Matrix protein 2 ectodomain (M2e) is considered an attractive component of a broadly protective, universal influenza A vaccine. Here we challenge the canonical view that antibodies against M2e are the prime effectors of protection. Intranasal immunizations of Balb/c mice with CTA1-3M2e-DD-generated M2e-specific memory CD4 T cells that were I-Ad restricted and critically protected against infection, even in the complete absence of antibodies, as observed in JhD mice. Whereas some M2e-tetramer-specific memory CD4 T cells resided in spleen and lymph nodes, the majority were lung-resident Th17 cells, that rapidly expanded upon a viral challenge infection. Indeed, immunized IL-17A-/- mice were significantly less well protected compared with wild-type mice despite exhibiting comparable antibody levels. Similarly, poor protection was also observed in congenic Balb/B (H-2b) mice, which failed to develop M2e-specific CD4 T cells, but exhibited comparable antibody levels. Lung-resident CD69+ CD103low M2e-specific memory CD4 T cells were αβ TCR+ and 50% were Th17 cells that were associated with an early influx of neutrophils after virus challenge. Adoptively transferred M2e memory CD4 T cells were strong helper T cells, which accelerated M2e- but more importantly also hemagglutinin-specific IgG production. Thus, for the first time we demonstrate that M2e-specific memory CD4 T cells are broadly protective.