Antioxidant Cerium Oxide Nanoparticle Coatings Impart Immunomodulatory Effects by Suppressing Antigen-Specific Cytotoxic T Cell Activation

Abstract

Cellular entrapment within biostable hydrogels can decrease immunological rejection by blocking direct contact between the host and transplanted cells; however, these implants are still susceptible to deleterious inflammatory and immunological responses that can dampen their therapeutic effect. Reactive oxygen species (ROS) are key agents that facilitate these responses. While ROS is commonly attributed to general inflammation and cytotoxicity, it also plays an important role in the activation of adaptive immune cells, as ROS-mediated pathways facilitate the efficient generation of effector T cells. Herein, we sought to explore if incorporating a potent antioxidant, specifically cerium oxide nanoparticles (CONP), onto the surface of a hydrogel-based microbead platform could deliver an immunomodulatory biomaterial capable of dampening antigen-specific effector T cell generation. To test this hypothesis, CONP-based coatings were applied to the surface of cell-containing alginate microbeads and co-cultured with immune cells. Quantification of the immune responses found that CONP-coatings decreased the generation of antigen-specific effector CD8⁺ T cells. Interrogation of T cell and antigen-presenting cell (APC) responses found suppression was likely driven by the modulation of CD8⁺ T cells, as APCs were only modestly impacted. Results provide insight into the capacity of CONP to deliver an immunomodulatory effect. These findings also highlight the general potential of antioxidant biomaterials to serve a dual role in protecting cells from ROS-mediated damage and suppressing adaptive immune cell responses.

1. INTRODUCTION

Cellular therapy is a promising approach for treating numerous diseases and conditions caused by pathological death or dysfunction of specialized cells. Cellular transplantation outcomes, however, are commonly hindered by poor cellular engraftment, inflammatory responses, and immunological rejection. One key factor that contributes to many of these deleterious responses is reactive oxygen species (ROS). Excess ROS, e.g., superoxide, hydrogen peroxide, and hydroxyl radicals, are generated immediately upon implantation due to tissue damage^1,2. This localized ROS further propagates inflammation by activating innate immune cells, which produce elevated pro-inflammatory cytokines that can impart significant cell death^3,4. In addition, ROS is a key signaling factor in adaptive immune responses that potently target and destroy the transplanted cells^5–7. As such, reducing ROS at the graft site should not only decrease inflammatory responses but may also modulate implant-targeted immunological rejection.

To mitigate ROS at the transplant site, our group has explored the utility of integrating cerium oxide nanoparticles (CONP) with hydrogels as a ROS scavenging agent. CONPs are potent ROS scavengers with unique catalytic properties⁸. Their antioxidant capacity, resulting from their ability to indefinitely switch between Ce(IV) and Ce(III) states, is theoretically inexhaustible and ubiquitous^8,9. At the same time, their nanoscale and high surface activity deliver efficient reactivity per unit volume^8,9. In our approach, CONP coatings were generated in a layer-by-layer manner onto the surface of alginate microbeads. Specifically, CONP can be deposited via cycling the incubation of CONP solutions with alginate to form CONP/Alginate bi-layers (Figure 1). In this manner, the antioxidant is localized to the surface of the hydrogel, while the underlying alginate microbead serves as a vehicle for entrapping the foreign cells. In our previous work, we established this fabrication approach and demonstrated the capacity of the CONP coatings to efficiently scavenge ambient ROS in a manner that protected the underlying entrapped cells from oxidative damage^10,11.

Figure 1. Generation of layer-by-layer coatings of cerium oxide nanoparticle (CONP) and alginate bi-layer coatings onto alginate microbeads containing cells.

A) Schematic of the layer-by-layer process, achieved via serial incubations in CONP and a diluted alginate solution to form bi-layers. This process is repeated for 3 or 6 cycles, with washes between steps, to achieve up to 6 or 12 CONP-alginate bi-layers. B) Light microscopy images of alginate microbeads either uncoated or coated with 6 or 12 CONP-alginate bi-layers.

In addition to protecting the entrapped cells from direct ROS-mediated cytotoxicity, CONP coatings could potentially help dampen adaptive immunological responses, as ROS is a potent signal in activating antigen-specific immune cells¹². Specifically, T cells require three signals for efficient activation: 1) engagement of their T cell receptor (TCR) via antigen presentation; 2) binding of co-stimulatory molecules between the T cell and the antigen-presenting cell; and 3) signaling via soluble “danger” or pro-inflammatory agents^13,14. In the absence of the third signal, antigen-specific T cell responses are inhibited, with decreased overall proliferation and duration of response, as well as increased presence of anergic T cells and depressed effector function^15,16. ROS production is linked to this third signal, as ROS amplifies the generation of proinflammatory cytokines¹⁷. In addition, intracellular ROS plays a key role in T cell activation, as blocking these pathways reduces effector cell function^12,18,19.

