Dynamic impact of bivalent COVID-19 vaccine boosters on systemic and mucosal antibody and T cell immunity

Abstract

COVID-19 vaccines were updated to address immune escape from variants of concern (VOC). We explored the impact of ancestral/BA.1 bivalent mRNA booster vaccination (Autumn 2022) on peripheral and nasal antibody and T-cell responses to SARS-CoV-2 in an observational cohort of 133 healthcare workers, building on previous longitudinal vaccination studies. We demonstrate that maintenance of antibody and T-cell responses up to eighteen months following the third vaccine is at least partially driven by intercurrent infection. Boosting with the bivalent vaccine increases the breadth of circulating and nasal antibodies to spike, which waned over time but was still detectable six months post-dose. T-cell responses are well maintained and highly cross-reactive to VOCs irrespective of booster vaccination. Vaccination strongly boosted nasal IgG, but this was short-lived compared to circulating antibodies. Overall, ongoing COVID-19 vaccination provides benefit, boosting immunity in individuals who have not been recently infected, but new strategies may be needed to provide longer-term nasal immunity.

Keywords: Antibodies; COVID-19; Mucosal immunity; SARS-CoV-2; T cells; Vaccine.

Fig. 1

Study overview and experimental design. Blood and nasal epithelial lining fluid were collected from individuals registered in the PITCH cohort study. Volunteers (n = 89) were previously vaccinated with either mRNA or AZD1222 (V1&V2) and a subsequent V3 mRNA dose and received a fourth dose of the Pfizer or Moderna bivalent BA.1 mRNA vaccine (BNT162b2 and mRNA-1273) (V4). A parallel group in the cohort (no V4, n = 44) who did not receive the bivalent vaccine served as the control group to assess the impact of V4. Samples (V4, no V4) were collected at V3 + 6 months (m) (n = 51, 25), V3 + 12 m (n = 27,12), V3 + 18 m (n = 44), V4 + 1 m (n = 78), and V4 + 6 m (n = 63). The study period coincided with the circulation of Omicron subvariants BA.1, BA.2, BA.5, BQ1, XBB.1.5, 1.9 and 1.16. Data for V3 + 1 m and additional data for V3 + 6 m were available from a previous study. Antibody binding was measured with the MSD Serology assays and neutralising antibody responses were assessed by focus reduction neutralisation test (FRNT). T cell responses to SARS-CoV-2 were measured by IFN-γ ELISpot and T cell proliferation. Figure created with BioRender.com.

Fig. 2

Impact of the ancestral/BA.1 bivalent booster dose on antibody and T cell responses to SARS-CoV-2. (a–d) Timecourse of circulating IgG antibodies to SARS-CoV-2 (a) ancestral (Wuhan) and (b) BA.1 spike protein by MSD serology assay for V3 + 1 months (m) (n = 22–272), V3 + 6 m (n = 31–108), V3 + 12 m (n = 30–31), V4 + 1 m (n = 19–20), and V4 + 6 m (n = 19–20). Timecourse of paired IgG antibodies to (c) ancestral and (d) BA.1 spike at V3 + 12 m, V4 + 1 m and V4 + 6 m (n = 19 for each timepoint). (e–h) Timecourse of circulating T cell responses (IFN-γ) to overlapping peptide pools of SARS-CoV-2 (e) ancestral and (f) BA.1 spike by IFN-γ ELISpot assay for V3 + 1 m (n = 125–194), V3 + 6 m (n = 121), V3 + 12 m (n = 29–30), V4 + 1 m (n = 52–57), and V4 + 6 m (n = 49). (i–l) The impact of the bivalent vaccine on SARS-CoV-2-specific circulating IgG antibodies to (i) ancestral and (j) BA.1 spike, and T-cell IFN- γ responses to (k) ancestral and (l) BA.1 spike peptide pools. Data generated from the MSD serology assays are expressed in arbitrary units (AU)/mL. The dotted lines in (a, c, i) represent thresholds for a positive response for SARS-CoV-2 ancestral spike (1120.589 AU/mL), based on the mean concentrations measured in 64 pre-pandemic sera + 3 standard deviations (SD). ELISpot values are expressed as spot-forming units per million (SFU/10⁶) PBMCs. Bars represent the median and interquartile range (IQR). Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and calculated as fold change of the median response for each group in case of unpaired data or expressed as the median of the fold change between individual paired data. Unpaired data was compared using Mann–Whitney (two groups) or Kruskal–Wallis test with Dunn’s multiple comparisons test (three groups). Paired data was compared using Friedman test with Dunn’s multiple comparisons test. The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.

Fig. 3

Broadening of circulating IgG to variant of concern (VOC) spike upon receiving the ancestral/BA.1 bivalent vaccine. (a–f) Timecourse of circulating spike-IgG to SARS-CoV-2 VOC (a) Alpha, (b) Beta, (c) Delta, (d) Omicron BA.2 and (e) BA.5 by MSD assay for V3 + 6 months (m) (n = 10), V3 + 12 m (n = 30), V4 + 1 m (n = 19), and V4 + 6 m (n = 18–19). (f–j) The impact of the bivalent vaccine on circulating IgG responses to SARS-CoV-2 spike VOC (f) Alpha, (g) Beta, (h) Delta, (i) Omicron BA.2 and (j) BA.5 in individuals who did (V4, V4 + 6 m, n = 13–19) or did not (noV4, V3 + 18 m, n = 14) receive the bivalent vaccine. Data were generated from MSD serology assays and are expressed in arbitrary units (AU)/mL. Bars represent the median and interquartile range (IQR). Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and calculated as fold change of the median response for each group. Unpaired data was compared using Mann–Whitney (two groups) or Kruskal–Wallis test with Dunn’s multiple comparisons test (three groups). The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.

