides stronger protection against murine lymphoma, than bolus vaccination. 53However, nanocellulose hydrogels are surgically implanted, whereas HA cryogelbased vaccination is administered with a syringe.Similar to other biomaterial-based vaccines, the HA cryogel-based vaccine allows for incorporating a range of adjuvants and antigens with the advantage of leveraging an established bioorthogonal crosslinking chemistry that reduces the potential for inactivating vaccine components. 54,55e results support the development of a HAC2-based vaccine for protection against infections or cancers, including in settings of immunodeficiency.However, owing to its xenogeneic origin, OVA can be inherently more immunostimulatory than antigens that might be associated with cancer cells.HAC2-based vaccine formulations could be designed to utilize tumor cell lysate as a source of tumor antigen, precluding the need to identify a specific tumor antigen and potentially allowing for testing the formulation in multiple tumors types. 56,57Alternatively, HAC2-based subunit vaccines could be developed that contain known tumor antigens such as Trp2 and gp100. 58,59HAC2-based cancer neoantigen vaccines would likely require optimization with multiple TLR agonists and multiple antigen targets to improve vaccine efficacy and breadth of the adaptive immune response.Further, efficacy of HAC2-based neoantigen vaccines could be assessed in combination with checkpoint inhibitor or chemotherapy. 52,60,61  (Cy5-HA-Nb) was synthesized following a previously described technique. 12A total of 0.8 mg of Cy5-Tz was reacted with 100 mg of HA-Nb at 0.2 wt/vol in DI water for 24 h at 37 C and purified by dialysis in DI water using a 12,000 Da MW cutoff dialysis sack for 48 h.
Dialysis water bath was changed every $8 h.The Cy5-HA-Nb solution was then frozen overnight and lyophilized for 48 h.

| Cryogel formation
We followed a previously described cryogelation method. 12To form cryogels, aqueous solutions of 0.6% wt/vol HA-Tz and Cy5-HA-Nb were prepared by dissolving lyophilized polymers into deionized water and left on a rocker at room temperature for a minimum of      Female mice are commonly used in mouse model of melanoma.Previous studies have shown sex-based differences in adaptive immune response to both vaccination and tumor growth rate. 63,64Females typically develop higher antibody titers than males and tumor growth rate in murine melanoma is slowed in female mice due to a more robust adaptive immune response.It is unknown if there are sexdifferences in the mouse HSCT model described in this work.Male mice were used in the gp91 phoxÀ studies as the mutation is X-linked and males but not females lack superoxide production. 25,26

F
I G U R E 3 HA cryogel degradation is independent of encapsulated adjuvants.(a) Overview schematic depicting for in vivo degradation study.(b-e) Representative IVIS fluorescence images of cryogel degradation and quantification by measuring total radiant efficiency normalized to initial day 3 timepoint of (b) OVA-HAC2, (c) GMCSF-OVA-HAC2, (d) CpG-OVA-HAC2, and (e) GM-CSF and CpG encapsulated OVA-HAC2 (GMCSF-CpG-OVA-HAC2). Data in (e-h) represents mean ± SEM of n = 5 mice.Data in e-h compared two-way ANOVA with Bonferroni multiple comparison test on prime vaccine degradation curves.Figure 3a created using Biorender.
To demonstrate the utility of CpG-OVA-HAC2 in mediating enhanced immunity, we measured the effect of treating immunodeficient mice post autologous hematopoietic stem cell transplant (HSCT) in B6 mice.Mice received lethal radiation followed by HSCT consisting of both 15 M whole bone marrow cells and 10 M splenocytes F I G U R E 6 CpG-OVA-HAC2 provides protection against B16-OVA melanoma.(a) Overview schematic for assessing prophylactic immunization in mediating protection against B16-OVA melanoma.(b) Progression-free survival, (c) tumor volume measured in individual mice, and (d) overall survival.Mice were inoculated with 100K B16-OVA melanoma cells administered subcutaneously either in unvaccinated mice, or after two-dose bolus, single dose CpG-OVA-HAC2, and two-dose CpG-OVA-HAC2 administered as a prime and boost.(e) Quantification of anti-OVA IgG1 antibody titers in serum of vaccinated mice 6-weeks post prime and 3-weeks post tumor inoculation.(f) Overview schematic for assessing therapeutic immunization in mediating protection against B16-OVA melanoma.(g) Tumor volume measured in individual mice and (h) overall survival.Data in (b,c,d) represents n = 10 mice.Data in (e) represents mean ± SD of n = 10 mice.Data in (g,h) represents n = 8-10 mice.Data in (b,d,h) were compared pairwise using log-rank test.Data in (c,g) was compared using Kruskal-Wallis test with Dunnett multiple comparison of area under the curve.Data in (e) was compared using one-way ANOVA with Dunnett multiple comparison.Figure 6a,f created using Biorender.

8 h
to allow for dissolution.The aqueous solutions were then precooled to 4 C before crosslinking to slow reaction kinetics.HA-Tz and HA-Nb solutions were mixed at a 1:1 volume ratio, pipetted into 30 μL Teflon molds which were pre-cooled to À20 C, and quickly transferred to a À20 C freezer to allow for overnight cryogelation.To form OVA-HA cryogels, HA-Tz and Cy5-HA-Nb were solubilized in 3.33 mg/mL OVA solution in DI water prior to crosslinking.To form HA cryogel vaccines, 1 μL/cryogel of CpG (100 mg/mL concentration) and/or GM-CSF (1 mg/mL concentration) was added to HA-Tz or Cy5-HA-Nb solubilized in OVA solution prior to crosslinking.

