The longitudinal kinetics of antibodies in COVID-19 recovered patients over 14 months

Cohort establishment and antibody measurement

To investigate the longitudinal kinetics and persistence of SARS-CoV-2 specific antibodies, we established a cohort of 200 COVID-19 patients who recovered from SARS-CoV-2 infection. The cohort included individuals that were positive to SARS-CoV-2 as determined by qPCR and were admitted to the HaSharon hospital, Rabin medical center, Israel. Days following onset of symptoms (DFS) was derived from self-reported symptom onset. COVID-19 recovered patients enrolled cohort consists of: median age of 53 years (range: 20 to 81 years), gender distribution of males and females was balanced, and disease severity included mild (83%) and moderate/severe (17%) cases. Blood samples were collected from all participants at visit 1 (V1, mean 90 DFS) with follow up collections at visit 2 (V2) from 55.5% participants (mean 184 DFS), visit 3 (V3) from 32.2% participants (mean 298 DFS) and visit 4 (V4) from 12.2% participants (mean 406 DFS) (Figure 1a-b; Table S1).

• Download figure
• Open in new tab

Fig. 1 Experimental design, cohort characteristics, and reproducibility of antibody level measurements.

a) 200 COVID-19 recovered patients were recruited to the study at V1 with 89 patients donating blood samples at subsequent visits (V2-V4) at intervals of approximately 3 months between visits. A subset of the recruited patients, received one dose of mRNA-vaccine at various time points DFS. Blood samples were fractionated for the isolation of plasma samples that were used to determine the RBD⁺ antibody levels and for the analysis of the longitudinal kinetics. b) The cohort of COVID-19 recovered patients included a balanced male/female population and age distribution below and above 50 years. The majority of the recovered COVID-19 patients exhibited mild symptoms. c) RBD⁺ measurements were carried out in duplicates and each experiment was repeated twice (designated Exp1 and Exp2). Spearman’s rank correlation (RS) was used to determine the reproducibility of the experiments. p values < 0.05 were considered significant and Spearman’s rank coefficient is indicated (R_S)

The sub-cohort of COVID-19 recovered patients (24%) were vaccinated with the BNT162b2 COVID-19 mRNA vaccine within a time window of approximately 222 DFS.

Additionally, we included a prospective cohort of 17 naïve individuals that received two doses of the BNT162b2 mRNA-vaccine. Samples from this cohort were collected at four visits following mRNA-vaccine 2^nd dose as follows (median, days following vaccination -DFV): 8, 35, 91, and 182 DFV.

RBD⁺ IgG, IgM and IgA levels were measured by semi-quantitative ELISA and antibody levels were normalized by calculating the ratio of the signal over the mean signal obtained in the negative controls used in each microwell plate (signal over background).

Each sample was measured in duplicates and each experiment was repeated independently twice showing high reproducibility (Figure 1c). In total, 200 samples were collected for V1 with another 145 subsequent samples for V2-V4 to determine the kinetics of RBD⁺ antibodies within 14 months following symptoms.

Antibody levels in association with age, gender, and disease severity

To test whether the robustness of the antibody response following SARS-CoV-2 infection is associated with patients’ age, gender and the severity of the COVID-19 disease, RBD⁺ antibody levels at V1 were stratified according to the above-mentioned parameters. We found that the IgG and IgA levels significantly differ between male and females and IgM was not significantly different between gender groups (Figure 2a).

• Download figure
• Open in new tab

Fig. 2 SARS-CoV-2 RBD⁺ antibody levels over the period of 14 months.

a) grouped scatter plots of RBD⁺ IgG/A/M levels in COVID-19 recovered patients at V1, by gender, age and COVID-19 severity. Y-axis - log₁₀ values of the ratio between the signal obtained over the mean signal for the negative controls in each microwell plate. Mean values are indicated by solid black lines. P values were determined with an unpaired, two-sided Mann-Whitney U-test; p < 0.05 was considered statistically significant b) scatter plot of RBD⁺ antibody kinetics. X-axis represents days following symptoms (DFS). A non-linear regression model was used to determine the regression coefficient. Negative regression coefficients (i.e., slope) were determined for IgG, IgA, and IgM, indicating that all isotypes are decaying over the period of 14 months c) bar plot comparing the regression coefficient obtained for IgG, IgA and IgM. For all plots, p values, R² and slopes are indicated. p values < 0.05 were considered significant.

The cohorts were also divided by age groups with 50 years as the differentiating pivot. Here we found the IgG and IgA levels were significantly elevated in the >50 group compared to <50 groups while IgM did not significantly differ between age groups. The severity during the active disease phase was found to be associated with higher IgG and IgA levels at V1 while IgM was not found to be associated with severity (Figure 2a).

