Immunity after COVID-19 and vaccination: follow-up study over 1 year among medical personnel

Employees of the Kliniken Südostbayern Hospital Network (Bavaria, Germany) who recovered from a RT-PCR-confirmed COVID-19 episode between April and June 2020 were asked to participate in the prospective cohort study. After written informed consent, directly after recovery, and after approximately 12, 30 and 48 weeks, participants were asked to provide a serum sample (S-Monovette, Sarstedt, Nümbrecht, Germany), and additionally at approximately 30 weeks also a heparin-anticoagulated whole-blood sample for analysis of cellular immunity (LH Monovette, Sarstedt, Nümbrecht, Germany). At each blood sampling date, participants were asked to report their COVID-19-specific symptoms in structured questionnaires.

Approximately 1 year after COVID-19, when vaccines had been approved by heath officials and become available, participants were offered a one dose booster vaccination against COVID-19, according to the recommendations of the German vaccination advisory board (STIKO) [13]. Those who agreed to be booster-vaccinated were asked to provide another serum and whole blood sample at least 14 days thereafter.

Healthy employees without evidence of prior COVID-19 according to symptoms, negative anti-SARS-CoV-2 antibodies and repeated consistently negative SARS-CoV-2 PCR-tests served as controls and underwent the standard two-dose vaccine schedule between February and April 2021 in accordance with STIKO recommendations. They were asked to provide a serum sample immediately prior to the second vaccination and a serum and a whole-blood sample at least 14 days thereafter.

Serum was obtained by centrifugation within 6 h after drawing the blood sample and stored at − 20 °C until analysis. Mononuclear cells (PBMC) were isolated by density gradient centrifugation (Leucosep^®-vials, Greiner Bio-One, Frickenhausen, Germany) und with lymphocyte-separation media (Anprotec, Bruckberg, Germany).

The study was approved by the University of Regensburg ethic committee (reference number 20-1896-101).

Anti-SARS-CoV-2-specific antibody levels in serum were detected by an ELISA utilizing the SARS-CoV-2-spike-protein’s receptor-binding domain (RBD) as antigen, as previously described [14]. The assay is able to detect IgM-, IgA- and IgG-SARS-CoV-2 antibody responses separately with high specificity and sensitivity and the detected antibody levels were shown to correlate well with the virus neutralization capacity of the respective serum sample. ELISA results were expressed as optical densities of the sample/background ratios (signal/cutoff; S/CO), and ELISA readings ≤ 1.0 S/CO were considered negative. 1.0 < S/CO ≤ 3.0, 3.0 < S/CO ≤ 6.0, and S/CO > 6.0 were considered low, intermediate and high antibody titers, respectively. S/CO differences > 0.1 were defined as antibody titer increase or decrease of the antibody level. The high antibody levels in serum samples after vaccination, respectively, booster vaccination were titrated in eight steps of twofold dilutions, starting at a dilution of 1:200.

MultiScreen-IP-plates (0.45 μm, Merck Chemicals, Darmstadt, Germany) were activated with 35% ethanol and coated with 10 µg/ml anti-human IFN-γ (1-D1K, Mabtech, Nacka Strand, Sweden). Prior to addition of PBMCs, the plates were blocked with RPMI-medium and 10% FKS for 1 h at 37 °C, 5% CO₂ in a humidified incubator. Stimulation of the cells was done with 1 µg/ml PepTivator SARS-CoV-2 Prot S (Miltenyi Biotec B.V. & Co. KG, Bergisch Gladbach, Germany). Stimulation with 6.7 µg/ml phytohemagglutinin (PHA-L, PAN Biotech, Aidenbach, Germany) served as positive control. After 16–18-h incubation, 1 µg/ml biotinylated anti-human INF-γ monoclonal antibody (7-B6-1, Mabtech, Nacka Strand, Sweden) was added and incubated for another 2 h at room temperature. This was followed by addition of streptavidin–alkaline-phosphatase-conjugate (Mabtech, Nacka Strand, Sweden) 1:1000 in DPBS and incubation for another hour. Finally, the plates were covered with substrate solution (BCIP/NBTplus, Oxford Immunotec, Abingdon, UK) for 7 min. Analysis of the dried plates was done with an ELISpot Reader System (iSpot Robot and ELISpot Reader 7.0, AID, Straßberg, Germany).

Between April 9, 2020 and June 03, 2020, 136 employees of the hospital network were enrolled in the cohort. These were majorly directly involved in patient care, of younger/middle age (median 38 (inter-quartile range, IQR 29–49) years, 64% female) and generally healthy (15/130 (12%) with a chronic condition according to self-assessment). None of the study participants required inpatient care during the acute COVID-19-illness. 10/130 (8%), 50/130 (38%) and 70/130 (54%) of study participants rated their symptoms during the acute COVID-19 illness as severe, moderate and minor, respectively. Detailed characteristics of the study cohort and the correlation between symptoms, demographic data and antibody levels over the first 8 months have already been reported in detail [3]. During the whole observation period, none of the study participants developed a symptomatic reinfection. The sampling protocol and the number of participants at each time point of blood sampling are shown in Fig. S1.

