Serum but not mucosal antibody responses are associated with pre-existing SARS-CoV-2 spike cross-reactive CD4⁺ T cells following BNT162b2 vaccination in the elderly

Introduction

Global efforts have been mounted to develop efficient vaccines against coronavirus disease 2019 (COVID-19) (1). As severe COVID-19 mainly affects older individuals, many vaccination campaigns have prioritized the elderly population (2). However, vaccination efficacy is known to be decreased in this age group, particularly for primary vaccination (3). For COVID-19 vaccination - given the distinct homology of certain antigen target regions of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) to human common cold coronaviruses (HCoV) - one possible explanation could be an age-related reduced number of pre-existing cross-reactive CD4⁺ T cells in old individuals (4, 5). To assess the immunogenicity of the COVID-19 vaccine in this particularly vulnerable age group and identify possible relations to pre-existing SARS-CoV-2-specific cross-reactivities, we examined systemic cellular and serological and salivary humoral SARS-CoV-2-specific immunity during the course of COVID-19 vaccination with BNT162b2 mRNA vaccine (Tozinameran™, Comirnaty™) in old and comorbid nursing home residents (n=18; mean age 83±6) and their middle-aged caregivers (n=14; mean age 47±10) at baseline (prior to first vaccination), at day 28 (d28, 7 days after second vaccination) and at day 49 (d49, 28 days after second vaccination). For comparison with naturally acquired immunity, we additionally analysed COVID-19 convalescents (of comparable age to the middle-aged cohort; mean age 36±11) after mild natural SARS-CoV-2 infection at ∼d28 (n=10), ∼d49 (n=16) or ∼d94 (n=11) after symptom onset.

Materials and Methods

Results and Discussion

In this study, we investigated age-related differences in systemic and mucosal immune responses to COVID-19 mRNA vaccine BNT162b2 and compared to COVID-19 convalescents. First, we analysed anti-S1 IgG and anti-S1 IgA antibody levels and S1 neutralization capacity in serum as well as frequencies of peripheral antigen-reactive CD40L⁺ 4-1BB⁺ CD4⁺ T cells after in vitro stimulation with the N-terminal part (S1, covered by peptide mix S-I) and the C-terminal part (S2, peptide mix S-II) of the spike glycoprotein (Fig. 1a-e). In middle-aged donors, BNT162b2 vaccination induced a prompt and homogeneous response of anti-S1 IgG, anti-S1 IgA, S1-specific functional neutralization in serum and spike-reactive CD4⁺ T cells. In comparison, we observed significantly lower anti-S1 IgG and anti-S1 IgA levels and S1 neutralizing capacity in serum at both time points and lower T cell reactivity to S-I and S-II at ∼d49 in COVID-19 convalescents of comparable age after mild infection, which underlines the strong immunogenicity of BNT162b2 vaccine (Fig. 1a-e). However, in old vaccinees, particularly humoral vaccination responses were delayed and more heterogeneously distributed compared to the middle-aged cohort (Fig. 1a-e). For example, at d28, we did not detect anti-S1 IgG and anti-S1 IgA in five (28%) and eight (44%) old donors respectively (Fig. 1a and b; Supplemental Fig. 2) whereas all middle-aged donors exhibited strongly positive anti-S1 IgG levels. At d49, anti-S1 IgG and S1 neutralizing capacity (including one non-responder) were still significantly lower in the old vaccinees (Fig. 1a and c). Regarding the cellular response, all middle-aged and 89% (S-I) and 94% (S-II) respectively of the old donors acquired CD4⁺ T cell reactivity to S-I and S-II at d28. However, S-II-reactive T cell frequencies increased more homogenously and reached a higher level in the middle-aged than in the old cohort. At d49, both age groups reached comparable S-I- and S-II-reactive T cell levels (Fig. 1d-e). Consistently, reduced humoral and cellular vaccination responses in old individuals have been described for vaccines against influenza, yellow fever and tetanus as well as for COVID-19 (12-15).