To evaluate if CONP coatings can modulate adaptive immune pathways in response to entrapped cells, we tested these coatings using our established in vitro antigen-specific screening platform²⁰. This system reproducibly detects the generation of antigen-specific effector cytotoxic OT-1 CD8⁺ T cells in response to alginate-entrapped mOVA cells²⁰. mOVA cells express ovalbumin on their cell membrane and present OVA peptides as self-antigen, while the OT-1 CD8⁺ T cell responders have high clonal specificity for the OVA peptide²¹. For this study, alginate microbeads containing the mOVA cells were coated with CONP-Alg layers before co-culturing with OT-1 CD8⁺ cytotoxic T cells. The capacity of the CONP coatings to impair effector CD8⁺ T cell generation was then evaluated. In addition, the mechanism behind this modulation was investigated via delineation of CONP’s specific impact on CD8⁺ T cells and dendritic cells (DC). DC were studied, as they were found to be key antigen-presenting cells (APC) in CD8⁺ T cell activation via cross-presentation^20,22. Results from this work can inform the biomaterial community on the potential of CONP-based biomaterials to protect cells from ROS-mediated cytotoxic effects and dampen immune-specific effector responses.

2. MATERIALS AND METHODS

2.1. Animals

All animal procedures were conducted under protocols approved by the Institutional Animal Care & Use Committee (IACUC) at the University of Florida (Protocol #202008595). Ovalbumin was employed as the model antigen to provide specificity in this study. mOVA mice (C57BL/6J(CAG-OVA)916Jen/J), where all beta-actin expressing cells express both membrane-bound OVA and OVA peptide derivatives as self-antigens²¹, were used as the donors for antigen-bearing stimulator cells. OT-1 mice were used as the donor-mimicking responder cells at 8–15 weeks of age. OT-1 mice have CD8⁺ T cells clonally specific to SIINFEKL peptide via transgenic T cell receptors. The stimulator and responder strains were bred in the UF rodent models breeding core and genotyped to confirm targeted gene expression before use in this study.

2.2. CONP-Alginate (CONP-Alg) Layer-by-layer (LbL) Coating on Alginate Microbeads.

Radical and ambient ROS-scavenging CONP-Alg LbL coatings were assembled onto alginate microbeads following a previously published protocol¹⁰. Briefly, alginate microbeads (cell-free or with mOVA cells, d ≈ 750 μm) were generated using sterile 1.6% (w/v%) UP MVG (cGMP grade, Pronova, NovaMatrix) using a parallel air flow bead generator (Biorep, Inc). Following microbead fabrication, beads were subsequently coated with alternating CONP aqueous dispersion (3 mg/mL in MOPS buffer, cerium (IV) oxide, 20% in H2O, pH = 4, d < 5.0nm, Alfa Aesar, MA) and alginate solution (3 mg/mL in MOPS buffer, cGMP grade, medium viscosity, high guluronic (MVG; Pronova, NovaMatrix) incubations to form a bi-layer coating of CONP and Alginate. The coating time for each layer is 30 seconds with agitation. Three washes with MOPS buffer were applied in between each coating. This process was repeated until the desired number of CONP-Alg bi-layered cycles (Figure 1) was assembled. Alginate microbeads with three bi-layers (n=3, 6L) and six bi-layers (n=6, 12L) were tested in this study, with uncoated alginate microbeads (n=0, 0L) serving as control. Based on our previously published characterization studies, the 12L CONP-Alg coatings were ~ 50 nm thick¹⁰.

2.3. Immune Cell Isolation and Culture with CONP-Alg LbL Coated Microbeads.

Splenocytes sourced from mOVA mice were used as the antigen source for antigen-specific stimulation. Splenocyte isolation and cell entrapment within alginate microbeads were conducted following our previous study²⁰. Briefly, mOVA spleens were collected in cold HBSS buffer and mechanically disrupted and filtered (40-μm cell strainer) to create a single-cell suspension. Erythrocytes were removed by 5 min treatment with ACK lysing buffer and the resulting splenocytes were harvested by spinning down at 500×g for 5 min. For assessment of the global generation of effector CD8⁺ T cells, OTI splenocytes (responders) were seeded 10×10⁵ cells per well in a 96-well round bottom plate (tissue-culture treated, Corning, NY) and stimulated with mOVA cells entrapped within alginate microbeads (total 10×10⁵ cells within 10 beads) coated with various layers of CONP-Alg LbL coatings (0L, 6L and 12L) for 48 hours. An equal amount of free mOVA cells (10×10⁵ per reaction), SIINFEKL peptide (10⁻⁷ M), cell-free alginate microbeads with CONP-Alg LbL coatings (0L, 6L, and 12L), and T cell media-only were used as control stimulators. To interrogate specific CD8⁺ T cell effects, CD8⁺ T cells were purified using EasySep^™ Mouse CD8+ T Cell Isolation Kit (StemCell Inc). These freshly isolated OTI CD8⁺ T cells (1×10⁵ cells per reaction) were treated with SIINFEKL peptide and cultured with 10–30 alginate microbeads (d ≈ 750 μm) with 0 or 12 layers of CONP-alginate coatings for 24–48 hrs. To evaluate APC impacts, splenic dendritic cells (DCs) were purified (EasySep^™ Mouse Pan DC Isolation Kit, StemCell Inc.) from OTI donors and co-cultured with alginate-entrapped mOVA microbeads (total 10×10⁵ cells for 10 beads) with various layers of CONP-Alg LbL coatings (0L or 12L) for 48 hours. Responder cells were collected for immunophenotyping and/or activation analysis via flow cytometry.

