Induction of Antigen-Specific Tolerance in Multiple Sclerosis Model without Broad Immunosuppression

Abstract

Multiple sclerosis (MS) is a severe autoimmune disorder that wreaks havoc on the central nervous system, leading to a spectrum of motor and cognitive impairments. There is no cure, and current treatment strategies rely on broad immunosuppression, leaving patients vulnerable to infections. To address this problem, our approach aims to induce antigen-specific tolerance, a much-needed shift in MS therapy. We have engineered a tolerogenic therapy consisting of spray dried particles made of degradable biopolymer, acetalated dextran, and loaded with an antigenic peptide and tolerizing drug, rapamycin (Rapa). After initial characterization and optimization, particles were tested in a myelin oligodendrocyte glycoprotein (MOG) induced experimental autoimmune encephalomyelitis model of MS. Representing the earliest possible time of diagnosis, mice were treated at symptom onset in an early therapeutic model, where particles containing MOG and particles containing Rapa+MOG evoked significant reductions in clinical score. Particles were then applied to a highly clinically relevant late therapeutic model during peak disease, where MOG particles and Rapa+MOG particles each elicited a dramatic therapeutic effect, reversing hind limb paralysis and restoring fully functional limbs. To confirm antigen-specificity of our therapy, we immunized mice against influenza antigen hemagglutinin (HA) and treated them with MOG particles or Rapa+MOG particles. The particles did not suppress antibody responses against HA. Our findings underscore the potential of this particle-based therapy to reverse autoimmunity in disease-relevant models without compromising immune competence, setting it apart from existing treatments.

Graphical Abstract

Introduction

Multiple sclerosis (MS) is a life-altering chronic disease characterized by autoimmune destruction of myelin in the central nervous system, resulting in a variety of symptoms including but not limited to vision impairment, muscle weakness, vertigo, memory deficits, and mood changes ^1–4. As of 2020, there are an estimated 2.8 million cases worldwide, with increasing prevalence, and like most autoimmune disorders, incidence is substantially higher in women than men ^{5, 6}. Current FDA-approved disease modifying therapies (DMTs) for MS include broadly immunosuppressive injectable IFN-β and oral immunosuppressants ^{7, 8}. MS patients may also be treated with injectable glatiramer acetate, a unique amino acid copolymer that achieves immunosuppressive effects through a multi-faceted mechanism (e.g. MHC II competition with myelin basic protein, Th2 skewing) ⁹. Recently, antibody therapies have been FDA-approved to ameliorate MS more precisely; however, this approach still suppresses major components of the immune system. The most common antibody therapies used to treat MS target CD20, thereby depleting B cells. Clinical trials have demonstrated decreased vaccine efficacy in MS patients receiving necessary standard of care therapies, especially B cell depleting therapies, indicating reduced protection in this vulnerable group ^10–12. Despite clinical benefit of current therapies, there is no cure for MS, and patients are often susceptible to infectious disease. There is a strong unmet need for a more specific therapy that would halt the autoimmune response without interfering with immune competence necessary to protect against infections.

Although the exact causes of MS are unknown, substantial efforts have been made to increase our understanding of the immunological mechanisms driving disease and immune tolerance. Classically, it is thought that self-reactive CD4 T cells infiltrating into the brain and spinal cord drive the development of MS, and the influx of autoreactive T-helper type 1 (Th1) and Th17 cells during disease progression is well documented ^{13, 14}. However, with recent successes of B cell depleting therapies in the clinic, autoreactive B cells have been shown to play an instrumental role in disease progression ^{15, 16}. Furthermore, B cells have been implicated in the recently discovered link between Epstein–Barr virus infection and MS ¹⁷. To combat autoreactive immune cells, it is imperative to induce immune tolerance to only the disease-associated antigen. This can be done by specifically depleting autoreactive cells or increasing the prevalence of regulatory phenotypes ^18–20. Regulatory T and B cells (T_regs and B_regs) are of particular interest, with T_regs having a longer history of study for the amelioration of autoimmunity ^18–20. T_regs are often classified based on expression of Foxp3 and are induced by tolerized antigen presenting cells (APCs), which present antigen to naïve T cells without secondary co-stimulation. B_reg induction can occur from different stimuli and is often characterized by expression of anti-inflammatory cytokines such as IL-10, IL-35, and/or TGF-β ²¹. T_regs and B_regs inhibit inflammatory responses in an antigen-specific manner by shifting the cytokine environment or by direct inhibition of reactive cells, and B_regs can further induce T_regs ^{18, 21, 22}. Targeting mechanisms of immune tolerance is a promising strategy to halt autoimmune responses without broad immunosuppression.

Therapeutic vaccines inducing antigen-specific tolerance hold substantial promise in enhancing the quality of life of patients living with MS and other autoimmune disorders. Two possible therapeutic vaccine approaches include tolerized cell therapies and strategies to deliver disease-associated peptides to APCs in a tolerogenic manner. Tolerized cell therapies, namely tolerized dendritic cells (tolDCs) and T_regs induced ex vivo, are a promising approach. Companies are exploring antigen-specific tolerogenic cell therapies, like tolDCs, with some already in clinical trials (NCT02903537, NCT04530318, NCT02283671, NCT02618902). However, cell-based therapies are costly and difficult to manufacture and translate into the clinic ^{18, 23}. In situ reprogramming of APCs with particle delivery platforms is more cost-effective and translatable ¹⁸.

Particles passively target phagocytic APCs due to their size and can be formulated to deliver various therapeutic cargoes²⁴. While there is interest in developing tolerogenic particle therapies for MS, none are currently in clinical trials, with company pipelines focusing on other autoimmune disorders. Examples include Selecta, which merged with Cartesian Therapeutics to focus on RNA cell therapies, and Vaccitech, now Barinthus Biotherapeutics, developing a particle therapy for celiac disease.

To deliver disease-associated antigenic peptide to APCs in a tolerogenic manner, additional pharmacological agents are often co-encapsulated with the peptide in the particles. A commonly employed tolerizing agent is rapamycin (Rapa). Also known as Sirolimus or Rapamune, Rapa is FDA-approved as an immunosuppressant or anti-proliferative therapy ²⁵. As a promising agent for tolerance induction, Rapa inhibits mTOR to modulate dendritic cell (DC) and T cell differentiation ^{26, 27}. Rapa can be used to induce tolDCs and has been shown to encourage the induction of T_regs both in vitro and in vivo ^{18, 28, 29}. Rapa-containing therapies have also demonstrated therapeutic benefit in pre-clinical models for autoimmune disease ^30–32. Consequently, Rapa was selected for use in the particles developed in the present work.

In this work, acetalated dextran (Ace-DEX) was utilized to formulate tolerogenic particles for use as a therapeutic vaccine in a murine MS model. Ace-DEX is a biocompatible polymer with substantial utility as a drug carrier. Due to its acid sensitivity, Ace-DEX readily degrades inside the endosome of phagocytic APCs but persists longer in the pH-neutral extracellular space ^{33, 34}. As a result, triggered release of particle cargo can be achieved inside of the targeted APCs. Furthermore, Ace-DEX particles have been previously applied in experimental autoimmune encephalomyelitis (EAE) models, the gold standard pre-clinical models for MS ³⁵, with one report of particles containing Rapa and antigenic peptide applied to a relapsing-remitting model and another report of particles containing dexamethasone and antigenic peptide applied to a progressive model ^{31, 36}. In both works, Ace-DEX particles loaded with both peptide and a tolerizing agent elicited a reduction in disease severity, but underlying immunological mechanisms were not well studied.

Building upon previous work, particle fabrication was adapted to a scalable spray drying method rather than batch emulsion methods. Biodistribution and uptake of spray dried particles by APCs in vivo were assessed, followed by dose optimization of therapeutic cargoes within the Ace-DEX particles. Additional analyses were performed to further our understanding of immunological mechanisms impacted by tolerogenic Ace-DEX particle therapies. Furthermore, the application of therapeutic Ace-DEX particles at multiple stages of disease in EAE is highly clinically relevant, and evaluating an influenza vaccine response concurrent with tolerogenic particle therapy provides a critical assessment of the treatment’s antigen-specificity. The present work aims to improve upon Ace-DEX particles formulated for EAE and to inform further development and future clinical translation of tolerogenic particles for MS and other autoimmune diseases.

