Longitudinal cellular and humoral immune responses after triple BNT162b2 and fourth full-dose mRNA-1273 vaccination in haemodialysis patients

2.1 Study design and sample collection

This is a follow-up study in haemodialysis patients and control individuals, for which the results for haemodialysis patients after a complete two-dose BNT162b2 vaccination (16) and subsequent decline (17) have been previously reported. Blood samples were taken before start of dialysis (n=50) or from healthcare workers (n=33), who participated in the COVID-19 contact (CoCo) study served (18) as non-dialysed control population. To be included in the study, participants had to be over the age of 18 and able to give written informed consent. For the current analysis, we only considered dialysis patients for which results from all time points after either three or four vaccine doses were available. All participants received the standard two-dose regimen of BNT162b2 three weeks apart, followed by a third BNT162b2 vaccination about six (dialysis) or 8·5 months (controls) after the second vaccination. Only dialysis patients were vaccinated a fourth time with 100 µg mRNA-1273 four months after the last BNT162b2 vaccination. The vaccination schedule and blood collection time points are depicted in Fig. 1 and Fig. S1. Participants with SARS-CoV-2 infection diagnosed by either PCR or anti-nucleocapsid IgG determined by MULTICOV-AB multiplex measurement (19) were excluded from the analysis. Demographic characteristics and medical information are listed in Table 1, S1 and S2. Plasma was obtained from lithium heparin blood (S-Monovette Plasma, Sarstedt, Germany). Whole blood samples were used immediately for interferon γ release assay (IGRA). For saliva collection, all individuals spat directly into a collection tube. To inactivate replication-competent SARS-CoV-2 virus particles potentially present in saliva samples, Tri(n-butyl) phosphate (TnBP) and Triton X-100 were added to final concentrations of 0·3% and 1%, respectively (20). Both plasma and saliva samples were frozen and stored at −80°C until further use.

• Download figure
• Open in new tab

Fig. 1. Participant recruitment scheme for longitudinal vaccination response analysis in haemodialysis patients after triple BNT162b2 and fourth full-dose mRNA-1273.

Patients on haemodialysis (n=50) and healthcare workers as controls (n=33) were triple-vaccinated with BNT162b2 (green syringe) followed by a 100 µg (full) dose of mRNA-1273 (blue syringe) for dialysed individuals only. Sampling and vaccination schedule is given in days and weeks as indicated.

2.2 Ethics statement

Ethical approval was obtained from the Internal Review Board of Hannover Medical School (MHH, approval number 8973_BO-K_2020, amendment Dec. 2020). Written informed consent was obtained from all participants prior to starting the study.

2.3 MULTICOV-AB

IgG binding and levels were analysed using MULTICOV-AB, a multiplex coronavirus immunoassay which contains the trimeric Spike B.1, its subdomains (S1, S2, RBD), nucleocapsid B.1 and RBDs of Delta and Omicron BA.1 antigens as previously described (9, 19). Briefly, antigens were immobilised on spectrally distinct populations of MagPlex beads (Cat #MC10XXX-01, Luminex Corporation) either by EDC/s-NHS coupling (21) or by Anteo coupling (Cat #A-LMPAKMM-10, Anteo Tech Reagents) following the manufacturer’s instruction (19). The combined MagPlex beads were then incubated with samples at an effective dilution of 1:3200 for plasma and of 1:12 for saliva. After a wash step to remove unbound antibodies, IgG was detected with R-phycoerythrin labelled goat-anti-human IgG (Jackson ImmunoResearch Labs, Cat #109-116-098, Lot #148837, RRID: AB_2337678) as secondary antibody. After another wash step and bead resuspension, samples were measured once on a FLEXMAP 3D instrument (Luminex Corporation) using the following settings: Timeout 80 sec, Gate: 7500-15000, Reporter Gain: Standard PMT, 50 events. Raw median fluorescence intensity (MFI) values or normalised values (MFI/MFI of quality control (QC) samples (19, 22) are reported. Three QC samples were measured per individual plate to monitor MULTICOV-AB performance.

2.5 Anti-SARS-CoV-2 QuantiVac ELISA

Plasma samples were additionally analysed using the Anti-SARS-CoV-2-QuantiVac-ELISA IgG (Cat #EI 2606-9601-10G, Euroimmun) as previously described (16).

