Observational study: mRNA-1010 recipients had larger increases in memory B cells, with germinal centres persisting to 26 weeks in 5 of 13 people

Background and open questions

Seasonal influenza viruses drift antigenically, so vaccines must be updated each year; prior exposure can also reduce the effectiveness of recall responses against circulating strains through antigenic imprinting. Germinal centres (GCs) promote selection of high-affinity B cell clones and generate plasma cells and memory B cells, and are therefore seen as a target for improving vaccine breadth and durability.

Earlier mouse and COVID-19 studies showed that mRNA vaccines can induce strong GC responses; but influenza vaccines elicit secondary recall responses, and it has been unclear whether mRNA influenza vaccines produce stronger GC responses in humans and how persistent GCs reshape the secreted IgG repertoire. This study samples GCs directly by ultrasound-guided lymph node fine-needle aspiration and combines this with serum IgG proteomics (Ig-seq) to answer the question.

Study design

WU397 is an observational cohort study, non-randomized and unblinded, with no prespecified sample size and no power design for between-group comparison. Across the 2022–2024 Northern Hemisphere influenza seasons it enrolled 75 healthy adults aged 20–50: 38 received mRNA-1010 and 37 received the quadrivalent split vaccine Fluarix, whose HA strains differ slightly. Blood was collected to 26 weeks after vaccination; 13 people in the mRNA-1010 group and 15 in the Fluarix group additionally underwent ultrasound-guided lymph node fine-needle aspiration (FNA).

Ig-seq covered 16 people from the 2022–2023 season (8 each for mRNA-1010 and Fluarix), analysing affinity-purified A/H3 (A/Darwin/9/2021)-specific serum IgG at weeks 0, 4 and 17/26; clonotypes already present at week 0 were recorded as "pre-existing" and those not detected then as "vaccine-induced". There was no protocol-specified primary endpoint, the central readouts being GC persistence, memory B cell increases and changes in serum clonotypes, with between-group comparisons by two-sided Mann–Whitney U tests and similar methods.

Key results

Peripheral antibodies and memory B cells

ELISpot in the 2022–2023 season showed HA-specific plasmablasts peaking 1 week after vaccination; frequencies of A/H1- and A/H3-specific IgG+ and IgA+ plasmablasts were significantly higher in the mRNA-1010 group than in the Fluarix group, while IgM+ cells were higher with Fluarix, with no significant difference for influenza B. Across the 2022–2024 seasons, increases in plasma IgG titres against A/H1 and A/H3 were significantly greater in the mRNA-1010 group at week 4 and were maintained to the final time point; increases in HA-specific memory B cells were significantly greater at week 4 and at weeks 17/26. Increases in HAI titres did not differ significantly between groups at week 4 or at weeks 17/26.

Germinal centre persistence and somatic mutation

FNA sampling showed that, across both influenza seasons, only mRNA-1010 recipients, 5 of 13 (about 38%), still had detectable HA-specific GC responses 26 weeks (about 6 months) after vaccination, with no persistent GCs in the Fluarix group in either season. Three mRNA-1010 recipients contributed 21 clonally related pairs of week-1 plasmablasts and week-26 GC B cells, with the latter showing significantly higher IGHV and IGLV mutation frequencies, although BLI and ELISA showed no significant increase in affinity.

Serum clonotypes and intraclonal expansion

Ig-seq detected 67–186 A/H3-specific serum IgG clonotypes per person; pre-existing clonotypes made up 83.9% in the mRNA-1010 group and 86.2% in the Fluarix group at week 4. The mRNA-1010 group had significantly more vaccine-induced clonotypes at week 4 with comparable numbers of pre-existing clonotypes, and this difference weakened to slightly higher by weeks 17/26. The proportion of pre-existing clonotypes undergoing intraclonal CDRH3 expansion (more unique CDRH3 peptides at week 4 than week 0) averaged roughly 3 times higher in the mRNA-1010 group, and their share of serum was also higher and persisted to weeks 17/26.