Results

Fluorescent Ace-DEX is Spray Dried to Form Particles that Target Antigen-Presenting Cells In Vivo

Dextran was fluorescently labeled (Alexa-DEX) and acetalated to evaluate biodistribution.³⁷ The depot of subcutaneously injected Alexa-DEX particles were visualized via IVIS up to 18 days post-injection (Fig. 1A, Fig. S1). Notably in the evaluation of particle biodistribution, the draining (inguinal) lymph nodes (LNs) from particle treated mice exhibited increased fluorescence compared to PBS treated mice (Fig. 1B, Fig. S2), indicating particle trafficking to a crucial site for immune cell education. Importantly, the fluorescent signals observed in the heart, spleen, liver, and lungs of Alexa-DEX treated mice were comparable to or below those in PBS controls (Fig. S3). Fluorescent signals observed in the kidneys were only marginally increased in early timepoints and not maintained at later timepoints. This indicates that particles traffic to critical immune sites like the lymph nodes, but do not accumulate in areas where off-target effects could be problematic.

Fig. 1: Spray dried fluorescent Alexa-DEX particles formed a subcutaneous depot and were taken up by antigen-presenting cells.

(A) PBS-normalized radiance from IVIS live-animal images of the Alexa-DEX particle (Alexa-DEX Ps) depot, where radiance = 1 for PBS controls. (B) PBS-normalized radiance from IVIS images of the draining (inguinal) LN from cohorts of animals collected on days 1, 3, 7, 14, and 18, where radiance = 1 for PBS controls. (C) Flow cytometry results from the subcutaneous depot from each cohort, tracking immune cell (CD45+) infiltrates and their Alexa-DEX P uptake. DCs = dendritic cells. iMCs = inflammatory macrophages. iMOs = intermediate monocytes. Neuts = neutrophils. RTMs = resident tissue macrophages. Data are shown as mean ± standard deviation.

To evaluate particle uptake by APCs in vivo, flow cytometry analysis was performed on the collected subcutaneous depots with a focus on CD45+ immune cells (Fig. 1C, Fig. S4). The percentage of particle-positive cells was tracked over time, showing sustained influx and uptake by immune cells, peaking on day 3 and resolving by day 18. Particles were detected in various APC populations, including B cells, DCs, resident tissue macrophages (RTMs), inflammatory macrophages (iMCs), neutrophils (neuts), and intermediate monocytes (IMOs). DCs were particularly prominent, comprising 11–35% of CD45+ immune cells across all timepoints and demonstrating sustained trafficking to the depot as well as particle uptake. At early timepoints, both DCs and neutrophils were prevalent in the depot and had the highest frequencies of particle uptake (Fig. 1C). Indeed, 3 days post-injection, 66% of DCs and 80% of neutrophils were particle-positive. Although, neutrophil levels and particle uptake declined by the next timepoint, DC trafficking to the depot and particle uptake persisted. Day 7 post-injection, DCs had the highest frequency of uptake, with 44% of DCs shown to be particle-positive. This robust and sustained uptake by APCs, particularly DCs, suggests that this particle platform is well-suited for inducing antigen-specific immune tolerance, as DCs play pivotal roles in orchestrating antigen-specific responses and promoting regulatory pathways.

Ace-DEX Particles Elicit Functional Tolerance and Tolerogenic T Cell Phenotypes in an Adoptive Transfer Model

Having established the ability of APCs to take up fluorescently labeled Ace-DEX particles in vivo, a range of Rapa and peptide cargoes were successfully formulated in spray dried Ace-DEX particles. Encapsulation efficiencies (EE) and encapsulate loadings were measured across particle batches (Table S1) and used to ensure consistent dosing of active compounds (Rapa and/or peptide) across studies. Particles exhibited a collapsed morphology and appropriate size to passively target phagocytic APCs²⁴ (Fig. 2A). In vitro release assessment of ovalbumin(323–339) (OVA) and myelin oligodendrocyte glycoprotein (35–55) (MOG) peptides each demonstrated minimal release over 1 week at physiologically relevant temperature and pH 7.4 (Fig. S5B). Previous reports demonstrate sustained Rapa release over the course of weeks from Ace-DEX particles at neutral pH and rapid release at pH 5 (mimicking the endosome) ³⁴. As such, Ace-DEX particles are designed for triggered release of therapeutic cargo after uptake by APCs.

Fig. 2: Particles loaded with 1% wt/wt of Rapa and OVA demonstrate induction of tolerance in an OT-II adoptive transfer model.

(A) Scanning electron microscopy of Ace-DEX particles. (B) OT-II adoptive transfer model schematic. The model was induced with transfer of OT-II CD4 T cells on day 0, and recipient mice received treatments on days 1, 4, and 7. The spleen on day 10 represents collection of splenocytes from recipient mice. (C-F) Assessment of particles with 1% loading in the OT-II adoptive transfer model. Validated Rapa doses of 13 μg and 5 μg were used, and OVA doses of 12 μg and 4 μg were used. Particles are noted as Ps in legend. Graphs display individual data points, mean, and SD. Significance relative to indicated control was determined via ANOVA (*p<0.05, **p<0.01, ****p<0.0001). (C) Splenocyte suppressor assay results, evaluating mechanistic tolerance in the whole splenocyte population via inhibition of proliferation of freshly isolated OT-II CD4 T cells (measured by dilution of CFSE stain). Significance is relative to blank particles. (D) T cell suppressor assay results, evaluating mechanistic tolerance in T cells isolated from spleens of treated mice via inhibition of proliferation of freshly isolated OT-II CD4 T cells (measured by dilution of CFSE stain). Irradiated splenocytes are included for antigen presentation. Significance is relative to uninhibited OT-II T cells, which contains CSFE-stained OT-II cells alone with irradiated splenocytes. (E) Flow cytometry results for frequency of Th17 (IL-17+) adoptively transferred cells in the spleen. Significance is relative to blank particles. (F) Flow cytometry results for frequency of T_regs (Foxp3+) in differentiated (CD44+) CD4+ host cells. Assay described in Schematic S1. Each maker represents a mouse, N=5.

Particle therapies were optimized in an OT-II adoptive transfer model wherein CD4+ T cells recognizing OVA were isolated and transferred to a congenic recipient (Fig. 2B). The quality control of the adoptively transferred cells demonstrated a 91.8–93.7% purity of CD4+ T cells, and 89.2% of CD4+ T cells were confirmed naïve on day 0 (Fig. S6A–D). To evaluate the induction of tolerance from particle therapies delivered by subcutaneous injection, suppressor assays (Schematic S1) were performed, where splenocytes or T cells from treated (with particle or control group) mice were co-cultured with freshly isolated and CFSE-stained OT-II CD4+ T cells in the presence of OVA. Proliferation of CFSE-stained OT-II CD4+ T cells was assessed by flow cytometry (Fig. S6E), where inhibition of proliferation indicates functional tolerance.

After an initial dose titration (Table S2, Fig. S7–S8), particles with 1% wt/wt loading of Rapa and OVA (Rapa+OVA particles) were selected for further evaluation due to significant (p<0.05) inhibition of proliferation of OT-II T cells in a suppressor assay and trends towards increased T_regs in the transferred cells, including central memory Tregs (CD44+ CD62L+ Foxp3+ CD4+ T cells), which are expected to be long lasting (Fig. S8). Evaluation of Rapa+OVA particles was done with additional controls with matched Rapa and OVA doses (13 μg Rapa, 12 μg OVA) as well as a group receiving an overall reduced dose of the same batch of Rapa+OVA particles (5 μg Rapa, 4 μg OVA) (Table S3). In a suppressor assay with splenocytes from treated mice, all particle treatments containing Rapa achieved significant (p<0.001, p<0.01) reduction of proliferation of OT-II T cells compared to a blank particle control (Fig. 2C). However, in a suppressor assay with purified T cells from treated mice, only the reduced particle dose containing 5 μg Rapa and 4 μg OVA achieved significant (p<0.05) inhibition of proliferation compared to an OT-II T cell control, indicating T cell-specific tolerance (Fig. 2D). Of note, particles were necessary to achieve significant tolerogenic effects as soluble Rapa+OVA treatment performed similarly to negative controls in both suppressor assays (Fig. 2C–D). To gain insight into tolerance mechanisms, flow cytometry was performed to evaluate T cell phenotypes from treated mice (Fig. S9–S10, Fig. 2E–F). In the transferred cells, there was a significant (p<0.01) decrease in the frequency of Th17 cells in mice treated with Rapa particles (13 μg Rapa) or either dose of Rapa+OVA particles (13 μg Rapa + 12 μg OVA or 5 μg Rapa + 4 μg OVA) compared to blank particles (Fig. 2E). There was also a trend towards higher T_reg counts and a decreased ratio of Th17:T_reg in the transferred cells of mice treated with either dose of Rapa+OVA particles (13 μg Rapa + 12 μg OVA or 5 μg Rapa + 4 μg OVA) (Fig. S10). In the host cells of mice treated with the reduced dose of Rapa+OVA particles (5 μg Rapa, 4 μg OVA), there was a trend towards increased frequency of T_regs in the differentiated (CD44+) CD4 T cell population (Fig. 2F). Together, suppressor assay results and flow cytometry analysis indicated the reduced dose of Rapa+OVA particles (5 μg Rapa, 4 μg OVA) was most favorable for an antigen-specific induction of tolerance.

