Missing Wuhan-Strain Imprinting Leaves Children Vulnerable to Omicron BA.3.2.2

The Omicron sublineage BA.3.2.2 has shown an unusual demographic pattern: while most circulating SARS-CoV-2 variants concentrate their transmission among adults, BA.3.2.2 has been disproportionately detected in children. A new study published in Nature Microbiology offers a mechanistic explanation for this age bias, tracing it to the absence of immune imprinting against the ancestral SARS-CoV-2 Wuhan strain. The researchers show that BA.3.2.2 evades the Omicron-specific antibodies that dominate in people without ancestral-strain exposure, while remaining vulnerable to the Wuhan-cross-reactive antibodies that are enriched in strongly imprinted adults, such as recipients of mRNA vaccines.

The team began by analysing global SARS-CoV-2 sequence metadata from the GISAID EpiCoV database, covering records collected between 1 December 2025 and 24 May 2026. During early 2026, BA.3.2.2 showed a marked increase in prevalence. Age-stratified analysis revealed that two contemporaneous lineages, XFG and NB.1.8.1, displayed conventional demographic profiles, with reported sequences increasing across adult age groups and peaking in older adults. BA.3.2.2, by contrast, was significantly enriched in individuals aged 17 years or younger, with the strongest enrichment in children aged 0 to 9 years. This paediatric skew was consistently observed across Europe, North America and Asia, and was reproduced at the individual-country level.

To examine whether this age bias coincided with differences in ancestral-strain exposure, the researchers compiled retrospective age-stratified COVID-19 vaccination coverage from the European Centre for Disease Prevention and Control, the US Centers for Disease Control and Prevention and the Government of Canada. Across the countries examined, children consistently showed substantially lower and later uptake of ancestral-strain vaccines than adults, with the lowest coverage in the youngest age groups. This epidemiological overlap suggested that the paediatric enrichment of BA.3.2.2 might be rooted in differences in the antibody repertoires shaped by prior exposure to the ancestral virus.

To test this directly, the investigators recruited two paediatric cohorts in China with distinct vaccination histories. The first comprised 16 unvaccinated children aged 8 years or younger, who had no exposure to the ancestral SARS-CoV-2 strain but had experienced natural Omicron infections, representing the immune landscape of most young children. The second comprised 16 children aged 8 to 14 years who had received priming with an inactivated Wuhan-strain vaccine, reflecting the typical immune landscape of school-aged children in China, where such primary vaccination has been recommended for children older than 3 years since 2021. Plasma neutralization was measured against a panel of SARS-CoV-2 variant pseudoviruses.

The results were striking. The younger, unvaccinated children showed significantly lower neutralizing titres against the ancestral D614G virus and, overall, stronger neutralization of JN.1 sublineages, notably XFG, confirming their weaker ancestral imprinting. Against BA.3.2.2, however, 13 of the 16 unvaccinated samples fell below the limit of detection, defined as an NT50 of 10. This diminished BA.3.2.2 neutralization was not simply a reflection of globally weaker Omicron responses: the vaccinated children neutralized BA.3.2.2 robustly, with titres exceeding those measured against XFG. The pattern aligned with the observed paediatric enrichment of the variant and raised the possibility that BA.3.2.2 preferentially exploits antibody repertoires shaped by weak ancestral-strain imprinting.

If imprinting strength dictates BA.3.2.2 neutralization, the phenomenon should also appear among adults with varying degrees of imprinting. The researchers therefore examined a cohort of 20 long-term residents of the same geographic region, stratified by vaccination history: 12 adults who had received three doses of the inactivated CoronaVac vaccine only, and 8 adults who had received at least one dose of an mRNA vaccine, BNT162b2 or mRNA-1273. Blood was sampled approximately 46 months after initial vaccination. Fluorescence-activated cell sorting showed that the mRNA-vaccinated group had a significantly higher frequency of Wuhan-cross-reactive memory B cells, and plasma depletion of the Wuhan receptor-binding domain confirmed stronger ancestral cross-reactivity after mRNA vaccination, whereas Omicron-specific antibodies dominated the inactivated-vaccine response.

Pseudovirus neutralization assays highlighted the divergent antibody profiles. The mRNA group’s titres peaked against D614G and declined against subsequent variants, while the inactivated-only group showed a flatter profile with significantly higher titres against recent variants such as XFG. BA.3.2.2 displayed the inverse pattern: it was neutralized significantly more efficiently by plasma from the mRNA-vaccinated group than by plasma from the inactivated-only group. Thus, plasma from both young children and weakly imprinted adults showed selectively reduced BA.3.2.2 neutralization, linking the paediatric phenotype to imprinting-associated differences in humoral immunity.