Fig. 4

Transient broadening of neutralising SARS-CoV-2 antibodies upon vaccination. (a-e) Timecourse of circulating neutralising antibodies (nAb) to SARS-CoV-2 (a) Victoria and variants of concern (VOC) (b) Omicron BA.1, (c) BA.2, (d) XBB.1.5 and (e) BA.2.86 determined by Focus Reduction Neutralisation Assay (FRNT) at V3 + 1 months (m) (n = 59), V3 + 12 m (n = 20), V4 + 1 m (n = 10), and V4 + 6 m (n = 23). (f–i) Circulating SARS-CoV-2 nAb profiles at V3 + 12 m (f), V3 + 18 m (g), V4 + 1 m (h) and (i) V4 + 6 m. (j–n) The impact of the bivalent vaccine on circulating nAb to (j) Victoria, and Omicron (k) BA.1, (l) BA.2, (m) XBB.1.5 and (n) BA.2.86 in individuals receiving the bivalent vaccine (V4, V4 + 6 m, n = 23) and those who did not (noV4, V3 + 18 m, n = 25). Pie charts represent proportion of individuals with nAbs above positive threshold (> 20). The percentage of focus reduction was calculated and IC50 was determined using the probit program from the SPSS package. Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and calculated as fold change of the median response for each group in case of unpaired data or expressed as the median of the fold change between individual paired data. Unpaired data was compared using Mann–Whitney (two groups) or Kruskal–Wallis test with Dunn’s multiple comparisons test (more than two groups). Paired data was compared using Friedman test with Dunn’s multiple comparisons test. The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.

Fig. 5

Nasal epithelial lining fluid (NELF) IgG to SARS-CoV-2 spike is boosted upon vaccination with IgA levels being maintained. Timecourse of (a, b) IgG and (c, d) IgA antibodies against SARS-CoV-2 ancestral (Wuhan) and BA.1 spike (S) by MSD assay at V3 + 12 months (m) (n = 13–20), V4 + 1 m (n = 13–15), and V4 + 6 m (n = 14–21) in NELF. (e–h) The impact of the bivalent vaccine on nasal antibody responses to ancestral and BA.1 spike for (e, f) IgG and (g, h) IgA antibodies in individuals who received the bivalent booster dose (V4, V4 + 6 m, n = 15–22), and those who did not (noV4, V3 + 18 m, n = 28–34). Timecourse of (i) IgG and (j) IgA responses to nucleocapsid (N) in NELF at V3 + 12 m (n = 20), V4 + 1 m (n = 15) and V4 + 6 m (n = 21). Anti-N IgG and IgA responses in individuals who received the bivalent booster (V4, V4 + 6 m, n = 22) and those who did not (noV4, V3 + 18 m, n = 34) at comparable timepoints. (m, n) Correlations showing the relationship of antibody responses to S and N protein for (m) IgG and (n) IgA in NELF. Data generated from the MSD serology assays are expressed in arbitrary units (AU)/mL. The dotted lines represent thresholds for a positive response for SARS-CoV-2 N (IgG: 5.87 AU/mL, IgA:10.41 AU/mL) and S (IgG: 1.69 AU/mL, IgA: 10.46 AU/mL) respectively, based on the mean concentrations measured in 4 pre-pandemic samples + 3 standard deviations (SD). Bars represent the median and interquartile range (IQR). Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and calculated as fold change of the median response for each group. Groups were compared using Mann–Whitney (two groups) or Kruskal–Wallis test with Dunn’s multiple comparisons test (three groups). Correlation analysis was performed using spearman’s correlation. The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.

Fig. 6

Increased breadth of nasal IgG to SARS-CoV-2 VOC spike. (a–f) Timecourse of nasal fluid IgG responses to SARS-CoV-2 VOC spike for (a) Alpha, (b) Beta, (c) Delta and Omicron (d) BA.2 and (e) BA.5 by MSD assay at V3 + 12 months (m) (n = 20), V4 + 1 m (n = 15), and V4 + 6 m (n = 21). (f–j) Impact of the ancestral/BA.1 booster dose on nasal fluid IgG responses to VOC spike for (f) Alpha, (g) Beta, (h) Delta and Omicron (i) BA.2 and (j) BA.5 in individuals who received the bivalent vaccine (V4, V4 + 6 m, n = 22), and those who did not (noV4, V3 + 18 m, n = 34). Data generated from the MSD serology assays are expressed in arbitrary units (AU)/mL. Bars represent the median and interquartile range (IQR). Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and calculated as fold change of the median response for each group. Groups were compared using Mann–Whitney (two groups) or Kruskal–Wallis test with Dunn’s multiple comparisons test (three groups). The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.

Fig. 7

Cross-reactive T cell responses to SARS-CoV-2 VOC spike. T cell IFN-γ responses to SARS-CoV-2 ancestral (Wuhan) and Omicron BA.1 and BA.2 at (a) V3 + 6 months (m), (b) V3 + 12 m, (c) V4 + 1 m, (d) V4 + 6 m, and (e) V3 + 18 m were measured by IFN-γ ELISpot assay. Fold change of response between ancestral and VOC T cell response for Omicron (f) BA.1 and (g) BA.2 over the same period. The dotted line indicates responses to the ancestral strain. Responses are expressed as spot-forming units per million (SFU/10⁶) PBMCs. Bars represent the median and interquartile range (IQR). Statistical significance is indicated by two-tailed P values < 0.05. Fold change between significantly different timepoints is given below the P values and expressed as the median of the fold change between individual paired data. Groups were compared using Friedman test with Dunn’s multiple comparisons test (for paired data) and Kruskal–Wallis test with Dunn’s multiple comparisons test (for unpaired data). The numbers above the x-axis are medians, the numbers in brackets under the timepoints indicate biological replicates.