To further validate the 1% loading and the reduced dose of Rapa+OVA particles (4.4 μg Rapa, 4 μg OVA), particles were applied in the OT-II adoptive transfer model again with additional dose-matched controls and further titration of Rapa (Table S4). In a suppressor assay with splenocytes from treated mice, both Rapa and OVA were required in the particle formulation to elicit significant (p<0.01) reduction of proliferation of OT-II cells compared to a blank particle control (Fig. S11). Furthermore, increasing the amount of Rapa in the Rapa+OVA particles did not show any added benefit towards functional tolerance, but when the Rapa loading was reduced, there was no significant inhibition of OT-II T cell proliferation (Fig. S11). Taken together, evaluations of Ace-DEX particles in the OT-II adoptive transfer model serve as a dose optimization of particle cargoes and demonstrate that Ace-DEX particles loaded with Rapa and antigenic peptide induce functional tolerance as well as tolerogenic T cell phenotypes.

Ace-DEX Particles Containing MOG and Rapa+MOG Elicit Significant Therapeutic Effects When Given at Disease Onset

Particle therapies were assessed in an early therapeutic model of progressive EAE with particle loading and dosing informed by the optimization in the OT-II adoptive transfer model, resulting in 4.4μg Rapa and 3.9μg MOG dosages (Table S5). Mice received treatments via subcutaneous injection beginning at symptom onset, and the study ended at the expected disease peak on day 20 (Fig. 3A). Treatment with MOG particles and Rapa+MOG particles resulted in significantly (p<0.01 and p<0.05, respectively) reduced clinical scores compared to blank particles at the end of the study timeline, outperforming soluble Rapa+MOG and Rapa particles (Fig. 3B). Soluble Rapa+MOG and Rapa particles resulted in reduced clinical scores compared to blank particles, but this reduction was not significant. Furthermore, the clinical score area under the curve (AUC) was significantly lower in the MOG particle and Rapa+MOG particle groups (p<0.0001 and p<0.001, respectively) compared to the blank particle and vehicle groups (Fig. 3C). Soluble Rapa+MOG treatment resulted in a significantly (p<0.05) lower AUC than blank particle treatment, but there was no significant difference in day 20 scores. Notably, in the soluble group, some mice appeared to respond to therapy, while some did not (3/5 responders, Fig. S12). Looking at individual scores within each treatment group, all therapy-loaded particles (Rapa particles, MOG particles, or Rapa+MOG particles) resulted in a 100% response rate, while all other groups had multiple mice succumb to disease due to weight loss (Fig. S12). Although the Rapa particle group did not have any mice succumb to disease, some mice had developed severe symptoms, with multiple mice at a score of 3 at the end of the study, indicating mice had developed full hind limb paralysis. By comparison, all mice from the MOG particle group and Rapa+MOG particle group had a clinical score ≤2.5 at the end of the study, indicating all mice had some functional use of their hind legs, and most mice in these two groups had a clinical score ≤1, indicating no hind limb inhibition (Fig. S12, Table S6). In addition to clinical score, weight serves as an indicator of animal health, and mice in the MOG particle and Rapa+MOG particle groups maintained higher body weights on average compared to other groups, with the MOG particle group demonstrating the highest weights (Fig. 3D). Overall, both the MOG particles and Rapa+MOG particles demonstrated a dramatic therapeutic effect, characterized by substantially improved mobility and a 100% response rate to therapy.

Fig. 3: Evaluation of particle therapy in early therapeutic model of EAE.

(A) EAE model timeline and legend depicting treatment groups. Particles are noted as Ps in legend. (B) Summary of clinical scores. Graphed: mean and SEM. Significance for clinical scores on day 20 relative to blank particles was determined via ANOVA (* p<0.05, ** p<0.01; N=5). (C) Summary of clinical score AUCs. Graphed: mean and SEM. Significance relative to blank particles (*) or vehicle (#) was determined via ANOVA (* p<0.05, ## p <0.01, *** p<0.001, **** or #### p<0.0001). (D) Summary of weights reported as percent of day zero (D0) weight. Graphed: mean and SEM. Mice which have been euthanized are removed from the data for subsequent timepoints but are included in the summary data for prior timepoints. Significance between groups is not assessed due to removal of euthanized mice from later timepoints of the weight assessment. (E-F) Flow cytometry results from spinal cords taken on D20. Graphs display individual data points, mean, and SD. Significance relative to blank particles was determined via ANOVA (* p<0.05, ** p<0.01). Frequency of (E) CD19+ cells (B cells), (F) CD4+ T cells, (G) CD8+ T cells in the cord. (H-I) Frequency of Foxp3+ cells out of (H) CD19+ cells, (I) CD4+ T cells, and (J) CD8+ T cells.

Flow cytometry analysis was performed on spinal cords and cervical LNs (cLN) collected on day 20 from all groups that had 3 or more survivors as well as age-matched healthy mice never induced with EAE (Fig. S13). Spinal cords of mice treated with particles containing a therapeutic (Rapa particles, MOG particles, or Rapa+MOG particles) had substantially reduced immune cell infiltrates compared to mice treated with blank particles. Reduced frequencies and counts of B cells (CD19+), CD4 T cells (CD4+TCRβ+), and CD8 T cells (CD8+TCRβ+) were all observed, with the reductions in B cells and CD8 T cells being statistically significant (p<0.0001 for frequencies, p<0.001 and p<0.01 for B and CD8 T cell counts) compared to blank particles (Fig. 3E–G, Fig. S14). Furthermore, of the immune cells present, a higher percentage were regulatory (Foxp3+) in mice treated with therapeutic particles (Rapa particles, MOG particles, or Rapa+MOG particles), with significantly (p<0.05) increased frequencies of Foxp3+ B cells (Fig. 3H) and a trend towards increased frequencies of Foxp3+ T cells, especially CD8 T cells (Fig. 3I–J). Additionally, a trend towards decreased Th1 and Th17 counts in mice treated with therapeutic particles was observed, and increased PD-1 expression as a marker for exhaustion was noted in T cells from mice treated with MOG particles, with significance (p<0.05) in CD8+ T cells compared to mice treated with blank particles (Fig. S14). In the cLN, a significant (p<0.05) increase was observed in the total number of cells in mice treated with MOG particles compared to blank particles, and no significant differences were observed in the frequencies of different immune cell populations within the cLN of all mice that had been induced with EAE (Fig. S15).

Ace-DEX Particles Containing MOG and Rapa+MOG Elicit Significant Therapeutic Effects when Administered at Peak Disease

Due to the significant therapeutic effect achieved from MOG particle and Rapa+MOG particle treatments in the early therapeutic EAE model, both particle formulations were evaluated in a late therapeutic model where mice were treated at peak disease and monitored for over 20 days after the end of the treatment period (Table S7, Fig. 4A). Compared to treatment with blank particles, a significant and sustained reduction in clinical score was achieved from both MOG particle and Rapa+MOG particle treatments (p<0.05 at earlier timepoints, p<0.01 and p<0.001 at intermediate timepoints, and p<0.0001 at later timepoints) (Fig. 4B). Furthermore, both therapeutic particle treatments demonstrated a significant (p<0.05) effect relative to blank particles as early as 2 days after the first treatment, and mice continued to recover after the treatment period had ended. By day 45, mice in the MOG particle and Rapa+MOG particle groups had average clinical scores of approximately 1 (day 45 average score = 1 for Rapa+MOG particles, and day 45 average score = 0.94 for MOG particles). This result indicates a dramatic recovery from severe disease—where mice had severe hind limb inhibition or full hind limb paralysis at the beginning of the treatment period—to mild disease characterized by inhibition of the tail and regained hind limb function. These results are further exemplified by the significantly (p<0.0001) lower clinical score AUCs for MOG particles and Rapa+MOG particles compared to blank particles (Fig. 4C). Looking at individual scores (N=8) in the MOG particle and Rapa+MOG particle groups, except for one mouse at a score of 2 in the Rapa+MOG particle group, all mice had a score in the range of 0.5–1.5 at the end of the study, indicating mice re-gained the ability to walk on all 4 limbs without dragging their toes (Fig. S16A–C, Table S6). Additionally, both therapeutic particle treatments resulted in 100% survival. In contrast, in the blank particle group, multiple mice succumbed to disease (score=5), and the remaining survivors had generally higher scores than the mice in the therapeutic particle groups, ranging from 1–2.5 (Fig. S16A–C). In the blank particle group, reductions in average weights were observed on days when mice succumbed to disease, and in all groups, surviving mice recovered in weight, indicating their conditions were stable (Fig. S16D).