The course of SARS-CoV-2-specific antibody levels over 1 year after COVID-19 symptom onset showed a decrease with wide interindividual variation. Only 7/136 (5%) of study participants developed no measurable antibody response directly after infection. Nearly 1 year after symptom onset (median 333 (IQR 320–345) days) the proportion of seronegative individuals increased to 21/98 (21%). Extrapolation of the decay of average IgG-anti-SARS-CoV-2 antibody levels allows the estimate that, assuming further linear decay of the antibody levels, nearly 600 days after the onset of symptoms half of the individuals in this study will have antibody levels below the cutoff of the ELISA assay (Fig. 1a).

Anti-SARS-CoV-2-IgA-antibody levels declined more rapidly than IgG-antibodies and were nearly 1 year after symptom onset above the detection limit in only in 19/98 (19%) of study participants (Fig. 1b). IgM-antibody levels were short lived and detectable only briefly after infection (Fig. 1c).

SARS-CoV-2-spike-protein peptide reactive, interferon γ-producing mononuclear cells (mainly memory T cells [15]) were detected by ELISpot at a median of 211 (IQR 202–221) days after COVID-19 infection in 66% of study participants. Counts of these cells did not correlate with the antibody levels specific for the same antigen and measured at the same point in time (Spearman’s rho = 0.18). At this date, approximately 7 months after the infection, only 8/118 (7%) of study participants did not show any measurable antibody-mediated or cellular immunity against SARS-CoV-2.

Between December 31, 2020 and April 29, 2021, 56/136 (41%) of study participants received a booster vaccination, 19 (34%), 17 (30%) and 20 (36%) the vaccines Vaxzevria (Oxford/Astra-Zeneca), Comirnaty (BioNTech) and Spikevax (Moderna), respectively. The 30 control subjects did not differ from the study cohort (median age 41 (IQR 27–50) years, 67% female) and all except two were vaccinated with the Comirnaty vaccine.

The booster vaccination induced among the study participants with previous COVID-19 a very strong increase in anti-SARS-CoV-2-IgG-antibody levels. Determined a median of 42 (IQR 22–59) days after receipt of the booster dose, antibody levels reached significantly higher levels as immediately after infection (Fig. 1a, p < 0.001; Fig. 2a).

Similarly, after booster vaccination 54/56 (96%) of the study participants developed anti-SARS-CoV-2-IgA-antibody, that were significantly higher than directly after the infection (Fig. 1b, p < 0.001). In particular, 36/56 (64%) showed high IgA-levels (Fig. 2c), the COVID-19-naïve controls, however, showed weaker IgA-responses after the standard two-dose vaccination (Fig. 2d). SARS-CoV-2-specific IgM-antibodies were detected in only 3/56 (5%) of study participants after booster vaccination and in none of the vaccinated controls (Fig. 1c).

While booster-vaccinated study participants generally developed high anti-SARS-CoV-2-IgG-titers, mRNA vaccine (Comirnaty and Spikevax) recipients showed even higher antibody titers than Vaxzevria recipients (Fig. 3a; Vaxzevria vs. Comirnaty p = 0.0099; Vaxzevria vs. Spikevax p < 0.001). Antibody titers in COVID-19-naïve controls with complete vaccination did not differ significantly from study participants after booster vaccination (Figs. 1a, 4) and were also significantly above the levels of study participants immediately after COVID-19 infection (p < 0.001).

Elevated SARS-CoV-2-S-protein-reactive memory T cell counts were detectable both in individuals with previous COVID-19 and in COVID-19-naïve vaccine recipients (Fig. 5a). A booster vaccination led in study participants to a significant increase in SARS-CoV-2-S-protein-specific memory T cell counts (Fig. 5c; p = 0.002). SARS-CoV-2-S-protein-reactive memory T cell counts were numerically highest after vaccination with Spikevax and lowest after Vaxzevria, but the differences were not statistically significant (Fig. 3b). SARS-CoV-2-naïve controls after complete vaccination showed numerically, but not statistically lower SARS-CoV-2-S-protein-specific memory T cell counts compared to study participants after infection or after booster vaccination (Fig. 5b).

SARS-CoV-2-S-protein-specific memory T cell counts did not correlate with antibody levels after COVID-19 infection, after booster vaccination of individuals with previous COVID-19 and after complete vaccination in controls.