Antibody levels decay over the period of 14 months

All sera samples from COVID-19 recovered patients that were collected during the study (n=345) were used to evaluate the persistence of the RBD⁺ antibodies over the period of 14 months. We examined the association between DFS and the RBD⁺ IgG/A/M levels by applying a regression model and the extent of change over time was determined by the regression coefficient (i.e., the slope of the regression). For all antibody isotypes, the RBD⁺ levels were found to decay over the period of 14 months (Figure 2b). The negative values of the regression coefficient indicated that all isotypes exhibit a decline trend, and the regression coefficients differ significantly between isotypes, where IgA showed significantly faster decay followed by IgM and IgG (Figure 2c). As a result of the antibody decay, 18.4%, 61.81%, and 54.86% of serum samples (IgG, IgA, and IgM respectively) were found to reach the antibody range determined for negative control samples.

The temporal decay of antibodies is associated with the robustness of the early antibody response

The longitudinal kinetics of RBD⁺ antibodies at the individual level was determined by measuring the RBD⁺ antibody levels in follow up samples from 89 recovered COVID-19 patients. We found a marked decay in all follow up patients. The fraction of patients that exhibited antibody levels at V1 within the range of the negative control was 14%, 49% and 40% for IgG, IgA, and IgM respectively. The fraction of patients that exhibited antibody levels at their last visit within the range of the negative control was 29%, 70% and 73% for IgG, IgA, and IgM respectively (Figure 3a). Next, we stratified the RBD⁺ antibody levels to quartiles by bootstrapping the obtained values from serum samples of the extended cohort (n=345), and quartile thresholds were set according to the mean values derived from the bootstrap iterations (Figure S1). The quartile threshold values were then used to designate each measured antibody level at V1 to its respective quartile range (i.e., from low to high Q1-Q4 respectively, Figure S2). Follow up samples from the corresponding patients were used to determine the regression coefficient for each quartile group. For example, patients that exhibited RBD⁺ antibody levels at V1 above the Q4 threshold, were used to determine the regression coefficient based on these patients’ follow up samples obtained in the subsequent visits (Figure 3b). We found that for all isotypes, patients that were designated to quartile Q4, Q3 and Q2, exhibit negative regression coefficient values indicating that the antibody levels are declining over time. The antibody levels for patients that were designated to quartile Q1, did not show a significant decrease in antibody levels over time. Additionally, the rate of decay as determined by the regression coefficient, suggests that the decay rate is significantly faster for those patients that the antibody levels at V1 was designated to Q4 (Figure 3c). This decay was further observed when we examined the regression coefficient for groups Q3 and Q2. For group Q1 there was no significant decay in antibody levels for all isotypes.

• Download figure
• Open in new tab

Fig. 3 Longitudinal kinetics of RBD⁺ antibodies in COVID-19 recovered patients.

a) line graph of RBD⁺ antibody levels in COVID-19 recovered patients with follow up samples obtained at 2-4 visits. Y-axis in log₁₀ and X-Axis, days following symptoms (DFS). Dashed red lines represent the maximum and minimum range of the antibody levels as measured in negative control samples (n=27). b) scatter plots of RBD⁺ antibody levels stratified by quartiles. The antibody levels at V1 determined the attributed quartile and follow up sample from the same patients were used to determine the regression coefficients (slope). Four quartiles for each isotype were analyzed. Quartiles Q4-Q1 thresholds are indicated by horizontal dashed black lines. Y-axis, log₁₀, X-axis days following symptoms (DFS). Dashed red lines indicate the upper and lower range of negative controls.

Antibodies in COVID-19 recovered patients decay slower than in naïve vaccinees

We were interested to examine if the longitudinal kinetics of antibodies following recovery differs from the kinetics following vaccination. To this end, we measured the antibody levels in samples collected at four visits from 17 naïve individuals who received 2 doses of mRNA-vaccine starting on an average of 8 DFV and follow up samples at an average of 35, 91 and 182 DFV. Based on the quartile stratification approach, the RBD⁺ IgG levels at the average of 8 DFV, in all 17 naïve individuals, was designated to the highest quartile. This was not the case for IgA and IgM that were distributed across quartiles. Furthermore, the IgG/A/M levels in all 17 individuals reached the negative control range at the approximate 180 DFV (Figure 4a). The rate of the antibody decay as determined by the regression coefficients reveal that, as with COVID-19 recovered patients, IgA exhibited the fastest decay followed by IgG and IgM (Figure 4b). As the antibody levels at V1 in naïve vaccinees was designated to Q4 we compared the regression coefficient for IgG as measured in the recovered patient group that was designated to Q4 at V1. We found that the decay rate in COVID-19 recovered patients was significantly slower compared to the decay in naïve vaccinees (Figure 4c-d). Moreover, 100% of the naïve vaccinees exhibited antibody levels lower than the upper limit of the control group. On the other hand, only 5% of the recovered patients reached the negative control range at the 182 DFV time point of naïve vaccinees and 19% at an average of 265 DFS in their V4 samples. Next, we examined the effect of vaccination in a subset of the recovered cohort (n=20). As recovered patients received a vaccine at various DFS, we extrapolated the antibody levels to the day that the patients received the vaccine using the regression formula. The antibody levels at 8 DFV were extrapolated as well by using the regression formulas obtained from naïve vaccinees (Figure 4e). Based on the calculated temporal kinetics we found that approximately 80% of recovered vaccinees experienced an increase in IgG levels reaching the highest quartile (Q4) as opposed to 50% of them being designated to Q4 following recovery (Figure 4f). In addition, we found that the predicted IgG levels at 8 DFV was significantly higher than the predicted IgG levels on the day of vaccination (due to continued decay following recovery) and not significantly higher than the antibody levels at V1 (Figure 4g). Still, the increase in IgG levels 8 DFV in recovered patients was not significantly higher than the antibody levels detected 8 DFV in naïve vaccinees who received 2 doses of the mRNA-vaccine (Figure 4g).

