Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mRNA vaccines are highly effective at preventing infection and especially severe disease. However, the emergence of variants of concern (VOCs) and increasing infections in vaccinated individuals have raised questions about the durability of immunity after vaccination.

To study immune memory, we longitudinally profiled antigen-specific antibody, memory B cell, and memory T cell responses in 61 individuals receiving mRNA vaccines from baseline to 6 months postvaccination. A subgroup of 16 individuals had recovered from prior SARS-CoV-2 infection, providing insight into boosting preexisting immunity with mRNA vaccines.

mRNA vaccination induced robust anti-spike, anti–receptor binding domain (RBD), and neutralizing antibodies that remained above prevaccine baseline levels in most individuals at 6 months postvaccination, although antibodies did decline over time. mRNA vaccination also generated spike- and RBD-specific memory B cells, including memory B cells that cross-bound Alpha, Beta, and Delta RBDs, that were capable of rapidly producing functional antibodies after stimulation. Notably, the frequency of SARS-CoV-2–specific memory B cells continued to increase from 3 to 6 months postvaccination. mRNA vaccines also generated a higher frequency of variant cross-binding memory B cells than mild SARS-CoV-2 infection alone, with >50% of RBD-specific memory B cells cross-binding all three VOCs at 6 months. These variant-binding memory B cells were more hypermutated than wild-type–only binding cells. SARS-CoV-2–specific memory CD4⁺ and CD8⁺ T cell responses contracted from peak levels after the second vaccine dose, with relative stabilization of SARS-CoV-2–specific memory CD4⁺ T cells from 3 to 6 months. T follicular helper cell responses after the first vaccine dose correlated with antibodies at 6 months, highlighting a key role for early CD4⁺ T cell responses. Finally, recall responses to mRNA vaccination in individuals with preexisting immunity led to an increase in circulating antibody titers that correlated with preexisting memory B cell frequency. However, there was no substantial increase in the long-term frequency of memory B and T cells. There was also no significant difference in the decay rates of antibodies in SARS-CoV-2–naïve versus –recovered subjects after vaccination, which suggests that the main benefit of recall responses to mRNA vaccination may be a robust but transient increase in circulating antibodies.

These findings demonstrate multicomponent immune memory after SARS-CoV-2 mRNA vaccination, with memory B and T cell responses remaining durable even as antibodies decline. Immune memory was resilient to VOCs and generated an efficient recall response upon antigen reexposure. These durable memory cells may be responsible for continued protection against severe disease in vaccinated individuals, despite a gradual reduction in antibodies. Our data may also inform expectations for the immunological outcomes of booster vaccination.

SARS-CoV-2–specific antibody, memory B, and memory T cell responses were measured at six time points after vaccination, highlighting a coordinated evolution of durable immunological memory. B cell memory was also resilient to VOCs and capable of producing new antibodies upon reactivation. IgG, immunoglobulin G; Ab, antibody; NTD, N-terminal domain; T_FH, T follicular helper cell; WT, wild-type.