2.4. Flow Cytometry

CD8⁺ T cell proliferation and activation and DCs maturation and costimulatory markers in response to the stimulated controls were measured via flow cytometry. Viability of responder cells was labeled with Live/Dead^® Fixable Near IR dye (Thermo Fisher). The proliferation of responder cells was labeled with CellTrace^™ violet dye (Thermo Fisher). Nonspecific Fc binding was blocked with anti-mouse CD16/32 (BD Biosciences, clone 2.4G2, 1:50). OVA-specific CD8⁺ T cell activation was characterized by labeling anti-mCD8a-PE, and anti-human/mouse granzyme B-APC (Table S1). OTI CD8⁺ T cell activation was defined as the percentage of proliferating granzyme B⁺ CD8⁺ effector cells (see gating strategy in Figure S1). The mean fluorescence index (MFI) was calculated for selected studies using the geometric mean of granzyme B expression after gating for proliferating CD8⁺ T cells (Figure S2). DCs were assessed by staining anti-mCD11c-PE, anti-CD8a-BV650, anti-mCD80-Pacific blue, anti-mCD86-PE/Cy5, anti-I-Ab/I-Eb- Pacific blue, and anti-H-2Kb-APC (Table S1). DC maturation was quantified as fold-change in maturation marker stimulation compared to the unstimulated control post a 48-hour culture (see gating strategy Figure S6). Samples were analyzed using a BD Celesta flow cytometer (BD Biosciences). Background signals were identified and excluded by fluorescence-minus-one controls. Data analysis was performed using FCS Express 6.05 software (DeNovo software).

2.5. Measurement of Intracellular ROS Accumulation

Intracellular ROS production of the OTI responder cells post-stimulation was determined via CellRox Deep Red oxidative stress reagent (5 μM; Life Technologies) following the manufacturer’s protocol. CellRox Deep Red dye was added at a final concentration of 5 μM to the coculture systems at the end of the culture period and then incubated for 30 min at 37°C. Subsequently, the CellRox Deep Red dye-containing medium was removed, and the cells were washed with PBS and labeled with SYTOX blue as the viability indicator. The resulting intracellular fluorescent signal was analyzed via flow cytometry.

2.6. Oxidative Stress-Related Gene Expression Profiling

Following the designated co-culture, total RNA of OTI CD8⁺ T cells was extracted and reverse transcribed using RNase Easy and RT2 First Strand Kits (Qiagen). qRT-PCR array (RT2 profiler PCR arrays Oxidative Stress, PAMM-065Z, Qiagen) was performed using a QuantiStudio 6 flex system (Life Technologies). See Table S2 for summary of genes. Data analysis was conducted by Qiagen PCR Array Data Analysis web-based tool using the 2−ΔΔCt method normalized to unstimulated CD8⁺ T cells and housekeeping genes. Differential gene expression of cells, compared to their controls, was considered significant if both the log2 Fold Change ≥ 1 or ≤ − 1 and p-value ≤ 0.05. For this study, the control group was either unstimulated CD8⁺ T cells or CD8⁺ T cells incubated with SIINFEKL and uncoated beads, as designated. Results were summarized using volcano plots.

2.8. Statistical Analysis

The power of tests and the statistical methods are described throughout the article and in the figure legends. Results shown are representative of N = 2–3 independent studies with n = 3–4 independent replicates per study, as designated in the legend. Data are summarized as the average ± standard deviation. Generally, statistical assessments were performed using one-way or two-way ANOVA with Tukey’s multiple comparison analysis using GraphPad Prism 8.4.3 software. Statistical difference was considered significant when the probability value (p) was <0.05. Difference was shown as *p<0.05; **p<0.01; ***p<0.001; **** p < 0.0001 and n.s. indicates not significant.

3. RESULTS

3.1. CONP layer-by-layer coatings onto alginate microbeads suppressed the generation of cytotoxic T cells in response to the entrapped antigenic cells

As outlined in Figure 1, CONP-based layer-by-layer coatings were fabricated onto alginate microbeads via cycling alternating incubations of CONP and alginate solutions using methods optimized in our previous work^10,11. Prototypes of beads coated with either 6 or 12 total layers of CONP-alginate (Figure 1B) were generated. As previously reported, the resulting coatings were ≤ 55 nm in thickness, and the amount of cerium deposited within the layers was 32,879 ± 6,348 versus 64,879 ± 3,168 ng/50 beads for the 6- and 12-layer coatings, respectively^10,11 Previous work also found that these coatings did not impair the capacity of the entrapped cells to receive nutrients or dynamically respond to external stimuli¹⁰.