Fig. 4: Evaluation of particle therapy in late therapeutic model of EAE.

(A) EAE model timeline and legend depicting treatment groups. Particles are noted as Ps in legend. (B) Summary of clinical scores (N=8). Graphed: mean and SEM. Across all timepoints, significance for clinical scores was determined via two-way ANOVA, where * indicates significance between blank particles and MOG particles, and # indicates significance between blank particles and Rapa+MOG particles (* or # p<0.05, ** or ## p <0.01, *** p<0.001, **** or #### p<0.0001). (C) Summary of clinical score AUCs. Graphed: mean and SEM. Significance relative to blank particles was determined via ANOVA (**** p<0.0001).

Flow cytometry analysis was performed on spinal cords, cLNs, and spleens collected on day 45 (Fig. S17–S20). Compared to the blank particle group, no significant differences were observed in the assessed cell phenotypes from mice treated with MOG particles or Rapa+MOG particles, though there was a trend towards increased IL-10 expression in B cells from mice treated with MOG particles, particularly in the spinal cord (Fig. S18). Despite no significant differences in phenotypes observed via flow cytometry, there were differences in cytokine secretion from splenocytes after stimulation with MOG peptide. There was a trend towards increased secretion of both inflammatory (IL-17) and anti-inflammatory (IL-4, IL-10) cytokines and a significant (p<0.05) increase in IFNγ from splenocytes taken from mice treated with MOG particles compared to the blank particle group (Fig. S21).

Ace-DEX Particles Containing MOG and Rapa+MOG Do Not Diminish Antibody Response to Influenza Vaccination

To assess if the observed tolerogenic effects of MOG particles and Rapa+MOG particles were antigen-specific, antibody responses to influenza vaccines were evaluated in mice receiving particle treatments, where particles were dosed the same as in the late therapeutic EAE model (Table S7). Mice were immunized with influenza hemagglutinin (HA) adjuvanted with MF59 mimic, AddaVax, ³⁸ prior to receiving MOG particles, Rapa+MOG particles, or PBS (Fig. 5A). After the treatment period, mice that had received MOG particles or Rapa+MOG particles had anti-HA IgG titers comparable to the PBS control (Fig. 5B). Furthermore, no significant differences between groups were observed in titers for IgG subclasses, IgG1 and IgG2c, nor in IgM titers (Fig. 5C–E). In addition to anti-HA IgG titers, hemagglutination inhibition (HAI) titers, were assessed against influenza strain A/California/7/09 (H1N1) after the treatment period. Of note, HAI titers are a common correlate of protection and are used as evaluation for FDA approval of annual influenza vaccines³⁹. No significant differences were observed in HAI titers (Fig. 5F). Moreover, an additional timeline was assessed wherein mice received particle treatments or a PBS control followed by an influenza vaccine (HA+AddaVax). For the timeline where treatment preceded vaccination, MOG particles or Rapa+MOG particles did not reduce titer levels from a subsequent influenza vaccine. Rather, at the end of the study period, total IgG, IgG1, and IgM titers to HA were significantly (p<0.05) increased from mice that had received Rapa+MOG particles compared to PBS control (Fig. S18), which indicate increased Th2 skewing (IgG1) and fast, long-lasting protective immunity (IgM, IgG) ^40–42. These results indicate that MOG particles and Rapa+MOG particles do not diminish antibody responses to prior or subsequent vaccinations, suggesting tolerogenic particle therapies are antigen-specific and do not compromise immune competence.

Fig. 5: Evaluation of immune competence via antibody responses to influenza vaccination administered prior to particle therapy.

(A) Experimental timeline. C57BL/6 mice receive an influenza vaccine (AddaVax+HA) in a prime-boost schedule on days 0 and 21 followed by particle treatments or PBS on days 28, 31, and 34. Indicated by red drops, blood is collected periodically to evaluate titers to COBRA HA. N=5. IM = intramuscular. SubQ = subcutaneous. P = particle. Day 42 titers for (B) total IgG, (C) IgG1, (D) IgG2c, and (E) IgM. (F) HAI titers against A/California/7/09 (H1N1) from blood collected on day 42. Significance was determined via ANOVA (ns = no significance).

Discussion

Tolerogenic Ace-DEX particles were optimized and evaluated for their therapeutic potential in alleviating autoimmunity without broad immunosuppression, assessing functional outputs of induced tolerance and antigen-specificity as well as cellular phenotypes in clinically relevant murine models. The spray drying fabrication method utilized herein is a scalable, continuous process commonly employed in the pharmaceutical industry to make dried powder formulations or controlled release particles ⁴³. Furthermore, the resulting particles are ideal for targeted immunomodulation as they passively target APCs due to their size.²⁴ In prior studies, similarly sized Ace-DEX particles were taken up by DCs and trafficked to LNs ^44–46. Consistent with these prior reports, in the present work, fluorescently labeled spray dried Ace-DEX particles (called Alexa-DEX particles) formed a subcutaneous depot post-injection (Fig. 1A), and APCs were found to enter the depot and take up the particles (Fig. 1C). These Alexa-DEX particle-positive APCs appeared to traffic to the draining LNs (Fig. 1B, Fig. S2). By trafficking to a major germinal center, Ace-DEX particles can contribute to immune education and memory. Additionally, due to the known acid-sensitivity of Ace-DEX, particles are expected to degrade inside the endosome of APCs, resulting in triggered release of particle cargos inside the cell ^{33, 47}. Indeed, previous reports have highlighted the utility of Ace-DEX particles in achieving intracellular delivery of cargo ⁴⁷. By targeting APCs, the particle therapies optimized in the present work are designed to encourage antigen presentation of the encapsulated peptide in a tolerogenic manner.

Previous studies have illustrated the utility of Ace-DEX particles incorporating a tolerogenic agent and disease-associated peptide to significantly induce antigen-specific tolerance ^{30, 31, 36}; however, this system has not been optimized. For this optimization, an OT-II adoptive transfer model was used (Fig. 2B), allowing for a mechanistic assessment of induced tolerance in a host mouse where a subset of CD4+ T cells have existing memory to an antigen of interest, OVA. This model is more relevant to autoimmune disease than directly treating OT-II mice, in which all CD4+ T cells are specific to OVA, and it provides an efficient timeline for optimization of particle therapy prior to application in a disease model. To assess the induction of tolerance from particle therapy, suppressor assays were performed. When presented with OVA peptide in vitro, newly isolated OT-II CD4+ T cells should proliferate like self-reactive cells driving an autoimmune response. However, if particle treatments induced tolerance to OVA, cells isolated from particle-treated mice are expected to suppress proliferation of the OT-II T cells. As such, inhibition of OT-II CD4+ T cell proliferation serves as a functional measurement of tolerance.

Across experiments, Rapa+OVA particles demonstrated functional tolerance in the suppressor assays, and flow cytometry analysis of T cell phenotypes provided insight to tolerance mechanisms. In the splenocyte suppressor assay, inhibition of proliferation of newly isolated OT-II T cells indicated functional tolerance from the whole splenic cell population. The significant results observed with multiple particle formulations in the splenocyte suppressor assay are encouraging, but the mechanism is unclear and could be due to multiple cell types. Only the optimized dose of the Rapa+OVA particles with 1% loading demonstrated significant results in both the splenocyte suppressor assay and the T cell suppressor assay (Fig. 2; Schematic S1), where the T cell suppressor assay more specifically assesses T cell dependent functional tolerance, which is expected to be antigen-specific and long lasting. Indeed, there has been a substantial research focus on the induction of tolerogenic T cell phenotypes (e.g. T_regs) for antigen-specific strategies against autoimmune disease¹⁸. Furthermore, flow cytometry analysis demonstrated trends in increased T_regs (Foxp3+ CD4+ T cells) in transferred cells across experiments (Fig. S4 and S6), including at the optimized dose of Rapa+OVA particles, where significant reductions in the inflammatory Th17 phenotype were also observed in the transferred cells (Fig. 2E). This result was encouraging, as the phenotypes reflected in the transferred cell population are antigen-specific to OVA. Furthermore, Th17 cells have been implicated in the development and progression of autoimmune diseases such as MS, and their reduction can be indicative of increased antigenic tolerance ^{14, 20}. In the host cells, the optimized dose of Rapa+OVA particles resulted in a trend towards increased T_regs in the differentiated (CD44+) CD4 T cells, which are expected to be long lasting and beneficial for the maintenance of tolerance ^{48, 49}. Taken together, the significant reduction in Th17 cells and trends towards increased T_regs may explain the functional tolerance observed in the splenocyte and T cell suppressor assays from the optimized dose of Rapa+OVA particles.

