The rapid evolution of SARS-CoV-2, combined with heterogenous and waning population immunity, sustains COVID-19 as an ongoing public health concern. Understanding of how repeated antigenic exposures influence immunity to variants is needed to inform future vaccination strategies. This study evaluated neutralizing antibodies in a longitudinal cohort with well-documented COVID-19 vaccination and SARS-CoV-2 infection histories to assess whether repeated and varied antigenic exposures enhance the breadth and durability of humoral immunity. Although small group sizes limited statistical comparisons, our findings provide insights into how cumulative antigenic exposures influence variant immunity and vaccine responses, helping address key knowledge gaps given the rarity of large cohorts with well-defined exposure histories.

Despite the recognized advantages of hybrid immunity in enhancing antibody durability and neutralization breadth [17-19], we found that neither prior Omicron breakthrough infection nor prior bivalent vaccination substantially altered the magnitude or durability of neutralization responses after XBB.1.5 vaccination. This finding suggests that the benefits of hybrid immunity may become saturated after multiple exposures. Although additional vaccine doses reliably increased antibody titers, titers waned, and the ongoing emergence of immune-evasive variants supports continued consideration of updated vaccine formulations for at risk individuals.

Another key finding, consistent with previous reports, was strong evidence of immune imprinting (also called back boosting) [26, 27]. The XBB.1.5 vaccination elicited higher neutralizing antibody titers against previously encountered antigens (ancestral/D614G and early Omicron variants) than against vaccine-matched XBB.1.5. Such immune imprinting, often referred to as Original Antigenic Sin [28], reflects the dominance of memory responses from earlier exposures. This phenomenon was also observed in a pre-Omicron study where a Beta variant vaccine in individuals previously primed with the ancestral vaccine elicited higher neutralization titers against the ancestral variant than against the Beta variant [29].

Additional reports similarly described limited response to novel epitopes in BA.5 following bivalent vaccinations [30, 31], suggesting the dominance of immune memory focused on shared epitopes in previously seen antigens rather than de novo responses to novel Omicron epitopes. A recent report further showed that individuals without prior bivalent vaccination exhibited greater boosting of neutralizing antibodies against XBB.1.5 and JN.1 after XBB.1.5 vaccination than those previously given bivalent vaccines, suggesting a dampening effect of prior bivalent exposure [32]. However, it remains unclear whether memory responses actively inhibit responses to novel epitopes. Memory responses could hinder responses to novel epitopes through mechanisms like accelerated antigen clearance or epitope masking, or they may simply dominate through superior memory recall and higher-affinity responses without directly suppressing primary responses. Nonetheless, a booster vaccination increases neutralizing antibody titers overall, irrespective of whether the boosted antibodies target conserved or novel epitopes. However, eliciting responses to a new variant’s novel epitopes would have the advantage of priming a rapid memory response for any subsequent exposure to that variant.

A related observation was that participants with two XBB exposures (XBB infection and vaccination) had the highest titers against XBB.1.5, suggesting a second exposure broadened responses to novel XBB.1.5 epitopes [33]. Participants with a long interval between their ancestral and XBB.1.5 vaccinations also had relatively high titers, consistent with findings that extended intervals between antigenic exposures enhance responses to novel antigens [34-36]. These observations underscore the importance of updating vaccine to enhance responses to new variants, while preserving the benefits of back boosting.

At six months post XBB.1.5 vaccination, GMTs declined by 36–62% across all variants and exposure groups, reflecting the natural contraction of humoral immunity following acute antigenic exposure that is dominated by short-lived plasma cells. The WTs-Bi-XBBPVI-XBB group exhibited the largest declines (46–82%), particularly against JN.1 variants, suggesting that homologous boosting, while generating strong peak responses, may not establish long-lived memory B cells capable of sustained high-level neutralizing antibodies.