Our established in vitro antigen-specific system was used to investigate the capacity of CONP-based coatings to modulate adaptive immune responses. Specifically, mOVA splenocytes were entrapped within alginate microbeads before applying a CONP-based coating via layer-by-layer assembly. CONP-coated, cell-containing alginate microbeads were then cultured with OT-1 splenocytes containing CD8⁺ T cells with a high clonal specificity to OVA antigen (Figure 2A). The OT-1 splenocytes also had other immune cells, including antigen-presenting cells and CD4⁺ T cells. The subsequent generation of viable, proliferating, and granzyme B⁺ CD8⁺ T cells in response to shed antigen was then quantified via flow cytometry, as these markers classically define an effector cytotoxic T cell (Figures 2B and S1)^23,24. As expected, the co-culture of OT-1 splenocytes with mOVA cells not entrapped within alginate resulted in the robust generation of antigen-specific effector CD8⁺ T cells (Figure 2C, black bar). The entrapment of the mOVA cells within alginate microbeads did not prevent cytotoxic CD8⁺ T cell generation (Figure 2C, white bar), as the indirect antigen recognition pathway remained intact due to diffusion of antigens out of the alginate, as we have previously reported²⁰. The addition of a CONP coating, comprised of either 6 or 12-layers, onto the cell-containing alginate microbead resulted in the significant suppression of effector CD8⁺ T cell generation, with ~50% reduction for both 6 and 12-layer CONP coatings when compared to CONP-free beads (fold change of 0.54 ± 0.039 and 0.49 ± 0.11 respectively; Figure 2B middle and right panels, respectively, and Figure 2C). No significant differences in effector CD8⁺ T cell generation were noted between the 6 or 12 L CONP-coated groups (p = 0.866, Tukey post-hoc). These results illustrate a suppressive impact of the CONP-based coatings on adaptive immune responses.

Figure 2. CONP coating onto alginate microbeads decreases the generation of cytotoxic effector CD8⁺ T cells in response to alginate-entrapped antigen-specific cells.

A) Schematic of the experiment, whereby OT-1 splenocytes were co-cultured with free or alginate-entrapped mOVA cells. Alginate microbeads containing mOVA cells were either left uncoated or coated with 6 to 12 layers of antioxidant CONP-Alginate. Antigen-specific OT-1 CD8⁺ T cell activation was quantified by the % of proliferating (cell trace) and granzyme B⁺ CD8⁺ T cells via flow cytometry analysis. B) Representative flow plots of OT-1 CD8⁺ T cell activation in response to mOVA cells entrapped within alginate microbeads that were either uncoated (“NO CONP”, left panel) or coated with 6 layers (“6L CONP”, middle panel) or 12 layers of CONP-Alg (“12L CONP”, right panel). X axis is cellular proliferation via cell trace, while y-axis is granzyme B staining. Cells shown are already gated for single cells, viability, and CD8⁺. C) Summary of fold change in proliferating granzyme B⁺ CD8⁺ T effector cells, as compared to uncoated alginate microbeads containing mOVA cells (i.e., “NO CONP”). The “Alginate Beads” group is an alginate-only (no cells) negative control, while the “mOVA Cells” group is the cell only (no alginate) positive control. Experimental groups include alginate microbeads containing mOVA cells coated with either 6 (“6L CONP”) or 12 layers (“12L CONP”) of CONP-Alginate. Bars indicate the average of individual data points from one representative experiment with standard deviation (n=3; N=2). Statistical significance was determined by one-way ANOVA with Tukey’s posthoc. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001

3.2. CONP layer-by-layer coatings suppress cytotoxic T cell effector phenotype

To investigate how CONP-based coatings may impart their immunosuppressive effects, the culture conditions were altered to examine only CD8⁺ T cells. For this study, purified OT-1 CD8⁺ T cells were isolated and cultured with their cognate antigen (i.e., SIINFEKL), resulting in the proliferation and activation of these cells. In addition, either uncoated or 12-layer CONP-coated alginate microbeads were included in the culture (Figure 3A). A dose-response was evaluated by adding either the base (10 microbeads) or three-fold (30 microbeads) dose. The response of CD8⁺ T cells to the defined stimuli was analyzed after 24- and 48-hours.

Figure 3. CONP coating alters the phenotype of CD8⁺ T cells responding to antigen.