Binding breadth and correlation with titres

In the 2022–2023 season cohort (mRNA-1010 n=14, Fluarix n=15), bead-based assays showed significantly greater week-4 increases in plasma binding to 9 A/H1 and 12 A/H3 strains in the mRNA-1010 group (FDR-corrected q<0.05); at week 26 (with some week-17 samples) A/H1 was no longer significant while A/H3 remained so. The proportion of newly arising CDRH3 peptide variants seen by Ig-seq correlated positively with the change in anti-A/H3 titres at week 4 (Spearman ρ=0.694).

Neutralization breadth

In microneutralization assays against homologous (A/Darwin/9/2021) and heterologous (A/Thailand/8/2022) A/H3N2, neutralizing titres rose in both groups, but the increases did not differ significantly between groups. In sequencing-based neutralization assays, the mRNA-1010 group had significantly greater week-4 increases against 11 of 13 A/H1N1 strains; against 83 A/H3N2 strains, increases were higher for most strains but did not reach statistical significance.

Mechanistic interpretation

Demonstrated in the paper: Flow cytometry and single-cell sequencing directly show that GC responses in the mRNA-1010 group can persist to 26 weeks, with both IgG+ and IgA+ cells detectable among HA-specific GC B cells, whereas the Fluarix group was predominantly IgG+. The median proportion of HA-specific GC B cell clones overlapping with week-1 plasmablasts was 87% in the mRNA-1010 group, while in the Fluarix group only 1 person showed overlap (20%).

Lineage tree analysis identified 3 A/H3 serum IgG clonotypes that were clonally related to GC B cells and showed expanded CDRH3 diversity, all from the mRNA-1010 group and none from Fluarix; GC B cell sequences (weeks 8–26) sat in subbranches corresponding to post-vaccination CDRH3 peptide variants. The proportion of newly arising CDRH3 peptide variants correlated positively with the change in titres at week 4 (ρ=0.694), which is correlative evidence.

Author hypotheses: The authors speculate that mRNA-1010 may deliver more antigen to draining lymph nodes and prolong antigen persistence, and may activate low-frequency memory B cells through the self-adjuvant effect of lipid nanoparticles, the higher valency of membrane-bound antigen and human cellular glycosylation patterns; repeated rounds of mutation in persistent GCs may provide "anticipatory mutations" for clonal diversification and binding breadth, and whether repeated vaccination broadens the memory repertoire remains speculative.

Limitations and uncertainties

• The study is observational and non-randomized, with no prespecified sample size, limited to healthy adults aged 20–50, and the two vaccines contain slightly different HA strains, so platform effects and strain differences cannot be separated; all between-group differences should be read as associations.
• The FNA subgroup comprised only 13 people in the mRNA-1010 group and 15 in the Fluarix group, GC persistence was a count comparison of 5/13 versus 0, and the paper reports no statistical test for it; follow-up extends to 26 weeks with no protective efficacy or clinical endpoint data.
• Single-cell and lineage analyses are very small in scale: scRNA-seq covered 4 mRNA-1010 and 3 Fluarix recipients, the 21 clonally related pairs came from 3 people, and only 3 serum clonotypes met the lineage criteria; Ig-seq covered only A/H3 in 8 people per group.
• Functional evidence is limited: there were no significant between-group differences in HAI or A/H3N2 microneutralization, and increases in neutralization across the 83 A/H3N2 strains did not reach significance, which the authors attribute to limited sample size; the cause of persistent GCs (which the authors say remains unexplained) and any causal link to breadth were not directly tested.

Clinical and industry implications

If these associations are reproduced in larger, randomized studies, mRNA influenza vaccines such as mRNA-1010 may broaden the binding repertoire of serum IgG by prolonging GC activity, which aligns with the broad binding sought for a universal influenza vaccine.

Current evidence is limited to immunological readouts: neither HAI nor A/H3N2 microneutralization differed between groups, so clinical protection cannot be inferred from this. FNA combined with Ig-seq offers a methodological framework for comparing the effects of vaccine platforms on GC persistence and the serum IgG repertoire; because the two vaccines contain different HA strains, comparative conclusions should be interpreted with caution.

Authors, source and verification

Evidence level: Full text; verification record: Europe PMC fullTextXML PMC13414572: read the abstract, introduction, all Results subsections, Discussion, Methods and all figure legends