The particle optimization performed in the OT-II adoptive transfer model highlights the importance of dosage in the induction of antigenic tolerance. Generally, lower doses demonstrated greater measures of immune tolerance. Although more than one component of the therapy was titrated together, the lower peptide dose likely contributed to the enhanced tolerogenic effect. It is possible the lower antigen dose delivered to APCs results in lower TCR engagement, which is known to be tolerogenic ⁵⁰. At the evaluated peptide and Rapa doses, incorporation of Rapa in the particle therapy was needed to achieve a measurable tolerogenic effect in the OT-II adoptive transfer model, but increasing Rapa beyond a certain point did not result in further tolerogenic effects (Fig. 2, S7). This result is consistent with reports of Rapa being utilized to enhance tolerogenic therapies in models of autoimmune disease ^30–32. Additionally, at the optimized dose, Rapa particles without antigen did not elicit significant effects in the splenocyte suppressor assay (Fig. S7), indicating that the tolerogenic effects of the optimized dose are likely antigen-dependent. Based on the results observed in the OT-II adoptive transfer model, encapsulated Rapa and peptide doses were established for therapeutic application in clinically relevant timelines of the EAE model. Across all studies performed after optimization in the OT-II model, the doses of Rapa and peptide remain consistent, and encapsulate loadings were used to ensure consistent dosing of active components (e.g. Rapa and/or peptide) across batches, resulting in only minor adjustments to particle dosing (Tables S1, S4, S5, S7). Ace-DEX particles are non-immunogenic and do not provide any therapeutic benefit when administered alone (Fig. 3B).

EAE is the gold standard prototypical model for MS used in pre-clinical translational research ³⁵. To be as clinically relevant as possible, Ace-DEX particles were applied in different timelines of EAE: an early therapeutic model and a late therapeutic model. In the early therapeutic model, mice receive treatments beginning at the onset of disease, and the study ends at the expected disease peak on day 20 (Fig. 3A). Compared to a commonly employed approach seen throughout literature of treating EAE mice before symptom onset, the early therapeutic model has higher clinical relevance as MS cannot be diagnosed before symptom development due to a lack of known biomarkers or observable measures of pre-symptomatic disease ⁵¹. It has been well established in the clinic that early treatment of MS leads to better outcomes, and this approach is mimicked in the early therapeutic EAE model; however, it is difficult to achieve early, accurate diagnosis as patients with MS can present with a variety of possible symptoms ^{51, 52}. Furthermore, MS is a chronic condition with no cure, and patients are living with all stages of disease. It is therefore imperative to evaluate possible new treatment approaches with consideration for multiple stages of disease. In the late therapeutic model, waiting until peak disease to begin treatment when mice are in the score range of 2.5–3.5 is relevant to patient populations living with severe symptoms and chronic disease. In addition to waiting until peak disease, it is imperative to assess the maintenance of tolerance beyond the treatment period as a long-lasting therapy is more translatable. In the late therapeutic model, mice were monitored for over 20 days after the end of the treatment period (Fig. 4A), translating to approximately 2 years for a human ⁵³.

In the early therapeutic EAE model, MOG particles and Rapa+MOG particles were the only treatments to elicit stable weights and a significantly lower clinical score compared to controls at the end of the study period, outperforming soluble Rapa+MOG and Rapa particles (Fig. 3). Accordingly, mice treated with MOG particles or Rapa+MOG particles demonstrated improved mobility, quality of life, and overall health compared to negative controls (blank particles, vehicle). These responses to therapy are highly impactful, and the therapeutic benefit of particles containing MOG alone is unprecedented. Looking at other reports of polymeric particle formulations applied to a MOG-induced EAE model, the inclusion of tolerogenic agents (e.g. Rapa and/or cytokines) appears essential to the prevention of symptom progression. In these reports, particles administered prior to or at the onset of symptoms minimize the worsening of disease, though some progression of symptoms is often observed^54–57. By comparison, Ace-DEX particles administered at symptom onset resulted in average scores less than 1 (0.9 for Rapa+MOG particles and 0.2 for MOG particles), demonstrating comparable or improved therapeutic benefit from a simpler formulation and, compared to some reports, a preferred route of administration (e.g. SubQ vs intranodal).

To gain insight into the mechanisms contributing to the significant therapeutic benefit achieved by MOG particles and Rapa+MOG particles, flow cytometry analysis was performed on spinal cords and cLNs, representing the site of disease and the respective draining LN. As immune infiltrates in the cord are known to contribute to the progression of EAE and both Th1 and Th17 phenotypes are implicated in autoimmunity ^{58, 59}, it is noteworthy that therapeutic particles lead to substantially reduced infiltrates, including inflammatory phenotypes, in the spinal cord (Fig. 3, S10). Although more work is warranted to determine to what extent particle treatments prevent autoreactive cells from entering the cord or stop them from persisting once in the cord, the latter is likely to occur since immune infiltration begins prior to the onset of symptoms and the treatment period evaluated herein. In the MOG particle group, significantly higher cell counts were observed in the cLN without a reduction in frequencies of assessed B and T cell phenotypes compared to mice treated with blank particles (Fig. S11), a sign of possible egress of immune infiltrates from the spinal cord. Egress has been previously described as a mechanism to combat autoimmunity in models of type 1 diabetes and MS, wherein pathogenic immune infiltrates are purged from the site of action to achieve remission ^60–62. Furthermore, in these reports, inflammatory populations have been observed in secondary lymphoid tissues after egress (i.e. spleen or LNs), but they do not appear to re-infiltrate the site of disease ^60–62. Our evaluation therefore suggests that treatment with MOG particles likely facilitates the withdrawal of inflammatory immune cells from the spinal cord into the local LN.

To further evaluate mechanisms of induced tolerance in the early therapeutic EAE model, the frequencies of tolerogenic phenotypes were also assessed in the spinal cords. The frequency of Foxp3 expression in B cells was significantly increased compared to the blank particle group (Fig. 3H). This is notable as Foxp3+ B cells have been utilized as a treatment in murine models of autoimmunity and demonstrated beneficial tolerogenic effects ⁶³. Regarding T cells, Foxp3 expression is an established hallmark characteristic of T_regs, especially in CD4 T cells but with recognition in CD8 T cells as well ⁶⁴. Compared to blank particles, there was a trend towards increased frequencies of Foxp3+ T cells, especially CD8 T cells, in spinal cords from mice treated with therapeutic particles (Fig. 3I–J), indicating a trend towards increased T_regs. In addition to regulatory phenotypes, exhaustion of autoreactive T cells can contribute to tolerance ⁶⁵. To evaluate T cell exhaustion, the median fluorescence intensity (MFI) of PD-1 was measured. Notably, spinal cords from mice treated with MOG particles demonstrate increased PD-1 MFI in both CD4 and CD8 T cells compared to the blank particle group, and the increase is significant within CD8 T cells. Together, increased frequencies of regulatory phenotypes (B_reg-like and T_regs) and increased T cell exhaustion in the spinal cords from particle therapy can both serve as signs of induced tolerance at the site of action.

In the late therapeutic EAE model, MOG particles and Rapa+MOG particles both elicited a significant and sustained therapeutic effect in a clinically relevant model for severe MS, with mice going from exhibiting hind limb paralysis to regaining functional use of their hind limbs (Fig. 4). This result is highly impactful and largely unprecedented in the literature. Across reports of polymeric particles applied to MOG-induced EAE, most therapies are administered prior to or at symptom onset, and in cases where therapies are applied at later stages of disease, the effect sizes are not as large or monitored as long as the ones observed herein ^54–57. Furthermore, these reports detail numerous particle formulations made with poly(lactic-co-glycolic) acid (PLGA). Although PLGA has an extensive history as a drug carrier, the striking reduction in clinical score observed herein and the previously described therapeutic effects³⁶ from Ace-DEX particles strongly outcompete PLGA systems (Table S8), and future work is planned to further compare the polymer systems in the induction of immune tolerance. Overall, given the difficulty in treating severe MS in the clinic ⁵¹, the reversal of disease achieved by tolerogenic Ace-DEX particles is incredibly promising and may inform future development of tolerogenic particles to combat autoimmune disease.