Single-B-cell V(D)J sequencing of memory B cells reactive to the JN.1 receptor-binding domain, chosen as a common contemporary antigen to capture both Omicron-specific and Wuhan-Omicron cross-reactive responses, revealed the underlying repertoire differences. The strongly imprinted mRNA-vaccinated cohort was heavily enriched for the immunoglobulin heavy-chain variable germline genes IGHV3-53 and IGHV3-66, which typically encode classical public Class 1 antibodies targeting the ACE2-binding site. These antibodies are readily elicited by ancestral-strain exposure but not by Omicron infection, partly because their germline-encoded CDRH1 and CDRH2 motifs can naturally bind the Wuhan receptor-binding domain. Conversely, the inactivated-only repertoire was dominated by IGHV5-51 and IGHV2-5 segments, which typify Omicron-specific Class 1/4 antibodies generated de novo after antigenic shift.

To validate these findings in a broader population, the team analysed 1,003 randomly collected blood samples from Beijing and Tianjin, obtained between February and August 2025. Using clustering based on neutralization titres, they classified vaccinated, high-neutralization samples into highly imprinted, D614G-biased groups and weakly imprinted, Omicron-specific groups. Consistent with the smaller cohorts, BA.3.2.2 showed stronger neutralization escape in both the weakly imprinted and unvaccinated groups than in the highly imprinted group, and unvaccinated individuals neutralized most Omicron variants robustly but showed markedly lower positivity against BA.3.2.2. V(D)J sequencing confirmed that highly imprinted individuals had elevated IGHV3-53/66 usage, largely absent in unvaccinated individuals. The researchers then tested a panel of 87 recombinant monoclonal antibodies: 45 IGHV3-53/66-encoded Class 1 antibodies broadly retained neutralization against BA.3.2.1 and BA.3.2.2, potentially because the variant lacks key Class 1 escape substitutions such as F456L, whereas of the 42 Class 1/4 antibodies dominated by IGHV2-5 and IGHV5-51, only two retained activity against BA.3.2.2.

The study also explored how BA.3.2.2 might evolve further. Using replication-incompetent pseudoviruses carrying additional receptor-binding-domain substitutions, the researchers found that most mutations substantially escaped Class 1 antibodies, especially the 455-456 ‘FLip’ mutation combination. Notably, a Y505H reversion restored neutralization for the majority of Class 1/4 antibodies that parental BA.3.2.2 had completely escaped, identifying Y505H as the key determinant of escape from Omicron-specific repertoires. Fitness assays indicated that F456L, alone or combined with L455F, maintained or slightly enhanced ACE2 binding and entry efficiency relative to parental BA.3.2.2, although L455S heavily impaired receptor engagement. The authors caution that sustained transmission in paediatric populations, which provide an immunological niche with minimal antibody resistance, could catalyse the emergence of secondary variants combining BA.3.2.2’s paediatric evasion features with Class 1 escape mutations capable of breaching adult imprinted immunity. They argue that future paediatric vaccination strategies should weigh variant-matched antigens against the long-term value of establishing a broad antibody baseline that includes ancestral-induced Class 1 neutralizing antibodies, while noting that the observational design and modest cohort sizes in the vaccination-history analysis are limitations acknowledged through the larger validation cohort.

Subject of Research: Immune imprinting and the paediatric enrichment of the SARS-CoV-2 Omicron sublineage BA.3.2.2

Article Title: Lack of ancestral SARS-CoV-2 imprinting promotes Omicron sublineage BA.3.2.2 infection in children

Article References: Niu, X., Xiong, D., He, P., Yu, L., Li, Y., Cai, X., An, R., Wang, Y., Kong, R., Yu, Y., Wang, J., Wang, Q., Huang, J., Zhang, B., Zhu, T., Zhao, T., Huang, J., Jian, F., Shao, F., … Cao, Y. (2026). Lack of ancestral SARS-CoV-2 imprinting promotes Omicron sublineage BA.3.2.2 infection in children. Nature Microbiology. https://doi.org/10.1038/s41564-026-02526-w

Image Credits: AI Generated

DOI: 10.1038/s41564-026-02526-w

Keywords: SARS-CoV-2, Omicron, BA.3.2.2, immune imprinting, children, neutralizing antibodies, IGHV3-53, IGHV2-5, IGHV5-51, pseudovirus neutralization, vaccination, viral evolution