Venous blood-based measurement of SARS-CoV-2 immunity
We used a combined SARS-CoV-2 T cell and IgG antibody assay based on venous whole blood to measure cellular and humoral immune responses in healthy donors (see Table 1 for participant characteristics), giving blood in the COVID-Immun study between September. 2021 and March 202214. Among vaccinated donors, SARS-CoV-2-specific T-cell responses, identified by measurement of plasma-derived interferon-gamma (IFN-γ) after blood stimulation whole with SARS-CoV-2 peptides (as previously described in refs.). 14,15,16,17,18) and nucleocapsid (N)-binding IgG responses were elevated in those who reported prior infection, although both responses were higher in those previously infected unvaccinated donors (Fig. 1a, b). IgG responses directed at the spike glycoprotein (RBD, S1, S2) were all higher in previously infected vaccinated donors (Fig. 1c–e).
Table 1 Characteristics of the participantsFig. 1: Magnitude of anti-SARS-CoV-2 IgG and T-cell responses measured in venous blood samples.
a SARS-CoV-2-specific IFN-γ+ T-cell responses were measured by venous whole blood assay and subdivided based on participants’ vaccination and previous SARS infection status -CoV-2 (PCR and/or lateral flow test confirmed). “Vac + /Inf +” n= 60 (green), “Vac + /Inf-” n= 82 (blue), “Vac-/Inf + ” n= 4 (yellow), “Vac-/Inf-” n= 1 (not represented). SARS-CoV-2-specific IgG binding responses targeting the nucleocapsid (‘N’) (b; ****p< 0.0001, **p= 0.0016), spike receptor binding domain (“RBD”) (c; **p= 0.0022, *p< 0.015), spike subunit 1 ("S1") (d; ***p= 0.0005, *(Vac + /Inf+ vs. Vac + /Inf-) p= 0.022, *(Vac-/Inf+ vs. Vac + /Inf-) p= 0.012) and spike subunit 2 (“S2”) (e) were measured using the venous whole blood assay and subdivided based on participants’ vaccination and previous SARS-CoV-2 infection status (PCR and/or lateral flow test confirmed). “Vac + /Inf +” n= 60 (green), “Vac + /Inf-” n= 71-82 (blue), “Vac-/Inf + ” n= 4 (yellow). Comparisons used Kruskal-Wallis tests with correction for multiple comparisons using Dunn’s tests. Data are presented as boxplots (centerline at median, upper limit at 75th percentile, lower limit at 25th percentile) with whiskers at minimum and maximum values. Each dot represents a donor. The source data is provided as a source data file.
After blood sampling, participants were asked to self-report positive PCR and/or lateral flow test results for COVID-19; Participants were assumed to have contracted the Delta (B.1.617.2) coronavirus variant if they tested positive between September 1, 2021 and December 29, 2021, and Omicron (B.1.1.529) after 29 December 2021, when this variant of concern became dominant according to Public Health Wales. Among the 148 evaluable donors, we observed an infection rate of 26.3% (39/148) in the 6 months after blood draw, 38 of which were breakthrough infections after a second or third dose of the COVID-19 vaccine (pfizer/BioNTech (BNT162b2)). mRNA vaccine or AstraZeneca vaccine (ChAdOx1 nCoV-19); an unvaccinated donor also became infected. The magnitude of the SARS-CoV-2-specific IFN-γ-positive T-cell response was significantly lower in those who reported a positive COVID-19 diagnostic test than in uninfected donors.p< 0.0001; Fig. 2a), mainly due to suboptimal induction of T-cell responses by vaccination among certain participants (p= 0.050; Fig. 1 supplementary). There was no correlation between the magnitude of the IFN-γ+ T-cell response and the time before a positive test result for COVID-19 (Supplementary Figure 2). In contrast, neither RBD-, S1-, S2-binding IgG responses (Fig. 2b–d) nor RBD-, S1-specific neutralizing antibody responses to wild-type SARS-CoV-2 or the delta variant (B.1.617) (Supplementary Fig. 3) could distinguish individuals at risk of infection. However, low anti-SARS-CoV-2 N-binding IgG response was associated with risk of COVID-19 (p= 0.0084; Fig. 2e); in fact, the odds of people testing positive for COVID-19 were 85% lower if they had a confirmed history of SARS-CoV-2 infection (p= 0.00035; OR 0.15, 95% CI 0.047–0.39; Fig. 4 supplementary).
Fig. 2: Magnitude of anti-SARS-CoV-2 adaptive immune responses measured in venous blood samples up to six months before a positive COVID-19 test.