2.6 Interferonγ release assay

SARS-CoV-2-specific T-cell responses from whole blood were analysed by measuring IFNγ production after stimulation with a peptide pool from the SARS-CoV-2 Spike S1 with the SARS-CoV-2 Interferon Gamma Release Assay (Cat #ET-2606-3003, Euroimmun) and the IFNγ ELISA (Cat #EQ-6841-9601, Euroimmun) as previously described (16). Background signals from negative controls were subtracted and final results calculated in mIU/mL using standard curves. IFNγ concentrations >200 mIU/mL were considered as reactive. Results from positive and negative controls were not statistically significant different between time point T1 and T2. We defined this arbitrary cut-off by using average background IFNγ activity without antigen-stimulation in all samples of T1 multiplied with 10 for the threshold for IGRA-positive. Using this cut-off, we found in all of the 15 controls taken from independent individuals before the COVID-19 pandemic negative IGRA results (24). The upper limit of reactivity was 2000 mIU/mL.

2.7 Data analysis and statistics

RStudio (Version 1.2.5001), with R (version 3.6.1) was used for data analysis and figure generation. Additionally, the R add-on package “beeswarm” was utilised to visualise data as stripcharts with overlaying boxplots and to create non-overlaying data points. A second R add-on package “gplots” was used to generate specific colours for plots. Figures were exported from RStudio and then edited using Inkscape (Inkscape 1.2). Spearman’s rho coefficient was calculated to determine correlation between IGRA results and ACE2 binding inhibition using the “cor” function from R’s “stats” library. Mann-Whitney-U test and Wilcoxon test were used to determine difference of signal distributions between dialysed and non-dialysed groups for unpaired and paired samples, respectively using the “wilcox.test” function from R’s “stats” library. To assess differences in the study population, Pearson’s Chi-squared test with Yates’ continuity correction was used for categorical characteristics using the “chisq.test” function from R’s “stats” library and Mann-Whitney-U test as above was used for difference in age. The type of statistical analysis performed (when appropriate) is listed in the figure legends. Pre-processing of data such as matching sample metadata and collecting results from multiple assay platforms was performed in Excel 2016.

2.8 Role of the funders

This work was financially supported by the Initiative and Networking Fund of the Helmholtz Association of German Research Centres (grant number SO-96), the EU Horizon 2020 research and innovation program (grant agreement number 101003480 - CORESMA), the State Ministry of Baden-Württemberg for Economic Affairs, Labour and Tourism (grant numbers FKZ 3-4332.62-NMI-67 and FKZ 3-4332.62-NMI-68) and the European Regional Development Fund (ZW7-8515131 and ZW7-85151373). The funders had no role in study design, data collection, data analysis, interpretation, writing or submission of the manuscript. All authors had complete access to the data and hold responsibility for the decision to submit for publication.

3.1 Inferior humoral responses in haemodialysis patients after triple BNT162b2 vaccination

To characterise the vaccination response after the third BN162b2 vaccination in 50 patients on maintenance haemodialysis, we had followed immunoglobulin levels longitudinally after the second dose of BNT162b2 using MULTICOV-AB, a multiplex immunoassay containing antigens from the Spike protein of SARS-CoV-2 and selected variants of concern (9). As a novel control group, 33 samples from healthcare workers with triple BNT162b2 vaccination were used for comparison. Detailed information on the study populations can be found in Table 1, S1 and S2. Consistent with our previous reports (16, 17), IgG responses towards the original B.1 isolate in vaccinated dialysis patients were significantly reduced (p=4·68*10⁻⁵, Mann-Whitney-U test) when compared to the control group and declined after the second vaccination to comparable levels in both groups (p=7·33*10⁻², Mann-Whitney-U test, Fig. 2a). A third BNT162b2 vaccination about six to eight months after the second increased the peak IgG RBD B.1 response in both groups but with higher variability in dialysis patients (p=4·02*10⁻², Mann-Whitney-U test, Fig. 2a). As an additional control, quantitative S1 IgG titres were measured using a commercial assay (Fig. S2), which led to a very similar pattern of significantly diminished antibody responses in dialysis patients compared to non-dialysed individuals after the second BNT162b2 dose, declining titres and a robust peak response increase after the third vaccination.

• Download figure
• Open in new tab

Fig. 2. Humoral immune response haemodialysis patients after a triple vaccination with BNT162b2.