Flow cytometry analysis was performed to gain mechanistic insight in the late therapeutic model. No significant differences in cellular phenotypes were observed in the spinal cord, cLN, or spleen; however, there was a trend towards increased IL-10 expression in B cells, which may indicate increases in a B_reg phenotype (Fig. S14–S16). It is possible the blank particle group exhibited survivor bias, but it is more likely that no significant differences were observed due to the disease progression timeline. Immune cell infiltrates, such as leukocytes in the spinal cord, are more prevalent at earlier stages of EAE through peak disease and less pronounced at later, chronic disease stages ⁵⁸. Though no significant differences were observed in cell phenotypes via flow cytometry analysis, antigen recall of splenocytes demonstrated increased secretion of both inflammatory and anti-inflammatory cytokines from mice treated with MOG particles (Fig. S17). Given the sustained remission of disease observed in the MOG particle group and the mechanistic findings of the early therapeutic EAE model, the antigen recall results may further indicate egress of reactive immune cells from the site of action to secondary lymphoid tissue, namely the spleen. Overall, our tissue analyses across both EAE timelines encompass major strides in investigating mechanisms of tolerance from Ace-DEX particle therapies, implicating multiple cell types and mechanisms, including egress and exhaustion of inflammatory cells as well as induction or maintenance of tolerogenic phenotypes. Future work including additional timepoints and antigen recall on cells isolated from the spinal cord and cLN would lead to even greater insight.

Comparing results across the murine models utilized herein, the inclusion of encapsulated Rapa was necessary to achieve a tolerogenic effect in the OT-II adoptive transfer model, but in both therapeutic timelines for EAE, both MOG particles and Rapa+MOG particles achieved similar significant therapeutic effects. The absence of a pre-existing inflammatory response in the OT-II model in contrast to the disease-relevant EAE model may explain the discrepancy in results observed from peptide-only particle therapy. Additionally, the discrepancy between models further supports egress as a mechanism contributing to the therapeutic effect of MOG particles applied in EAE as egress has been reported to occur in an inflammation-dependent manner ⁶⁰. Despite the discrepancy between models, the dose optimization performed in the OT-II adoptive transfer model likely contributed to the success of MOG particles combatting EAE. In previous reports of antigen-loaded Ace-DEX particles applied subcutaneously to EAE models, co-encapsulation of tolerizing agents, such as Rapa or dexamethasone, were necessary to enhance tolerogenic effects of the particle therapy, though it should be noted that the particles in these reports were made with a different fabrication technique (emulsion versus spray drying) ^{31, 36}. Similarly, in additional reports of polymeric particle treatments injected into LNs, particles containing MOG alone achieved a therapeutic effect in an EAE model, but the inclusion of Rapa notably improved the effect ³². Compared to these previous reports, the therapeutic peptide dose utilized in the present work is notably lower (3.9μg MOG vs 14.3–17.6μg MOG or PLP ^{31, 32, 36}), and lower antigen doses are expected to be more tolerogenic as antigen dose affects T cell differentiation and tolerogenic/inflammatory bias ^{66, 67}. Overall, the results in the late therapeutic model stand out in the field of particle therapies applied to EAE due to the clinically relevant timeline, efficacy of particles loaded with peptide alone, and dramatic effect size. It is a common practice to treat EAE mice before the onset of symptoms, and when treatments are administered at later timepoints, the difference in scores before and after treatment is rarely as dramatic as the score difference >2 observed herein ^{20, 32, 54, 68, 69}, indicating the results observed herein are unprecedented and highly impactful.

The efficacy of therapeutic particles in clinically relevant timelines of EAE is incredibly encouraging, but it is also imperative to assess immune competence in mice receiving tolerogenic therapy. Notably, DMTs currently approved for MS often cause patients to become immunocompromised, and many MS patients receiving necessary standard of care therapies are unable to develop adequate antibody responses to vaccines, further leaving them vulnerable to infection ^10–12. As such, tolerogenic Ace-DEX particles and influenza vaccines were both applied in vivo to determine if mice receiving particle treatments can maintain a vaccine response comparable to influenza-vaccinated mice not receiving particle therapy. Two timelines were utilized: 1) influenza vaccine followed by particle treatments or PBS and 2) particle treatments or PBS followed by influenza vaccine.

When particle treatments were administered after an influenza vaccine series, mice treated with MOG particles or Rapa+MOG particles had a similar antibody response to influenza antigen HA as mice that were given PBS after influenza vaccination. After the treatment period, no significant differences in HA titers were observed in different antibody isotypes, including total IgG, IgG1, IgG2c, and IgM (Fig. 5). Overall, the observed titers indicate a strong humoral response to the influenza vaccine, and the higher IgG1 titers compared to IgG2c indicate Th2 skewing, which is a common response to squalene adjuvants like MF59 mimic, AddaVax ⁷⁰. Furthermore, no significant differences were observed in HAI titers against A/California/7/09 (H1N1) (Fig. 5F), indicating that mice in all groups had comparable levels of neutralizing or inhibitory titers against A/California/7/09 (H1N1) and are expected to maintain comparable levels of protection against this strain of influenza virus ^{71, 72}. In the other timeline, when particle treatments were given before an influenza vaccine, no decreased antibody titers to HA were observed from particle treated mice compared to PBS control (Fig. S18). On the contrary, mice pre-treated with Rapa+MOG particles had increased IgG, IgG1, and IgM titers compared to mice pre-treated with PBS before influenza vaccination. The increased IgG and IgM titers indicate a potent humoral response, and the increased IgG1 indicates further Th2 skewing. These results are in line with previous reports of pre-treatment with mTOR inhibitors like Rapa leading to increased antibody responses to influenza vaccine antigens. In a randomized control trial with elderly participants (age 65 years or older), patients who received a combination of mTOR inhibitors 2 weeks prior to an influenza vaccine had an increased serologic response and a reduced self-reported infection rate compared to a placebo group who received influenza vaccines without mTOR inhibitor pre-treatments ⁷³.

Both timelines evaluating tolerogenic Ace-DEX particles and influenza vaccines represent a crucial assessment for future clinical translation of tolerogenic particle therapies to treat autoimmune disease without compromising immune competence. Our results indicate that the significant therapeutic effect observed in the EAE model from MOG particles and Rapa+MOG particles is likely antigen-specific with strong potential to induce immune tolerance without broad immunosuppression. More work is warranted to evaluate influenza vaccine efficacy together with the maintenance of restored antigenic tolerance in particle-treated EAE mice, but the current results demonstrate promise for the development of an antigen-specific therapy to combat autoimmune disease without interfering with immune competence.

Conclusion

There is a strong unmet clinical need for antigen-specific therapies to combat autoimmune diseases without broad immunosuppression. In the present work, a particle-based therapeutic vaccine was optimized for successful applications in a prototypical model of MS. MOG particles and Rapa+MOG particles achieved significant therapeutic effects in the EAE model – both when mice were treated at symptom onset and when mice were treated during peak disease, representing multiple clinically relevant stages of disease. The therapeutic effect achieved in the late therapeutic model is particularly exciting as it is often challenging to control severe disease, and the dramatic effect size observed herein is largely unprecedented. Furthermore, to our knowledge, this is the first report of peptide-only particles generating a significant tolerogenic response. Additionally, due to the antigen-specificity of the particle therapies, tolerogenic immune memory is expected to be long lasting and not to interfere with immune responses against infectious disease. Indeed, MOG particle and Rapa+MOG particle therapies did not inhibit antibody responses to influenza vaccination. To our knowledge, this is the first report of polymeric particles inducing tolerance in a clinically relevant model of MS while also demonstrating compatibility with influenza vaccines without measurable loss of protection. With further development, tolerogenic Ace-DEX particles hold promise in alleviating autoimmune disease without the broad immunosuppression that currently makes patients vulnerable to infection.

Methods

Unless otherwise noted, all materials were obtained from Sigma Aldrich (St. Louis, MO).

