SARS-CoV-2 infection imprints neutralising antibody responses in the absence of vaccination

Immune imprinting refers to a phenomenon where a first viral encounter shapes subsequent immune responses against antigenically distinct variants, favouring the recall of memory B cells against cross‐reactive epitopes in lieu of de novo responses. This bias can lead to suboptimal neutralising antibody protection when subsequent infections occur with highly divergent viral lineages such as the SARS-CoV-2 Omicron lineage viruses. Given the rapid rollout of Wu-1-based vaccines in many countries before the emergence of Omicron, prior studies have demonstrated that vaccination with ancestral Wuhan-Hu-1 (Wu-1)-based vaccines results in diminished neutralisation of the Omicron variants compared to Wu-1, even after recent Omicron exposure. The effects of imprinting induced by natural infection in the absence of vaccination is unknown, however, particularly in populations where Omicron infection preceded Wu-1-based vaccination. This presents a critical gap in our knowledge given the different dominance hierarchies of neutralising antibody responses between vaccines and natural infection. We therefore assessed humoral responses in a population exposed to pre-Omicron (ancestral) and post-Omicron variants prior to receiving any vaccine. Using a binding antibody assay to distinguish variant-specific exposures, we found that 93% of participants had serological evidence of ancestral SARS-CoV-2 infection, while 58% had evidence of Omicron exposure. In individuals exposed solely to pre-Omicron variants, neutralisation titres against Wu-1 were significantly higher than those against Omicron variants, as expected. Paradoxically, participants with evidence of both pre-Omicron and Omicron infections also exhibited higher neutralisation titres for Wu-1 relative to Omicron BA.1, despite the ancestral exposure occurring ~18 months earlier and Omicron being the most recent infection. We confirmed these findings in another independent Nigerian cohort where Omicron exposure was even more prevalent (94%), and all participants had evidence of pre-Omicron infection. The findings suggest imprinted immunity from the ancestral pre-Omicron lineage viruses, and remarkably these early responses to Wu-1 were able to dominate over more recent, likely multiple, Omicron lineage infections. To directly assess imprinting, we performed serum antibody depletion experiments using Wu-1 spike protein as bait. In nearly all participants, depletion of Wu-1-specific antibodies resulted in complete abrogation of serum neutralising activity against both Wu-1 and Omicron spike pseudotyped viruses. Furthermore, during prospective follow-up, additional Omicron infection and contemporaneous administration of Wu-1-based vaccine boosted neutralising responses across variants, partially equalizing titres between Wu-1 and Omicron. However, Omicron responses did not surpass ancestral responses, suggesting only partial mitigation of imprinting. These data demonstrate the presence of immune imprinting against SARS-CoV-2 in the absence of vaccination and its persistence thereafter. The dominance of ancestral responses in unvaccinated populations in the post-Omicron era suggests the preferential use of an Omicron variant-based vaccine in lieu of Wu-1-based vaccines to override imprinting and achieve broader protection, particularly in vulnerable populations such as the elderly or those with compromised immunity. These data also inform preparation against future pandemic zoonoses with rapidly evolving viruses, cautioning against over-reliance on vaccines against ancestral viruses.