IgG response (a) and ACE2 binding inhibition (b) towards the SARS-CoV-2 B.1 RBD isolate were measured in plasma from haemodialysis patients (blue circles, n=50) and controls (orange circles, n=33) using MULTICOV-AB (a) or an ACE2-RBD competition assay (b) after double or triple vaccination with BNT162b2 at the indicated time points. Data is displayed as normalised median fluorescence intensity (MFI) signal (a) for IgG binding or as % ACE2 binding inhibition where 100% indicates maximum inhibition and 0% no inhibition (b). Samples with an ACE2 binding inhibition of less than 20% (dashed line) are classified as non-responders. Boxes represent the median, 25th and 75th percentiles, whiskers show the largest and smallest non-outlier values. Outliers were determined by 1·5 times IQR. Mean sampling time in days after two-dose BNT162b2 vaccination as Δt is displayed on the x-axis. Statistical significance was calculated by two-sided Mann-Whitney-U test. P-values for relevant comparisons are given above the sample groups. Significance was defined as p<0·05. Response data from dialysed individuals from day 21 and day 113 after the second BNT162b2 dose were already published before as part of Strengert et al. (16) and Dulovic et al. (17).

For a functional characterisation of vaccine-induced antibodies towards the original B.1 RBD isolate, RBDCoV-ACE2, we used a multiplex competitive inhibition assay (23). ACE2 binding inhibition was significantly reduced in dialysed compared to non-dialysed individuals (p=2·42*10⁻⁶, Mann-Whitney-U test) after the second vaccination (Fig. 2b). Responses were comparably diminished in both groups four to eight months after the second vaccination, with only 12% and 6% of samples being above the 20% responder threshold in patients on haemodialysis and controls, respectively. However, comparable to IgG binding levels, the third BNT162b2 vaccination restored and even augmented ACE2 binding inhibition against the B.1 variant in both populations.

3.2 Strong immune responses after a fourth mRNA-1273 vaccination in haemodialysis patients

Next, we followed the anti-Spike RBD IgG levels in haemodialysis patients after the third vaccination over time and after a fourth vaccination with a full 100 µg dose of mRNA-1273, which was considered by German guidelines for immunocompromised individuals. As expected, IgG responses against the original B.1 isolate had again declined within approximately 4 months after the third vaccination (Fig. 3a, Fig. S2) as did the ACE2 binding inhibition activity as a surrogate for virus neutralisation (Fig. 3b). While the decline was not as severe as after the second BNT162b2 dose with now 64% of samples remaining above the 20% ACE2 binding inhibition threshold, only the fourth vaccination with mRNA-1273 markedly raised both anti-Spike RBD IgG levels (Fig. 3a, Fig. S2, Table S3 for a complete statistical evaluation) and ACE2 binding inhibition (Fig. 3b) towards the B.1 isolate above levels seen at peak response after the second and third dose of BNT162b2. 96% of samples from individuals on haemodialysis were now classified as above the 20% ACE2 responder threshold. Further, we also analysed the longitudinal development of ACE2 binding inhibition towards the dominantly circulating SARS-CoV-2 of 2021 (Delta) and 2022 (Omicron) (Fig 4c, d). ACE2 binding inhibition towards the Delta Variant was slightly reduced over time to levels observed with the B.1 isolate. Overall, the third dose resulted in a clear increase in Delta ACE2 responder rates from 24% after two-dose BNT162b2 scheme to 64%, which was further increased to 94% after the subsequent dose of mRNA-1273 (Fig. 4c). Importantly, neutralisation against the Omicron BA.1 variant, which was largely absent after the second vaccination and only transiently above threshold in 38% of dialysis patients after the third vaccination, reached high levels of ACE2 binding inhibition with an 80% responder rate at peak response after the fourth vaccination with mRNA-1273, which coincided with Omicron being the dominant SARS-CoV-2 variant circulating in Germany (Fig. 4d).

• Download figure
• Open in new tab

Fig. 3. Longitudinal humoral immune response in haemodialysis patients after a triple vaccination with BNT162b2 and a fourth full-dose of mRNA-1273.