Venous blood samples obtained from healthy donors (n= 148) were assessed for the magnitude of SARS-CoV-2-specific IFN-γ+ T-cell responses (a; ****p< 0.0001) and SARS-CoV-2-specific IgG binding responses targeting the spike receptor binding domain ('RBD') (b), spike subunit 1 (“S1”) (c), spike subunit 2 (“S2”) (d) and nucleocapsid (‘N’) (e; **p= 0.0084). Participants who self-report a positive test for COVID-19 (PCR and/or lateral flow test) are highlighted; all cases of infection occurred within 6 months of blood draw. Comparisons used two-sided Mann-Whitney tests. Data are presented as boxplots (centerline at median, upper limit at 75th percentile, lower limit at 25th percentile) with whiskers at minimum and maximum values. Each dot represents a donor. ns not significant. f Heatmap demonstrating Spearman’s rank correlations between variables in the specified data set. Comparisons that were not statistically significant were excluded from the matrix and are represented by empty boxes. The source data is provided as a source data file.
The previously defined cut-offs for diagnostic positivity14 were considered too arbitrary to assess the risk of reinfection, therefore quartile intervals were established to establish absolute risk parameters. A statistical model that included only variables shown to have a significant impact on outcome revealed that the magnitude of the SARS-CoV-2-specific IFN-γ+ T-cell response was the most significant immunological biomarker for establishing the odds of an individual testing for COVID-19. 19 positive (Fig. 2f and Supplementary Fig. 4). Participants with SARS-CoV-2-specific IFN-γ+ T-cell responses in the third (194–489 pg/ml IFN-γ) and fourth (>489 pg/ml IFN-γ) had a 65 % (p= 0.055; OR 0.35, 95% CI 0.11–1.00) and 90% (p= 0.0050; OR 0.098, 95% CI: 0.014–0.42) smaller odds respectively than those in the first quartile (Supplementary Figure 4). Overall, participants with a SARS-CoV-2 venous blood-derived T cell-specific response ≤79 pg/ml IFN-γ had a 43.2% risk of progressive infection at 6 months, whereas those with a response >489 pg/ml IFN-γ had a 5.4% risk of infection (Table 2).
Table 2 Absolute risk of SARS-CoV-2 infection within 6 months after blood sampling
Capillary blood-based measurement of SARS-CoV-2 immunity
Venous whole blood testing has limited scale due to the need for a phlebotomist to obtain samples. In order to increase the accessibility of the SARS-CoV-2 T-cell and IgG test, an alternative capillary blood sampling technique was developed that allows participants to obtain the blood sample at home with a finger prick To our knowledge, there are no previous reports measuring antigen-specific T cell functionality in capillary blood samples. A strong correlation between lymphocyte counts obtained using matched capillary and venous blood samples has previously been demonstrated19. In addition, whole blood-based assays measuring SARS-CoV-2-specific T cell responses have been reported using as little as 320 μl of venous blood20, thus mitigating concerns about the frequency of T cells precursors within capillary blood samples.
We used this high-throughput, whole capillary blood-based, combined SARS-CoV-2 T cell and IgG antibody assay to measure cellular and humoral immune responses in participants with a heterogeneous range of comorbidities and states prior vaccination/infection ( Table 1 ), recruited across the UK between 24 January and 14 March 202214. The majority (90.9%) of finger prick blood samples were obtained correctly and were sent to the laboratory within 24 hours of collection. In some cases, samples arrived up to 48 h after blood draw, although none of these samples failed quality controls or affected overall T-cell or antibody measurements (Figure supplementary 5). Although some individuals showed variation in the magnitude of SARS-CoV-2-specific IFN-γ+ T-cell responses measured in matched capillary and venous blood samples, in general there was no significant difference (p= 0.88; Fig. supplementary 6).
Among vaccinated individuals who also reported prior infection, SARS-CoV-2-specific IFN-γ+ T-cell responses were significantly elevated (p= 0.0001), although not significantly higher than previously infected unvaccinated donors (p= 0.19; Fig. 3a). IgG responses directed to the spike glycoprotein (RBD, S1, S2) were all significantly higher in vaccinated versus unvaccinated donors, regardless of previous infection status (Fig. 3b–d). Interestingly, mean N-binding IgG responses were higher in previously infected unvaccinated and unvaccinated participants, although this did not reach significance (Fig. 3e). Among self-reported unvaccinated and uninfected donors, 15/37 (40.5%) participants had positive N-binding IgG responses above a pre-defined cutoff14 of 2.0 BAU/ml; 12 of these 15 participants had positive IFN-γ+ T-cell responses above a pre-defined cutoff of 22.7 pg/ml IFN-γ14. As such, it is highly likely that these participants were previously infected with SARS-CoV-2 and were not tested for COVID-19 due to personal choice, lack of availability of PCR devices, and/or or lateral flow, or they were asymptomatic. Although there was a significant correlation between the IFN-γ+ T-cell response and the level of N-binding IgG in unvaccinated donors (p= 0.0044; Fig. Supplementary Fig. 7), N-binding IgG responses declined at a greater rate in vaccinated versus unvaccinated donors, whereas IFN-γ+ T-cell responses were maintained regardless of …