A) Schematic of the experiment, whereby purified CD8⁺ T cells were treated with cognate antigen (SIINFEKL, 10 nM) and either alginate-only microbeads (orange bars) or alginate microbeads coated with 12L of CONP-Alg at 10 (blue bar) or 30 (green bar) bead doses for up to 48 hours. B) Fold change in the percentage of viable and proliferating CD8⁺ T cells for each group after 24 and 48 hours of culture, compared to uncoated bead controls. C) Granzyme B expression, quantified via fold change in MFI from uncoated controls, after 48 hour culture. D) Example histograms of intracellular ROS staining of groups, including FMO control. E) Fold change in ROS MFI for CD8⁺ T cells responding to designated treatment after 24 and 48 hour culture. Results were normalized to levels measured for 24 hr unstimulated controls. F-G) Volcano plot of RT-PCR results for CD8⁺ T cells cultured with antigen and either 10 (F) or 30 (G) microbeads for 48 hours. The dashed line on the Y-axis designates p = 0.05. Gene expression levels are reported as fold change from controls of SIINFEKL + uncoated beads. Results in B, C, and E are shown as bars, which indicate the average of individual data points (n=3; N=3) with standard deviation. Statistical significance was determined by two-way ANOVA with Tukey’s posthoc. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 ns=not significant

Without the addition of the cognate peptide, CD8⁺ T cells were minimally activated due to the lack of antigen stimulation, regardless of their culture conditions or materials present, Figure S3A. This demonstrated that the alginate microbeads, uncoated or CONP-coated, did not activate the CD8⁺ T cells. As expected, the inclusion of the cognate antigen into the media resulted in the robust proliferation and conversion of OT-1 CD8⁺ T cells towards an effector phenotype, when compared to non-peptide treated controls (Figure S3B; p < 0.0001, 1-way ANOVA).

The addition of CONP-coated alginate microbeads at the 10 or 30 bead dosage did not significantly alter the number of viable and proliferating CD8⁺ T cells compared to the uncoated alginate microbead control group (Figure 3B, p = 0.49, 1-way ANOVA). However, the presence of the CONP coating significantly reduced (by ~50%) the generation of effector cytotoxic T cells, as measured via granzyme B expression, when compared to uncoated alginate bead controls (Figure 3C; p < 0.0001, 1-way ANOVA). CONP bead dosage did not significantly affect this metric (p = 0.93; Tukey post-hoc). This CONP-mediated suppression was not due to cytotoxicity, as there were no significant alterations in cell viability or death when comparing cells cultured with peptide and uncoated or coated beads (Figure S4). Overall, this data indicates that the mechanism of action of the CONP coating’s impact was attributed to suppression the conversion of CD8⁺ T cell to an effector phenotype and not due to cytotoxic effects.

It was suspected that the CONP’s suppressive response would correlate to ROS signaling, given that the activation of ROS pathways promotes the T cell effector phenotype^7,25. Quantification of intracellular ROS found that the activation of CD8⁺ T cells via antigen resulted in a robust ROS increase when compared to untreated controls, with a 4.44 ± 0.22 and 4.06 ± 0.12-fold increase at time points 24 and 48 hours, respectively (Figure 3D & E). The inclusion of the CONP coating onto the microbead significantly dampened intracellular ROS levels within CD8⁺ T cells in a dose-dependent but time-independent manner (p < 0.0001 and p = 0.59 for dose and time factors, respectively; 2-way ANOVA).

To examine the impact of CONP-based coatings on the expression of genes associated with oxidative stress and antioxidant defense, RNA from CD8⁺ T cells cultured with their cognate antigen and either uncoated or 12L CONP-coated microbeads was collected after 48 hours. The impact of CONP dose was examined by increasing the number of 12L CONP-coated microbeads per well from 10 to 30. Over 80 genes were screened, as summarized in Table S2.

Globally, the addition of SIINFEKL to CD8⁺ T cell cultures resulted in the alteration of a total of 17 oxidation-related genes when compared to unstimulated controls (Figure S5); the most prominent downregulated genes (> 3 fold) being Mpo, Ccl5, Hspa1a, while the most prominent upregulated genes (> 3 fold) being Prdx4, Prnp, Scd1, and Recql4. These results implicate alterations in these genes during antigen-stimulated CD8⁺ T cell expansion and activation.

Including the CONP coating onto the microbeads resulted in significant and substantial changes in four genes, when compared to peptide-stimulated but uncoated microbead controls (Figure 3F & G). Specifically, Vim was significantly downregulated, and Scd1, Serpinb1b, and Txnrd1 were significantly upregulated in CD8⁺ T cells incubated with both doses of CONP-coated beads. A comparison between the two doses found no significant changes in the expression of the genes tested.