Acetalated Dextran Synthesis

Ace-DEX was synthesized according to previously established methods ³³. The final Ace-DEX polymer was lyophilized and stored at −20°C. Cyclic acetal coverage (CAC) was assessed by nuclear magnetic resonance (Inova 400 MHz). Ace-DEX synthesized with a reaction time of 2.25 hr yielded polymer with 60–65% CAC, herein referred to as 60 CAC and 65 CAC batches. Alexa-DEX was formed as described previously.³⁷ Briefly, dextran was aminated via a microwave reaction with diaminobutane which was then reacted with Alexa-Fluor 647-NHS-ester (Thermo Fisher) for 2 hr in 0.1 M MES MES (20sulfonatoehyl)morpholin-4-ium) buffer (pH = 6). This formed “Alexa-DEX”, dextran with a fluorescent terminus. Alexa-DEX was acetalated with a 20:80 wt% of Alexa-DEX:dextran with the same reaction conditions as 60–65% CAC Ace-DEX to form Alexa-DEX polymer.

Microparticle Fabrication

A variety of Ace-DEX particles were made via spray drying, encompassing a range of peptide and Rapa loadings. Peptides ovalbumin (323–339) (OVA) and myelin oligodendrocyte glycoprotein (35–55) (MOG) were used. Particles were all made with a spray drying fabrication process with a B-290 spray dryer (Buchi, New Castle, DE). Feed solutions were made for each particle batch by dissolving Ace-DEX at 2.5 mg/mL and including other cargoes in the solution according to their solubility and desired loading. The solvent systems utilized to create feed solutions were selected based on peptide solubility and applicability to the spray drying process. For particles used in studies with Alexa-DEX, 99% EtOH and 1% dimethyl sulfoxide (DMSO) were used. For particles used in studies with OVA, 10% v/v basic water in ethanol was used, and for particles used in studies with MOG, 1–2% v/v phosphate buffered saline (PBS) in ethanol was used. Feed solutions were sprayed with the following settings: 10mL/min feed flow rate, 75°C inlet temperature, and 55mm Q-flow meter. Particles were sprayed directly into a falcon tube for efficient collection then suspended in a sucrose solution (10mg/mL particles in 20mg/mL sucrose), aliquoted, lyophilized, and stored at −20°C.

Particle Characterization

Particle characterization included assessments of peptide and Rapa EEs, visualization of particle morphology, and detection of endotoxin. EE assessments were done in triplicate and endotoxin assessments were done in duplicate. EE of peptides were assessed via a fluorescamine assay. Particles and peptide standards were dissolved in 0.1N HCl, then 0.1M borate buffer was added to adjust the pH for the fluorescamine reaction. Peptide concentration was measured according to fluorescence (excitation/emission = 390nm/460nm) with respect to peptide standard curves, and blank particles were subtracted as background. Rapa EE was determined via high performance liquid chromatography (HPLC, 1100 Series, Agilent Technologies, Santa Clara, CA) with a C18 column (Aquasil 77,505–154,630 C18 Column, 5 μm Pore; 150 mm L x 4.6 mm ID, Thermo Fisher Scientific) using an 80:20 acetonitrile:water solvent system. Particle morphology was visualized via scanning electron microscopy (SEM, Hitachi S-4700 Cold Cathode Field Emission). To evaluate endotoxin, a Pierce LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fisher Scientific, Waltham, MA) was used according to the manufacturer’s instructions, and all samples were below the limit of quantification (<0.1 EU/mg).

Assessment of Peptide Release from Particles

Peptide release from 1%Rapa 1% OVA particles was evaluated in triplicate in vitro. Particles were suspended in PBS (pH 7.4) at 1mg/mL in Eppendorf tubes. Tubes were placed on a shaker plate and maintained at 37°C, and timepoints were taken at 0, 0.5, 2, 4, 8, 24, 48, 72, 96, and 168 hr. At each timepoint, suspensions were briefly vortexed before an aliquot was collected. Aliquots were then centrifuged, and the supernatant and pellet were collected separately. The remaining peptide in the pellet was measured by fluorescamine assay. Percent released over time was calculated by subtracting the remaining peptide from the original peptide loading.

Animal Study Design

Tolerogenic Ace-DEX particles were evaluated in three murine models: OT-II adoptive transfer, EAE, and influenza vaccination models (details on each below). The three models were employed to evaluate tolerance mechanisms, the potential for therapeutic benefit against MS, and immune competence, respectively. Two to three experiments were conducted in each model, allowing for repeated evaluation of the particle therapy under a range of conditions. Sample sizes of 5 or more mice were employed according to standard in the field. Power calculations were performed for EAE considering possible effect sizes at different timepoints of disease progression. All collected data was included in the analyses, and no outliers were excluded. All experiments were randomized. When assigning groups in the EAE studies, mice were first distributed evenly based on weight and/or clinical score as applicable prior to the start of treatment. Groups were then assigned with a random number generator. Clinical scoring was performed according to metrics from Hooke laboratories in a single-blinded manner (Table S6). Alexa-DEX particles were evaluated in healthy mice to investigate APC uptake and biodistribution of the particle platform employed herein. All animal procedures and humane endpoints were in accordance with and approved by the Institutional Animal Care and Use Committee (IACUC) of the University of North Carolina at Chapel Hill.

Biodistribution and Alexa-DEX Particle Trafficking

Alexa-DEX particles (2 mg) were injected in each subcutaneous flank of 10–12-week-old Balb/c mice (Jackson Laboratory). Five cohorts of animals were humanely euthanized at different timepoints (1, 3, 7, 14, and 18 days) to evaluate Alexa-DEX biodistribution within tissues, with each cohort containing N=3 Alexa-DEX particle-treated mice and N=2 PBS-treated mice. At each timepoint, spleens, kidneys, hearts, lungs, livers, and draining (inguinal) LNs were imaged on an In Vivo Imaging System (IVIS) Lumina. The subcutaneous depot was collected as described previously and processed into single-cell suspension for flow cytometry analysis.^{37, 74} The cohort for the day 18 timepoint was imaged on days 0, 1, 2, 3, 4, 7, 10, 14, and 18 for progression of the subcutaneous depot of particles.

OT-II Adoptive Transfer Model Induction and Dose Optimization Studies

To induce the OT-II adoptive transfer model, CD4 T cells were isolated from OT-II mice and transferred to a congenic recipient. OT-II mice were originally purchased from Jackson Labs (Bar Harbor, ME) and bred in house. Both male and female OT-II mice were used. For congenic recipients, male C57BL/6 (B6) Thy1.1 mice (Jackson Labs) were used. On day 0, spleens and inguinal lymph nodes (LN) were collected from 8–12 week old OT-II mice. Tissues were processed to create a single cell suspension. CD4 T cell purification was then performed using a CD4+ T cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s protocol. For quality control, flow cytometry analysis was performed to determine purity of the isolated cells. Purified OT-II CD4 T cells were stained with cell trace violet (CTV), then 2.5x10⁶ cells were injected intraperitoneally (IP) into congenic recipient mice, concluding day 0.

For the dose optimization studies, recipient mice (N=5) were treated on days 1, 4, and 7 then euthanized on day 10. Particle treatments were prepared by suspending particle aliquots in PBS. Treatments were all administered subcutaneously (SubQ) as 2 100μL injections in the flanks. At the end of the study timeline, spleens were collected post-euthanasia. Suppressor assays and flow cytometry analysis were performed to assess functional tolerance and T cell phenotypes, respectively.

Suppressor Assays

All cells used in the suppressor assays were obtained from murine spleens post-mortem. A splenocyte suppressor assay and a T cell suppressor assay were both done, defined by which cells were isolated from treated mice (Schematic S1). Where applicable, a CD4+ T cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) was used to isolate T cells. In both the splenocyte and T cell suppressor assays, CD4 T cells were isolated from untreated OT-II mice and stained with carboxyfluoroscein succinimidyl ester (CSFE). For the T cell suppressor assay, to fulfill the role of APCs in the culture, splenocytes were also obtained from a wild type (WT) B6 Thy1.1 mouse and irradiated (rs-2000 x-ray irradiator). In the splenocyte suppressor assay, splenocytes from treated mice and CSFE-stained OT-II CD4 T cells were plated in a 9:1 ratio and incubated with 33.33 μg/mL OVA for 72h. In the T cell suppressor assay, CD4 T cells from treated mice and CSFE-stained OT-II CD4 T cells were plated in a 1:1 ratio and incubated with 33.33 μg/mL OVA for 72h. At the end of the incubation period, flow cytometry analysis was performed, and proliferation of the OT-II T cells was assessed by dilution of CSFE to determine %divided (also referred to as %proliferating).