• Download figure
• Open in new tab

Fig. 1: Serological antibody responses, neutralizing capacity and CD4⁺ T cell reactivity to S-I and S-II.

a and b, anti-S1 serum IgG (a) and IgA (b) OD ratios in the old (red dots) and middle-aged (white) vaccinees at BL, d28 and d49 and in COVID-19 convalescents (blue) at ∼d28 or ∼d49. c, quantification of SARS-CoV-2 receptor (ACE2) binding inhibition in serum in percent in the old and middle-aged vaccinees at BL, d28 and d49 and in COVID-19 convalescents at ∼d28 or ∼d49. d and e, Stimulation Indices of S-I (d) and S-II (e) peptide pool-specific CD40L⁺ 4-1BB⁺ CD4⁺ T cells in the old and middle-aged vaccinees at baseline (BL), d28 and d49 and in COVID-19 convalescents at ∼d28 or ∼d49. f and g, anti-S1 serum IgG OD ratios (f) and inhibition in serum (g) in the old and middle-aged vaccinees at d28 grouped according to S-II-specific CD4⁺ T cell reactivity at BL (Stimulation Index > 1.5). Positivity thresholds: antibody OD ratio (dotted lines) > 1.1; neutralizing capacity (dotted lines) > 30%; CD4⁺ T cells Stimulation Index (dotdash lines) > 1.5 (positive, above background) and (dashed lines) > 3.0 (certainly positive; 6). Grey lines connect follow-up samples. p ≤ 0.05 = *, p ≤ 0.01=**, p ≤ 0.001=***, p ≤ 0.0001=**** according to the Mann-Whitney U test.

We have recently demonstrated that, in contrast to the N-terminal part, the C-terminal part of the spike glycoprotein contains highly conserved domains and triggers CD4⁺ T cell cross-reactivity to SARS-CoV-2 (11). Pre-existing T cell reactivity to S-II at baseline was significantly lower in old individuals than in the middle-aged cohort (Fig. 1e; Supplemental Fig. 3a). Possible explanations for this age-related reduction of cross-reactivity could be cellular senescence or impaired (oronasal) mucosal immunity (16-18). Remarkably though, in these old individuals, high levels of S-II-specific, but not S-I-specific, CD4⁺ T cells at baseline were associated with higher anti-S1 IgG and consistently with elevated S1 neutralizing capacity in serum at d28 (Fig. 1f-g; Supplemental Fig. 3b). In the light of the increased risk of the elderly for severe COVID-19 and the current discussions on their need of booster vaccinations, it is essential to identify and evaluate possible predictors of low vaccination efficiency particularly in this age group. Several studies have supported the notion of beneficial effects of pre-exposure SARS-CoV-2 cross-reactivity (4, 19, 20). However, whether this phenomenon has a direct effect on BNT162b2 immunogenicity in the elderly was unclear. Our findings here show that increased frequencies of pre-existing S-II-specific CD4⁺ T cells were associated with the efficiency of anti-S1 IgG and S1 neutralizing vaccination responses in the elderly. Cross-reactive CD4⁺ memory T cells expand faster upon antigen reactivation post-vaccination to aide B cell activation and class-switch and thus mount a more efficient antibody response.

