A SINGLE-CHAIN GMCSF-MOG TOLEROGENIC VACCINE EXPANDS MOG-SPECIFIC CD25+ FOXP3+ REGULATORY T CELLS THROUGH LOW-EFFICIENCY ANTIGEN RECOGNITION EVENTS TO INHIBIT EXPERIMENTAL AUTOIMMUNE ENCEPHALOMYELITIS

Abstract

Previous studies showed that tolerogenic vaccines comprised of single-chain GMCSF-neuroantigen (NAg) fusion proteins inhibited experimental autoimmune encephalomyelitis (EAE) in rodents. The studies detailed here provide evidence that GMCSF-NAg vaccines elicited tolerance through the expansion of preexisting NAg-specific regulatory T cells (Tregs) via low-efficiency antigen recognition that was below the CD40/CD40L activation threshold. GMCSF-NAg-induced tolerance was dependent upon vaccine-induced Tregs, because treatment of mice with a Treg-depleting mAb reversed vaccine-induced tolerance. Vaccine-induced T cell responses were investigated using T cell receptor (TCR) transgenic OTII-FIG mice which recognize OVA323-337 as a high-efficiency antigen, and 2D2-FIG mice which recognize MOG35-55 as a low-efficiency antigen and NFM13-37 as a high-efficiency antigen. Subcutaneous vaccination of 2D2-FIG mice with the low-efficiency GMCSF-MOG vaccine elicited a major Treg population that appeared within 3 days, was sustained over several weeks, expressed canonical Treg markers, and was present systemically in the blood, spleen, and lymph nodes. The GMCSF-MOG vaccine required covalent linkage because a vaccine that contained GM-CSF and MOG35-55 as separate molecules did not elicit Treg responses. GMCSF-MOG vaccination elicited Tregs when introduced either subcutaneously or intravenously as well as in the proinflammatory adjuvants CFA and alum. The GMCSF-MOG-induced Tregs were immunosuppressive and prevented the proliferation of MOG35-55-specific T cells. The GMCSF-MOG vaccine not only elicited Tregs but also induced a desensitized MOG35-55-specific (2D2) T cell repertoire because the vaccine decreased the number of 2D2 CD3+ T cells, reduced the overall expression of the 2D2 TCR, and increased the CD4- T cell compartment.

The ability of GMCSF-NAg vaccines to induce Tregs was dependent upon the efficiency of T cell antigen recognition, because treatment of OTII-FIG and 2D2-FIG mice with the high-efficiency GMCSF-OVA and GMCSF-NFM vaccines respectively, did not elicit Treg responses. The high-efficiency GMCSF-NFM vaccine induced a vigorous T conventional cell (Tcon) memory response and activated the CD40L/CD40 co-stimulatory pathway. In contrast, the low-efficiency GMCSF-MOG vaccine elicited Tregs and lacked sufficient TCR signal strength to activate CD40L/CD40 pathway. Activation of the CD40L/CD40 pathway using an agonistic anti-CD40 mAb precluded Treg expansion with the low-efficiency GMCSF-MOG vaccine in 2D2-FIG mice. Therefore, the strength of the TCR stimulus and the downstream activation or exclusion of the CD40L/CD40 costimulatory pathway was the switch that controlled Tcon versus Treg responses respectively. Remarkably, the low-efficiency GMCSF-MOG vaccine retained Treg expansive activity when co-administered with the high-efficiency GMCSF-NFM vaccine in 2D2-FIG mice.

The GMCSF-MOG vaccine appeared to predominantly drive Treg expansion rather than Treg induction because the emergence of Tregs was delayed in 2D2-FIG-Rag-/- mice which have reduced frequencies of pre-existing Tregs as compared to 2D2-FIG mice. Pre-existing Tregs were also required for tolerance because GMCSF-MOG was encephalitogenic in 2D2-FIG-Rag1-/- mice but not in 2D2-FIG mice. Likewise, GMCSF-MOG was an effective prophylactic in Treg-sufficient C57BL/6 mice and prevented active EAE. Overall, these studies provide evidence that GM-CSF is an effective tolerogenic adjuvant when combined with low-efficiency peptides that fall below the CD40L/CD40 triggering threshold. Thus, a subthreshold CD40L/ CD40 response delimits a critical parameter needed for antigen-specific tolerance and expansion of pre-existing Treg populations.