4. 5 | 4 . 6 | 62 4. 7 |
Endotoxin testingEndotoxin testing was conducted using a commercially available endotoxin testing kit (88282, Thermo Fisher Scientific, lot: VH310729) and following manufacturer's instructions.HA-Tz and Cy5-HA-Nb from both suppliers were solubilized at 0.3 wt% in endotoxin free water and samples were tested in technical triplicates.To calculate endotoxin content of a single HA cryogel, the EU/mL concentration for HA-Tz and Cy5-HA-Nb were multiplied by 0.03 (30 μL of volume per HA cryogel).Other components of the HA cryogel vaccine formulations tested for endotoxin included CpG (100 μg/mL concentration), OVA (100 μg/mL concentration), GM-CSF (1 μg/mL concentration), and RO water which was used to solubilize all vaccine components.Pore size analysis of HA cryogels For scanning electron microscopy (SEM), frozen HA cryogels were lyophilized for 24 h in their molds.Lyophilized HA cryogels were adhered onto sample stubs using carbon tape and coated with iridium in a sputter coater.Samples were imaged using secondary electron detection on a FEI Quanta 250 field emission SEM in the Nano3 user facility at UC San Diego.Fluorescence images of Cy5-labeled HA cryogels were acquired using a Leica SP8 All experiments were performed at the UC San Diego School of Medicine Microscopy Core.Pore size quantification of SEM images was quantified using FIJI image processing package.LPS doping of HAC2LPS was added to HA-Tz and HA-Nb solutions from supplier 2 at either 12.5 ng/100 mg or 1.25 ng/100 mg.The mixtures were stirred for 24 h and then frozen overnight and lyophilized for 48 h.Lyophilized polymers were rehydrated to assess endotoxin concentration prior to conducting in vivo studies.

4. 8 |
HA cryogel pore-interconnectedness analysisCy5-labeled HA cryogels were incubated in 1 mL of FITC-labeled 10 μM diameter melamine resin micro particles (Sigma Aldrich) at 0.29 mg/mL concentration on a rocker at room temperature overnight.Fluorescence images of HA cryogels with FITC-labeled microparticles were acquired using a Leica SP8 confocal.Interconnectedness of the HA cryogels was determined by generating 3D renderings of confocal z-stacks using FIJI imaging processing package and assessing fluorescence intensity of both the Cy5 and FITC channels with depth starting from the top of the HA cryogel.All experiments were performed at the UC San Diego School of Medicine Microscopy Core.

4. 9 | 4 . 10 |
In vitro degradation of HA cryogelsCy5-labeled HA cryogels were placed into individual 1.5 mL microcentrifuge tubes (3448, Thermo Scientific, lot: 20430852) with 1 mL of 100 U/mL Hyaluronidase from sheep testes Type II (HYAL2, H2126, Sigma Aldrich, lot: SLBZ9984) in 1Â phosphate buffered saline (PBS).Degradation studies were conducted in tissue culture incubators at 37 C. Supernatant from samples were collected by centrifuging the samples at room temperature at 2000g for 5 min and removing 0.9 mL of supernatant.HA cryogels were resuspended by adding 0.9 mL of freshly made 100 U/mL HYAL2 in 1Â PBS.Fluorescence measurements were conducted using a Nanodrop 2000 Spectrophotometer (Thermo Fisher Scientific) and these values were normalized to sum of the fluorescence values over the course of the experiment.In vitro release assays OVA-HAC1, OVA-HAC2, CpG-OVA-HAC2, or GMCSF-OVA-HAC2 were placed into individual 1.5 mL microcentrifuge tubes in either 1 mL of 1Â PBS or 100 U/mL HYAL2 solubilized in 1Â PBS.Quantification of in vitro OVA release was conducted using a Micro BCA kit (Thermo Fisher Scientific, lot: UD277184) on supernatant samples.Absorbance values of 100 U/mL HYAL2 solution was assessed and subtracted from absorbance values in supernatant containing HYAL2 solution.Quantification of in vitro CpG release was conducted using OliGreen Assay kit (Thermo Fisher Scientific, lot: 2471822).Quantification of in vitro GM-CSF release was conducted using Murine GM-CSF ELISA kit (Peprotech, lot: 1010055).

4. 12 |
In vivo mouse experimentsAll animal work was conducted at the Moores Cancer Center vivarium at UC San Diego and approved by the Institutional Animal Care and Use Committee (IACUC) under protocol number S17160.All animal experiments followed the National Institutes of Health guidelines and relevant AALAC-approved procedures.Female C57BL/6J mice (B6, Jax no.000664) and BALB/cJ mice (Jax no.000651) were 6-8 weeks at the start of the experiments.Male B6 and B6.129S-Cybbtm1Din (gp91 phoxÀ , Jax no.002365) mice were 6-8 weeks old at the start of experiments.All mice in each experiment were age and sex matched and no randomization was performed.This work did not assess if there are sex-differences associated with the observed effects.
Endotoxin content of HA suppliers.