Flow Cytometry Analysis

Collected tissues were processed to create single cell suspensions. Spleens and LNs were processed according to standard protocols, where tissues were mashed and filtered. For spleens, ACK lysis buffer was used. To process spinal cords, tissue grinders were used to create an initial cell suspension. Suspensions were filtered and a dead cell removal kit (Miltenyi Biotec) was used to remove debris. For panels that included cytokine staining, a 4 hr incubation was performed with PMA and ionomycin (Biolegend, San Diego, CA) to stimulate cells prior to staining. Unstimulated controls were also prepared for each sample. Staining was performed with flow antibodies purchased from Biolegend and used according to manufacturer protocols. Panel details and gating strategies are reported for each analysis (Fig. S4, S6, S7, S9, S13, S17). Acquisition was performed on a BD LSRFortessa flow cytometer (UNC Flow Cytometry Core), and data was analyzed in FlowJo v10.7.1.

Experimental Autoimmune Encephalomyelitis Model Induction and Treatments

EAE was induced in female 12–13 week old murine pathogen free B6 mice (Taconic Farms, Hudson, NY). Mice were pre-acclimated to the facility for 10 days prior to EAE induction. To induce EAE, kits from Hooke Laboratories (EK-2110 kit with MOG_35–55 antigen sequence MEVGWYRSPFSRVVHLYRNGK, Lawrence, MS) were used according to the manufacturer’s protocols. In short, on day 0, mice were weighed then immunized SubQ with an emulsion of complete Freund’s adjuvant and MOG, followed by a 200ng dose of pertussis toxin (PTX) administered IP. PTX was administered again on day 1. Daily assessments of clinical score and weights began on day 7 and continued until the end of the study. Detailed clinical scoring criteria adapted from Hooke Laboratories are summarized in Table S6. Mice were euthanized immediately if they developed any front limb paralysis, developed a body condition of 2 or lower (Table S9), or lost ≥20% of their weight relative to the start of the study. Mouse hydration levels were monitored and small petri dishes with Dietgel Boost (Lab Supply, Durham, NC) were maintained in the cages throughout the EAE studies. Supplementary fluids were administered (1mL PBS IP) to combat dehydration if mice lost ≥ 1 g of weight in a 24 h period or ≥ 0.5 g of weight per day for multiple consecutive days.

Particles were evaluated in two timelines of EAE: an early therapeutic model and a late therapeutic model. In the early therapeutic model, treatments were administered beginning at the onset of disease, when symptoms first emerged in any EAE-induced mice. Prior to administering treatment, a random number generator was used to assign cages to each treatment group (N=5). Particle treatments were each prepared as a suspension in PBS, and all treatments were administered SubQ as 2 100μL injections in the flanks. Mice were treated on days 11, 14, and 17 and monitored until day 20. Spinal cords and cLNs were then collected for flow cytometry analysis. In the late therapeutic model, treatments were administered during peak, severe disease. Prior to administering treatment, mice in the score range of 2.5–3.5 were evenly distributed based on score and, where possible, approximate weights. Only mice in this score range were entered into the study, and treatment groups were randomly assigned (N=8). Mice were treated on days 17, 20, and 23 and monitored until day 45. Spinal cords, cLNs, and spleens from all groups with ≥3 mice remaining at the end of the study were collected for analyses (e.g. flow cytometry).

Antigen Recall Assessment

Splenocytes isolated at the end of the EAE late therapeutic model were plated in 96 well plates at 1x10⁶ cells/well and stimulated with 5μg/mL of MOG for 72 h (media: RPMI 1640 supplemented with 10% FBS, 1% pen/strep, 1% NEAA and 0.1% β-mercaptoethanol). Unstimulated samples were also prepared from each mouse. Supernatant was collected and a multiplex ELISA was performed to measure secretion levels of the following cytokines: IFNγ, IL-17, IL-2, IL-4, and IL-10. The multiplex ELISA was performed according to the manufacturer’s instructions (custom Mcytomag kit, MilliporeSigma, Burlington, MA) and acquisition was performed on a Luminex Magpix. Background correction was applied by subtracting the detected cytokine levels of unstimulated samples from stimulated samples.

Assessment of Immune Competence in Particle Treated Mice

Female 12 week old C57BL6/J mice were obtained from Jackson labs. Mice received influenza vaccines and Ace-DEX particle therapies. Particles were dosed and administered as done in the EAE model. For influenza vaccination, mice were immunized intramuscularly (IM) with 1 μg influenza hemagglutinin (HA) adjuvanted with AddaVax (InvivoGen, San Diego, CA). HA and AddaVax were mixed at 1:1 (v:v) ratio and incubated for 30 minutes according to the manufacturer’s instructions. A full volume of 50 μL was used per immunization (2 x 25 μL injections). The HA used in this study was computationally optimized broadly reactive antigen (COBRA) Y2 kindly provided by the Ross Group at Cleveland Clinic ⁷⁵. Influenza vaccination followed a prime-boost schedule where the second immunization was administered 3 weeks after the first. Particles or a PBS control were administered subcutaneously as done in the EAE model either before or after the influenza vaccination schedule.

Antibody response to the influenza vaccine was assessed in serum. Blood was collected by submandibular bleeding prior to study start, in between influenza immunizations and particle treatments, and at the end of the study. Serum was separated from whole blood by spinning in serum collection tubes (Greiner, Monroe, NC) then stored at −80°C until analysis. Anti-Y2 titers were determined by ELISA as previously described ⁷⁶. The titer cutoff value was calculated by fitting a curve of the background adjusted absorbance value vs the serum dilution using a “log(inhibitor) vs response-variable slope (four parameters)” model in GraphPad Prism 9 and interpolating for a titer cutoff value ⁷⁷. Hemagglutination inhibition (HAI) titers were determined as previously described against A/California/7/09 (H1N1) ⁷¹.

Statistical Analysis

Statistical analysis was performed in GraphPad Prism 7.00. For the late-therapeutic EAE model, a two-way ANOVA with Dunnet’s multiple comparisons post-test was performed to compare all groups to the blank particle group at each timepoint. For the immune competence experiments with influenza vaccinated mice, a one-way ANOVA with Tukey’s multiple comparisons was performed to compare across all groups. In all other cases, a one-way ANOVA with Dunnet’s multiple comparisons was performed to compare all groups to a negative control group. For the early-therapeutic EAE clinical scores and the immune competence experiments, statistical analysis was performed at the final timepoint. Significance is denoted as follows: * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001.

Supplementary Material

Fig. S1 includes IVIS images of subcutaneous depots of Alexa-DEX particles. Fig. S2 shows the draining (inguinal) LNs at multiple time points post Alexa-DEX injection. Fig. S3 includes biodistribution of Alexa-DEX particles to the spleens, hearts, lungs, kidneys, and livers of animals at multiple time points post Alexa-DEX injection. Fig. S4 includes the gating strategy used in the Alexa-DEX trafficking study. Table S1 includes characterization of all Ace-DEX particles. Fig. S5 includes SEM micrographs of particles and peptide release curve. Schematic S1 includes a schematic of suppressor assays. Fig. S6 includes quality control and suppressor assay gating strategy applied to optimization studies in the OT-II adoptive transfer model. Table S2 includes treatment groups and dosing for initial titration in the OT-II adoptive transfer model. Fig. S7 includes the gating strategy, and Fig. S8 includes the results (suppressor assay and flow analysis) from the initial titration in the OT-II adoptive transfer model. Table S3 includes treatment groups and dosing for secondary evaluation of 1% loaded particles in the OT-II adoptive transfer model. Fig. S9 includes the gating strategy, and Fig. S10 includes flow analysis from secondary evaluation of 1% loaded particles in the OT-II adoptive transfer model. Table S4 includes treatment groups and dosing for final dose confirmation in the OT-II adoptive transfer model, and Fig. S11 includes the suppressor assay results. Table S5 includes treatment groups and dosing for the EAE early therapeutic study. Fig. S12 includes clinical scores from individual mice in the early therapeutic study. Table S6 includes EAE scoring criteria. Fig. S13 includes the gating strategy, and Figs. S14 and S15 include flow cytometry analysis (spinal cord and cLN, respectively) from the early therapeutic study. Table S7 includes treatment groups and dosing for the EAE late therapeutic model. Fig. S16 includes the clinical scores from individual mice and weight summary in the late therapeutic study. Fig. S17 includes the gating strategy, and Figs. S18, S19, and S20 include flow cytometry analysis (spinal cord, cLN, and spleen, respectively) from the late therapeutic study. Fig. S21 includes antigen recall of spleens from the late therapeutic study. Fig. S22 includes evaluation of immune competence to influenza vaccination administered after particle therapy. Table S8 includes a comparison of literature for polymeric particles applied to EAE. Table S9 includes body condition scoring.