We additionally investigated the presence of anti-S1 secretory IgA (sIgA) and S1 neutralization capacity in the saliva as a potential correlate of local mucosal protection from SARS-CoV-2 infection. We assume that the presence of S1-specific antibodies and S1 neutralizing capacity in the saliva may contribute to protection against SARS-CoV-2 infection and reduce local replication (21, 22). At d28 following vaccination, anti-S1 sIgA levels increased above their age groups’ maximum pre-vaccination level in all middle-aged (0.45 OD ratio) but only 60% of the old donors (0.31 OD ratio; Fig. 2a). Consistently, an increase in S1 neutralizing activity in saliva was detected in most vaccinees at d28 (Fig. 2b). This response, however, was transient and anti-S1 sIgA and salivary S1 neutralization dropped to pre-vaccination levels in all vaccinees within four weeks (d49) after the second vaccination (Fig. 2a and b). There was no correlation between anti-S1 sIgA or salivary S1 neutralizing capacity and pre-existing cross-reactive CD4⁺ T cells (Fig. 2c and d). Intriguingly, compared to aged-matched middle-aged vaccinees, COVID-19 convalescents exhibited significantly higher anti-S1 sIgA levels and S1 neutralizing capacity in the saliva at ∼d49 after symptom onset (Fig. 2a and b). Unlike salivary S1 neutralizing capacity, anti-S1 sIgA remained significantly increased in convalescents at ∼d94 indicating that anti-S1 sIgA does not correspond to neutralizing activity following infection at later time points. However, we found a correlation between anti-S1 sIgA levels and salivary S1 neutralization in convalescents at ∼d28 and at ∼d49, which was not observed in vaccinees (Supplemental Fig. 4). This suggests that neutralizing capacity in the saliva following vaccination may not only rely on anti-S1 sIgA but possibly anti-S1 IgG, which is consistent with reports on detectable anti-S1 IgG in the saliva of vaccinated individuals (23). In COVID-19 convalescents, anti-S1 sIgA secretion in salivary glands (and salivary S1 neutralizing activity) is likely induced by locally primed B and T cells in nasopharyngeal lymph nodes and/or tonsils (24). This is underlined by an increase in salivary neutralization capacity and sIgA in convalescents from d28 to d49 after infection (Fig. 2a and b) and may indicate generation of tissue-resident plasma cells after mucosal priming. In contrast, the more transient presence of anti-S1 sIgA in the saliva of vaccinated individuals could be the result of transfusion of serum-derived anti-S1 IgA through the endothelium into the oral mucosa (25). Currently, vaccines for intranasal application are in development, which may fill the gap in mucosal immunity observed here (26).

• Download figure
• Open in new tab

Fig. 2: Salivary immune responses.

a, anti-S1 saliva sIgA OD ratios in the old (red dots) and middle-aged (white) vaccinees at BL, d28 and d49 and in COVID-19 convalescents (blue) at ∼d28, ∼d49 or ∼d94. b, quantification of SARS-CoV-2 receptor (ACE2) binding inhibition in saliva in percent in the old and middle-aged vaccinees at BL, d28 and d49 and in COVID-19 convalescents at ∼d28, ∼d49 or ∼d94. c and d, anti-S1 saliva sIgA OD ratios (c) and inhibition in saliva (d) in the old and middle-aged vaccinees at d28 grouped according to S-II-specific CD4⁺ T cell reactivity at BL (Stimulation Index > 1.5). Salivary anti-S1 sIgA and ACE2 binding inhibition are displayed as means of up to three saliva samples on consecutive days (one per day for three days). Grey lines connect follow-up samples. ns = non-significant, p ≤ 0.05 = *, p ≤ 0.01=**, p ≤ 0.001=*** according to the Mann-Whitney U-test.

Taken together, our findings indicate that the presence of anti-S1 sIgA and S1 neutralizing capacity in the saliva after vaccination is of shorter duration and lower magnitude than after natural infection highlighting the need to determine the role of mucosal immunity, e.g., in the form of sIgA in saliva, for evaluation of SARS-CoV-2 immunity and its transmission. Furthermore, we demonstrate that BNT162b2 induces strong immune responses in middle-aged as well as most old and comorbid individuals. However, for some old individuals, the serological response to vaccination is hampered and may leave these individuals at higher risk of infection and severe disease courses, thus promoting recommendations for regular immune status check-ups and further vaccination boosts. Importantly, we show here that pre-existing SARS-CoV-2 spike glycoprotein cross-reactive memory T cells are associated with vaccination efficiency in the elderly and may generally contribute to the high responsiveness to COVID-19 vaccines.