IgG response (a) and ACE2 binding inhibition (b, c, d) towards the SARS-CoV-2 RBD of B.1 (a, b), δ (c) and Ο BA.1 (d) isolates were measured in plasma from haemodialysis patients (n=50) using MULTICOV-AB (a) or an ACE2-RBD competition assay (b, c, d) after immunisation with a triple dose of BNT162b2 (green syringe) and a fourth full-dose of mRNA-1273 (blue syringe). Data is displayed as normalised median fluorescence intensity (MFI) signal for IgG binding (a) or as % ACE2 binding inhibition where 100% indicates maximum inhibition and 0% no inhibition (b, c, d). Samples with an ACE2 binding inhibition of less than 20% (dashed line) are classified as non-responders (b, c, d). Interconnecting lines represent samples from the same individual. Sampling time points in days after the standard complete two-dose BNT162b2 vaccination is stated below the graph. Statistical significance was calculated by two-sided paired Wilcoxon rank test. Significance was defined as p<0·05. All p-values for relevant comparisons are listed in Table S3.

• Download figure
• Open in new tab

Fig. 4. Impact of triple vaccination with BNT162b2 and a fourth full-dose of mRNA-1273 on cellular immune response in haemodialysis patients.

(a) Whole blood from longitudinally-sampled vaccinated haemodialysis patients (n=50) was ex vivo stimulated using a SARS-CoV-2 Spike S1-specific peptide pool. Supernatant fractions were then analysed by interferon γ release assay (IGRA). Interconnecting lines represent samples from the same individual. Sampling time points in days after two-dose BNT162b2 vaccination is displayed on the x-axis. Statistical significance was calculated by two-sided paired Wilcoxon rank test. Significance was defined as p<0·05. All p-values for relevant comparisons are listed in Table S3. Response data from dialysed individuals from day 21 after the second BNT162b2 dose were already published before as part of Strengert et al. (16) and Dulovic et al. (17). (b) T-cell responses assessed by IGRA and B-cell responses assessed by ACE2-RBD competition assay towards the RBD of B.1, Delta, Omicron BA.1 isolates were plotted for correlation analysis (b). Correlation was calculated using Spearman’s coefficient rho. Dashed lines indicated the respective responder thresholds for IGRA (IFNγ release >200 mIU/mL) and RBD-ACE2 binding assay (20%). Responder rates (%) for both cellular and humoral response are shown in the upper right quadrant.

We also analysed IgG binding longitudinally after a triple dose of BNT162b2 towards the RBD of B.1, Delta and Omicron BA.1 VoC in saliva of haemodialysis patients to determine protection levels at the primary side of SARS-CoV-2 replication. Although anti-RBD specific IgG was readily detectable both in the peak and plateau response phase following the complete two-dose and the third booster dose of BNT162b2, IgG binding towards the Delta and Omicron BA.1 RBD was significantly reduced compared to the B.1 RBD across all time points (Fig. S3). Interestingly, saliva responses across vaccinated individuals were much more widespread in saliva than in plasma.

As clinical studies suggested that both cellular and humoral response can confer protection from COVID-19 (25), we also assessed vaccination-induced T-cell responses by IFNγ release assay longitudinally. Overall, these responses were more stable over time (Fig. 4a). After two BNT162b2 vaccinations, IFNγ release after in vitro re-stimulation was readily detectable in haemodialysis patients, but declined slightly thereafter. The third BNT162b2 vaccination increased cellular responses to levels comparable to after the second vaccination. Similar to the humoral responses, the fourth vaccination with mRNA-1273 further increased IFNγ release after Spike S1 peptide restimulation of T-cells (Fig. 4a, Table S3 for a complete statistical evaluation).

Finally, we correlated B- and T-cell responses after each vaccination within our longitudinal cohort of haemodialysis patients. We overall observed moderate correlation between peak T-cell responses (measured by IGRA) and B-cell responses [determined by % ACE2 binding inhibition of the B.1 variant (Spearman’s rho=0·561, Fig. 4b, upper panel)], which did not increase after the third (Spearman’s rho=0·405) and fourth (Spearman’s rho=0·371) vaccination. We further described responder rates for T- and B-cell response by a combined cut-off as displayed in Fig. 4b. Notably, responder rates among haemodialysis patients strongly increased to 72% after the triple BNT162b and further to 86% after the fourth full-dose mRNA-1273 dose. Importantly, whilst we observed a similar trend for the correlation coefficient between Delta and Omicron BA.1 % ACE2 binding inhibition and T-cell responses (Fig. 4b; middle and lower panel, Table S3 for a complete statistical evaluation), dual cellular and humoral responders levels equally strongly increased for both VoCs after the third and fourth vaccination to a final 84% and 74%, respectively.