3.3. CONP layer-by-layer coatings impart minimal impacts on dendritic cell activation

While our purified CD8⁺ T cell experiments indicate CONP directly dampens the generation of effector cytotoxic T cells via modulation of ROS pathways, our previous work also established a key role of antigen-presenting cells, specifically dendritic cells (DC), in cross-presenting antigen to CD8+ T cells to generate effector T cells within our mOVA-OT-I co-culture model system²⁰. Thus, to further investigate how CONP coatings may modulate effector CD8⁺ T cell generation, additional studies sought to isolate the potential impacts of CONP on dendritic cells. Specifically, CD11C+ pan DCs were isolated from OT-1 splenocytes and cultured with alginate microbeads containing mOVA cells (Figure 4A). The cell-containing microbeads were uncoated or coated with 12 layers of CONP-Alginate. DC phenotype was characterized via flow cytometry, with a focus on markers associated with DC presentation to T cells for subsequent activation (i.e., CD80 and CD86)^26,27, as well as maturation (i.e., MHC Class I and II)²⁸.

Figure 4. CONP coating minimally impacts dendritic cell phenotype.

A) Schematic of the experiment, whereby purified pan dendritic cells (DC) were co-cultured with alginate-entrapped mOVA cells without (orange bars) or with 12L of CONP-Alg (blue bar) for 48 hours. Dendritic cell phenotyping explored differences in the expression of CD80 (B), CD86 (C), MHCI (D), MHCII (E), and ROS (F). Results are shown as bars, indicating the average individual data points (n=3; N=2) with standard deviation. Statistical significance was determined by one-way ANOVA with Tukey’s posthoc. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 ns=not significant

As summarized in Figure 4B–C, the culture of DCs with alginate microbeads containing mOVA cells resulted in the upregulation of CD80 and CD86 compared to untreated controls (p < 0.0001 and = 0.003, respectively; Tukey post-hoc). While MHC Class I was not upregulated in these DC cultures, MHC Class II was increased compared to untreated controls (p = 0.075 and 0.002, respectively; Tukey post-hoc), indicating a maturation of effector phenotype associated with exogenous antigen presentation²⁸. A minimal effect of the CONP coating on DC activation markers was noted, with only a modest but significant decline in CD80 expression compared to untreated controls (p = 0.025; Tukey post-hoc). No significant changes in CD86 or MHC Class II expression were noted between uncoated and 12L CONP-coated microbead groups.

An examination of the potential role of intracellular ROS on DC phenotype found disparate results. While a classic LPS exposure increased intracellular ROS within the DCs (p < 0.0001; Tukey post-hoc), the incubation of DCs with alginate microbeads containing mOVA cells did not alter this metric (p = 0.885, Tukey post-hoc). Furthermore, the inclusion of the CONP coating imparted no distinct effect. These results indicate that the upregulation of ROS signaling is not a phenotype associated with DC cross-presentation to CD8⁺ T cells within this experimental system.

4. DISCUSSION

The entrapment of cells within hydrogels has long been investigated as an approach to prevent immunological recognition and rejection. However, its efficacy in the clinic is limited, with evidence of robust innate and adaptive immune responses to the entrapped cells²⁹. Animal models and recent in vitro studies provide further evidence that hydrogel barriers, while effective in damping direct antigen recognition of the entrapped cells, are insufficient in completely suppressing adaptive immune responses^20,30,31. This is attributed to the activation of indirect antigen recognition pathways. While the resulting effector immune cells cannot directly interact with the implanted cells due to the biomaterial barrier, they can still impart deleterious effects via the secretion of soluble cytotoxic agents and/or the promotion of fibrotic encapsulation^20,31,32. Ass such, additional material modifications to suppress these adaptive pathways could provide a benefit³³.

One approach that could suppress immune responses to cellular implants is scavenging reactive oxygen species (ROS). Intracellular and extracellular ROS play important roles in directing adaptive immune cell responses, including APC activation, T cell and APC signaling, and subsequent T cell activation, as well as in regulatory mechanisms³⁴. Imbalances in ROS lead to deleterious adaptive immune conditions, from hyperactivation (e.g., autoimmune responses) to suppression (e.g., tumors)^35,36. While elevated local ROS can disrupt immune cell pathways, typically via induction of cell death, translating this approach to cellular implantation would create an undesired microenvironment, as high ROS levels can permeate the hydrogel and destroy the graft. An alternative approach is suppressing local ROS by including antioxidants in the graft site. This approach impairs adaptive immune cell activation and provides additional protection to the implanted cells by mitigating local oxidative species.

We and others have explored the integration of antioxidant coatings onto biomaterial systems to protect the entrapped cells, in particular pancreatic beta cells for treating type 1 diabetes^{10,11,37–41}. For beta cell transplantation, ROS-scavenging is particularly impactful, as beta cells are highly susceptible to oxidative stresses⁴². In our laboratory, we have integrated cerium oxide nanoparticles (CONP) into biomaterials as a potent antioxidant, as it scavenges a variety of oxidants and does not expire or degrade like other agents (e.g., catalase, ascorbic acid, tannic acid). Previous studies have demonstrated the capacity of CONP, embedded or coated onto biomaterials, to repeatedly scavenge ambient oxidative agents, such as hydrogen peroxide, resulting in the protection of the co-entrapped beta cells^10,11,41.

