Biodistribution of self-antigens and regulatory cues in lymph nodes induce memory-like Treg to drive antigen-specific tolerance

Introduction: : Multiple sclerosis (MS) in an autoimmune disease in which the immune system attacks myelin on nerves, leading to loss of motor control. No treatments are curative, and even novel therapies such as monoclonal antibodies cannot distinguish between healthy and disease-causing cells. Alternatively, engineering immune cells—such as regulatory T cells (Treg)— to suppress myelin-attacking cells could lead to long-lasting, vaccine-like control of disease without immunosuppression. Notably, exposing cells to the immunomodulatory drug rapamycin (rapa) in the presence of myelin antigen (MOG) is known to promote myelin-specific Treg. In parallel, several tolerizing approaches involving direct delivery of soluble cues to lymph nodes (LNs) are gaining traction for allergy shots and other immunotherapies. Our studies have coupled intra-LN (iLN) delivery with polymer microparticle (MP) depots encapsulating rapa and MOG. These MPs are too large to drain from LNs, and instead, retain cues in LNs to locally condition immune cells. We have shown that this therapeutic iLN MP treatment induces antigen-specific Treg and reverses paralysis in mice with long term control of disease. However, the mechanisms through which MPs control timing, concentration, and combination of signals to induce long-lasting MOG-specific Treg is unknown. Thus, revealing these mechanisms is crucial toward engineering myelin tolerizing cells with long-lasting functionality. Here, we used spatial mass spectrometry and transgenic mouse models to show MPs extend the location, timing, and concentration of cues in LNs. We test the hypothesis that immune cue composition in MPs plays a critical role in efficacy and promoting markers of long-lasting Treg responses.

Materials and

Methods: : Poly(lactic-co-glycolic acid) PLGA MPs encapsulating MOG peptide and rapa were synthesized using a double emulsion technique. Rapa distribution was characterized using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry to analyze 10μm LN sections 0.01 and 48 hours after iLN MP treatment. Potassiated rapamycin signature (952 m/z) was detected. To test if MPs mediate efficacy by controlling immune cue concentration, mice were therapeutically treated with specific ratios of rapa:MOG 10 days following preclinical disease induction (EAE) and evaluated for clinical score and body weight. To test if the combination of immune cues promotes markers correlated with long-lasting T cell responses, adoptively transferred MOG TCR specific CD4+ T cells were isolated from MP treated LNs after seven days and analyzed via flow cytometry for Treg (CD4+ CD25+ FoxP3+) as well as early markers of persistence including memory cytokine signaling (CD127) and central memory phenotype (CD44, CD62L). Data was analyzed using an ANOVA and Tukey post-hoc test with significance defined as p< 0.05.

Results, Conclusions, and Discussions:: To define how the composition and concentration of cargo in iLN depots impacts efficacy in EAE, mice were therapeutically treated at day 7 with a single injection of MPs containing a fixed quantity of rapa and 10-fold serially diluted quantities of MOG (10, 1, 0.1ug/mgMPs). Notably, MPs containing 1μg/mgMPs MOG (green) significantly reduced disease (i.e., lower clinical scores) compared to 0.1μg/mgMPs (red), demonstrating dose-dependent efficacy (Fig 1A). To investigate the dynamic relationship between dose and controlled release, we used MALDI mass spectrometry imaging to visualize the spatial distribution and concentration of rapa. Mice were treated with rapa-loaded MPs or dose matched soluble control. These studies showed MPs increase the concentration of rapa (yellow) in the LN over two days (Fig 1B). We next tested how MP conditioning of the LN tissue connects to changes in therapeutically-relevant immune cell populations (e.g. MOG-specific Treg). In particular, we tested how the combination of immune signals biases MOG-specific Treg with phenotypes for long-lasting suppressive function. This involved adoptive transfer of MOG-specific T cells into mice treated iLN with MPs. The treated LNs were then analyzed six days later to enumerate the transferred cells. Our analysis revealed co-localization of MOG and rapa was necessary to bias significant expansion of MOG-specific Treg (Fig 1C). Further analysis of these Treg showed increased percentages in LNs with markers of central memory phenotype (CD44+CD62L+) which is connected to persistence and immunological maintenance (Fig. 1D). We also show an increased number of central-memory Treg that express the CD127 receptor- a key mechanism of cytokine signaling for memory cells. Together, these results indicate the composition and concentration of co-encapsulated cues plays a critical role in treatment efficacy and expression of longevity markers in Treg.

Conclusions: Long-lasting therapeutic control of MS without immunosuppression is a critical unmet need. This work helps define rules by which MPs can control the combination and concentration of immune cues in LNs to bias MOG-specific Treg memory phenotypes for long-lasting therapeutic tolerance. Future studies will build off of these results to test the function of these Treg and show MP depots bias similar phenotypes in human cells.

Acknowledgements and/or References (Optional)::