Imprinting remained prominent at six months. While titers against the ancestral strain (D614G) remained relatively high, the lowest titers at both one and six months were consistently against the JN.1 sublineage variants, confirming their enhanced immune evasions. A recent study reported waning patterns after XBB.1.5 boosting, but breadth against some of the newer variants was not assessed [26]. The rapid decline of cross-neutralization against highly evasive JN.1 variants highlights the difficulty of sustaining broad protection.

Consistent reduction in neutralization across JN.1 lineage variants harboring mutations that alter glycosylation in the NTD led us to investigate these glycan modifications as potential immune evasion mechanisms. The DelS31 mutation in KP.3.1.1, which has been observed in other sarbecoviruses [37], introduces a glycan at N30 [25], while the T22N mutation creates a new glycosylation site [38]. Sera showed slightly lower neutralization against these NTD-modified variants, suggesting that XBB.1.5 boosting in this cohort primarily induced neutralizing antibodies that target epitopes outside the NTD.

Evolution from BA.2.86 to JN.1 and its JN.1 sublineages reflects progressive acquisition of recurrent mutations, including R346T, L455S, F456L, and Q493E that collectively enhance viral fitness and antibody escape, with L455S playing a major role [33]. The recurrent ancestral residues, such as the N30 glycan and Q493E mutation, also found in other sarbecoviruses, suggests evolutionary optimization balancing immune escape with viral fitness [37].

Interestingly, XBB.1.5 vaccination elicited neutralization against BA.3.2 at levels comparable to XBB.1.5 and JN.1, with similar durability to early Omicron variants. Although BA.3.2 currently appears to lack replication fitness relative to some other variants [39], its 44 unique spike mutations compared with LP.8.1.1 that is included in recently updated vaccines (Fig. S1B), raise concern for potential antigenic shift. However, it still shares approximately 19 mutations with XBB.1.5 and JN.1 sublineages (Fig. S1B). Recent studies show that KP.2-formulated COVID-19 vaccination elicits comparable neutralization of BA.3.2 and JN.1 [40], while others report limited BA.3.2 neutralization after BA.5 and JN.1 exposure [39]. The higher titers observed in our cohort may reflect the participants’ extensive and distinct antigenic exposure histories, assay differences, or both.

Overall, the rapid decline in cross-neutralizing antibodies and persistent immune imprinting reinforce the challenge of maintaining durable immunity with current monovalent mRNA vaccines and underscore the potential benefit of updating vaccine formulation for eligible persons. The emergence of variants with highly balanced immune evasive and fitness profiles emphasizes the need for continuous, antigenic surveillance and real-time data to guide these updates. The marked reduction in titers against highly evasive variants supports considerations for timely boosting to maintain protective antibody levels, especially for high-risk individuals. Long-term success will likely require new vaccination strategies, such as novel antigenic designs or alternative delivery platforms, to improve neutralization breadth and durability, as well as enhance mucosal immunity.

This study has several limitations, including small subgroup sizes, uncontrolled timing intervals between antigenic exposures, lack of pre-XBB.1.5 vaccination sera, and assessment limited to humoral responses. Our pseudovirus assay is less sensitive to NTD-focused antibodies than those targeting the receptor binding site. Finally, the six months follow-up period precludes conclusions about long-term memory B-cell durability.

In summary, in individuals with extensive prior COVID-19 antigenic exposures, monovalent mRNA XBB.1.5 vaccination generated broadly neutralizing antibodies heavily influenced by immunological imprinting, with substantial waning within six months. Lower peak titers against the most antigenically divergent JN.1 sublineages resulted in very low neutralizing titers at six months, suggesting limited residual antibody protection. Additional antigenic exposures added minimal durability or breadth against immune-evasive variants. The marked decline in neutralization against newer variants over typical booster intervals underscores the need for ongoing antigenic surveillance and potentially strategic refinement of booster timing and vaccine composition to maintain protection against evolving variants, especially for populations at highest risk.