Delving further into the benefits of the antioxidant CONP, we sought to explore if CONP-coated biomaterials can also alter immunological pathways activated in response to hydrogel-entrapped cells. Recently, we developed a co-culture in vitro platform that provides a valuable screening tool for interrogating the benefits of immunomodulatory approaches. Specifically, this system allows for delineating and examining direct and indirect antigen activation pathways in response to cells entrapped within hydrogels²⁰. Leveraging this platform by adding distinct immune cell populations (e.g., whole splenocytes, purified CD8+ T cells, purified dendritic cells), we gained new insight into the capacity of CONP coatings to modulate different immune responses.

Culturing of the alginate-entrapped target cells with antigen-specific whole splenocytes resulted in the expected expansion of antigen-specific CD8+ T cells and their subsequent conversion to an effector cytotoxic phenotype (as measured via granzyme B expression). The inclusion of a CONP coating, built using alternative layer-by-layer deposition of alginate and CONP, resulted in substantial suppression in CD8+ T cell effector cell generation. As this OT-1 & mOVA system is highly reactive and generates robust cytotoxic CD8⁺ T cells in only two days, the degree of suppression observed was impactful. Also, as whole splenocytes were used in this experiment, results illustrate a benefit imposed by the CONP coating even when helper cells (e.g., APCs and CD4⁺ T cells) were present. Results observed herein correlate with other antioxidant approaches that observed decreased effector T cell generation^43,44.

Further interrogation into the specific impacts of CONP coatings on immune pathways was feasible using our co-culture system. Specifically, the culture of purified CD8⁺ T cells and relevant antigen (i.e., SIINFEKL) results in the efficient generation of effector T cells. Including alginate beads, with or without the outer CONP coating, allows for investigation into the specific effects of CONP on T cell phenotype without the compounding factors of helper T cells or APC. As shown, CONP coatings suppressed T cell conversion towards an effector phenotype, but not due to cytotoxicity. ROS scavenging was the likely mechanism, as intracellular ROS was significantly suppressed in the CD8⁺ T cell population. Activation of TCR on T cells is known to induce rapid intracellular ROS⁴⁵; a phenomenon validated by our results. This elevated ROS subsequently promotes T cell maturation towards an effector phenotype^46,47. Thus, CONP suppression of these pathways is the likely mechanism behind the observed decrease in T cell maturation. Other publications exploring antioxidant effects on CD8⁺ T cell activation, isolated from APC and CD4⁺ T cell support, found similar suppression of granzyme B expression of viable CD8⁺ T cells treated with a soluble and transient antioxidant⁴³. Given CONP’s renewable antioxidant nature, it is postulated that this suppressive effect is more durable, although future studies are needed to validate this.

Investigation of the genes altered by our experimental conditions revealed interesting trends. Generally, the stimulation of OT-1 CD8⁺ T cells via OVA peptide resulted in an expected genetic signature. Upregulation of oxidation-related genes Prdx4, Prnp, and Scd1 upon T cell activation was expected, per published reports^48–50. Correlations between Recq14 upregulation and T cell activation have not been previously reported, although the upregulation of Recq14 generally correlates to increases in cellular oxidative stress⁵¹. Genes downregulated upon antigen-specific stimulation included CCL5, MPO, and Hspa1a. CCL5 (RANTES) is commonly upregulated with T cell activation; however, this occurs 3–5 days after introducing the stimulating antigen⁵². As cells in this study were characterized 48 hours after adding the stimulatory peptide, the timeline for CCL5 upregulation was not aligned. While MPO is more commonly associated with neutrophils and macrophages, it has been detected in T cells; however, its role and/or expression patterns during expansion and/or activation are not reported⁵³. Finally, Hspa1a is primarily upregulated in exhausted or stressed CD8+ T cells⁵⁴, thus the lack of upregulation of Hspa1a in freshly stimulated CD8⁺ T cells was expected.

