Both higher- and lower-affinity self-reactive CD4+ T cells are expanded in autoimmunity; however, their individual contribution to disease remains unclear. We addressed this question using peptide-MHCII chimeric antigen receptor (pMHCII-CAR) T cells to specifically deplete peptide-reactive T cells in mice. Integration of improvements in CAR engineering with TCR repertoire analysis was critical for interrogating in vivo the role of TCR affinity in autoimmunity. Our original MOG35-55 pMHCII-CAR, which targeted only higher-affinity TCRs, could prevent the induction of experimental autoimmune encephalomyelitis (EAE). However, pMHCII-CAR enhancements to pMHCII stability, as well as increased survivability via overexpression of a dominant-negative Fas, were required to target lower-affinity MOG-specific T cells and reverse ongoing clinical EAE. Thus, these data suggest a model in which higher-affinity autoreactive T cells are required to provide the "activation energy" for initiating neuroinflammatory injury, but lower-affinity cells are sufficient to maintain ongoing disease.