Examining changes in the gene expression profile of OT-1 CD8⁺ T cells during stimulation in the presence of CONP coatings revealed a unique signature, in lieu of a simple switch of altered genes back to non-activated cells. This indicates that CONP-mediated modulation of CD8⁺ T cell activation is a unique process. The most dominant gene upregulated in CD8⁺ T cells cultured with CONP coatings was Scd1. Scd1 encodes the generation of the enzyme Stearoyl-CoA desaturase-1(Scd-1), which is responsible for maintaining fatty acid homeostasis within the cell⁵⁵. It is also correlated with the modulation of immune phenotype, with upregulation of Scd-1 associated with suppression of immune phenotype across various cells, including CD8⁺ T cells⁵⁰. In fact, the inhibition of Scd-1 in T cells has been found to enhance the conversion of CD8⁺ T cells to effectors, with measurable increases in the proportion of effector cells post-stimulation and increased cell expression of granzyme B, perforin, and INF-gamma^50,56. It is postulated that depletion of Scd-1 elevates oxidative phosphorylation, which leads to increased expression of effector molecules such as granzyme B. By upregulating Scd-1 expression, the increased presence of Scd-1 likely reduces this oxidative state. Changes in Scd-1 within CD8⁺ T cells in response to antioxidants have not been reported, to our knowledge. Thus, further investigation into the commonality of this CONP-induced modification to other antioxidants is of interest. Serpinb1b was also upregulated in response to CONP coatings. As effector CD8⁺ T cells express granzymes, which are serine proteases, the upregulation of Serpinb1b, a serine peptidase inhibitor, may result in the observed impaired CD8⁺ T cell effector function^57,58. Other highly upregulated genes following the addition of CONP coatings include Srxn1 and Txnrd1. Both genes encode endogenous antioxidants that suppress T cell conversion towards an effector phenotype and protect T regulatory cells^59–61. The prominent downregulation of Vim (vimentin) in response to CONP coatings was unexpected, as antioxidant regulation of Vim in CD8⁺ T cells has not been reported. Vimentin is a major intermediate filament within leukocytes and plays an important role in immune cell function, particularly related to migration and tissue infiltration^62,63. While the role of vimentin in T cell activation is not widely studied, limited reports indicate that the loss of protein within CD8⁺ T cells results in dysfunctional TCR clustering and synapse formation during stimulation, which can lead to insufficient effector conversion and efficacy^64,65. Given that Vim modulation using antioxidants is not widely reported, exploring the modulation of this gene is another area of future interest.

Our previous work established that the primary pathway for generating antigen-specific, effector CD8⁺ T cells in response to entrapped cells was via cross-presenting dendritic cells²⁰. Thus, to examine the effect of CONP-coated microbeads on DC phenotype, DCs were purified from splenocytes and co-cultured with mOVA cell-entrapped microbeads that were uncoated or coated with CONP. The effect of CONP on DC phenotype was modest, with limited impact on DC activation or ROS production. Contextualizing our results with ROS literature on DC indicates that modulation of DC cross-presentation via antioxidants may be limited. While DCs exposed to oxidative stress exhibit changes in phenotype with increased secretion of soluble cytokines, these changes were not shown to increase downstream CD8⁺ T cell activation or effector conversion⁶⁶. Published work has also demonstrated that the impact of oxidative stress on DCs varies, depending on their subtype, i.e., conventional (cDC) or plasmacytoid (pDC), with conventional being more resistant to oxidative stress effects^67,68. Since our isolation method and tissue source yielded a mixture of pDC and cDC, further work is needed to investigate these effects. The sensitivity of innate immune cells to oxidative stress has been well established, with macrophages and neutrophils identified as key responders and generators of ROS. Since these APC subsets generate high ROS for local release and cell signaling, antioxidant approaches can dampen their effector phenotype^69–71. ROS generation in DC, however, is unlike that observed for other antigen-presenting cells, as high levels of ROS are not typically observed during antigen processing and can even be detrimental to effective cross-presentation^72,73. Given that macrophages are likely present in our whole splenocyte experiments, further investigation into the impacts of CONP on macrophages could elucidate additional benefits.

These CONP-based coatings uniquely impact the immune cells surrounding the alginate-entrapped cells, leading to the dampened activation of effector T cells. Decreasing the number of activated CD8+ T cells surrounding a cell-containing biomaterial implant would likely impart benefits to the survival of the entrapped cells, as well as decrease generalized foreign body responses. However, future work is needed to validate this hypothesis. It may be postulated that the benefits of the CONP coatings are due to the coatings decreasing the permeability of the overall material, which would lead to reduced antigen shedding. Our previous work did not find alterations in glucose-stimulated-insulin-release for entrapped beta cells¹⁰, indicating that the embedded cells can both receive glucose and release insulin at a similar timescale to those within non-CONP coated alginate microbeads. Additionally, we did not observe dosage effects of the CONP layer on downstream CD8⁺ T cell activation, which would likely be the case if the layers altered the overall permeability of the system. Future work can further investigate these hypotheses.

5. CONCLUSION

In conclusion, this study illustrated the multi-faceted benefits of CONP-based biomaterials in modulating deleterious ROS-mediated responses. Specifically, this work investigated the ability of CONP coatings to inhibit the generation of effector T cells in response to antigenic stimuli. Using an antigen-specific cell co-culture system, potent suppression of effector T cell generation was observed, without measurable cytotoxic effects. Delineation of T cell and DC pathways found that CONP coatings impart their most significant impacts on T cell phenotype, suppressing granzyme B expression and intracellular ROS. CONP coatings also generated unique alterations in CD8⁺ T cell gene expression, which provides unique genes of interest for further study. Based on these results, the addition of CONP coatings to hydrogel platforms could impart multi-pronged benefits, as ROS scavenging can protect the underlying cells and dampen adaptive immune responses.