• Download figure
• Open in new tab

Fig. 4 RBD⁺ antibody kinetics differ between recovered patients and naïve vaccinees.

a) RBD⁺ IgG/A/M kinetics of naïve vaccinees following 2 vaccine doses. Nonlinear regression model was applied, and the regression coefficient was determined. The 95% CI is plotted in grey. Y-axis log₁₀ values. X-axis represents days following vaccination (DFV). Dashed red lines indicate the upper and lower limit of negative control samples. Horizonal dashed black lines indicate the quartiles. b) Bar plot comparing the regression coefficient obtained for IgG, IgA and IgM. For all plots, p values, R² and slopes are indicated. p values < 0.05 were considered significant. c) RBD⁺ IgG/A/M kinetics in naïve vaccinees in comparison with the kinetics in recovered patients whose IgG levels were designated as Q4 at the first visit (V1). Y-axis log₁₀. X-axis represents days following symptoms for recovered patients (red) and days following vaccination for naïve vaccinees (blue). d) Multiple comparison between the regression coefficients in recovered patients and naïve vaccinees. Y-axis represents the absolute regression coefficient value. e) Utilizing each patients’ regression formula pre-vaccination and the known vaccination day, the RBD⁺ IgG level was extrapolated to predict the IgG level on the day of vaccination (i.e., day 0). Post-vaccination antibody levels were used to extrapolate the antibody levels on day 8 following vaccination. Y-axis represents the RBD⁺ antibody levels in log₁₀. X-axis represents days in relation to the day of vaccination (i.e., day 0) at each serum collection time point and expected antibody levels. Dashed red lines indicate the upper and lower limit of negative control samples. f) Distribution of the recovered-vaccinated patients among quartiles as measured for their antibody levels at V1 compared to the extrapolated antibody levels at day 8 post-vaccination. Y-axis represents the percentage of recovered patients that were designated to each quartile. X-axis represents quartile groups. g) Comparison between the RBD⁺ IgG levels in recovered patients at V1, the extrapolated RBD⁺ IgG levels at day 0 (day of vaccine), day 8 following vaccination, and the RBD⁺ IgG levels of naïve vaccinees at an average of 8 DFV. Y-axis in log₁₀. X-axis represents the relative days from vaccination of each sub-cohort. Dashed red lines indicate the upper and lower limit of the antibody levels in the negative control group.

Human subjects

All patients and naïve individuals provided informed consent to the use of their data and clinical samples for the purposes of the study and blood collection was performed under institutional review board approvals number 0001281-4 and 0000406-1. Blood samples were collected at the Hasharon Hospital, Rabin Medical Center under ethical approval number 0265-20. A total of 345 sera samples from 200 recovered patients, all confirmed qPCR positive for SARS-CoV-2, were collected, with follow up samples from 89 COVID-19 recovered patients at approximately 90 days intervals at 3-4 visits (V1-V4). Additionally, 20 of the follow-up COVID-19 recovered patients received a single dose of BNT162b2 vaccine. Samples from 17 naïve individuals that received two doses of the BNT162b2 mRNA-vaccine were collected at four visits following mRNA-vaccine 2^nd dose as follows (median, days following vaccination - DFV): 8, 35, 91, and 182 DFV. Negative control included sera from 27 healthy individuals that were collected prior to the COVID-19 pandemic.

All blood samples were collected into BD vacutainer K2-EDTA collection tubes. Isolation of serum and peripheral blood mononuclear cells (PBMCs) was performed by density gradient centrifugation, using Uni-SepMAXI⁺ lymphocyte separation tubes (Novamed) according to the manufacturer’s protocol.

COVID-19 recovered participants were classified into levels of disease severity according to the classifying parameters given by the Israeli Ministry of Health (https://www.gov.il/en/departments/general/corona-confirmed-cases). Accordingly, patients that were confirmed for SARS-CoV-2 and exhibit fever, cough, malaise, and loss of taste and smell were considered as “mild”. Patients that were confirmed for SARS-CoV-2 and had pneumonia were considered as “moderate”. Patients that were confirmed for SARS-CoV-2 and had either respiratory frequency of more than 30 breaths per minute, or 93% saturation or less without supplements or problems in oxygenation was considered “severe”. Any case that showed more severe parameters, such as the requirement of ventilation and/or suffers from multiple organ dysfunction, will also be in the “severe” category for our study. COVID-19 recovered participants were stratified into two groups: one consists of all “mild” cases and another of the “moderate” and “severe” cases.

Expression and purification of recombinant protein

The plasmids for expression of recombinant SARS-CoV-2 receptor-binding domain (RBD) was kindly provided by Dr. Florian Krammer, Department of Microbiology, Icahn School of Medicine at Mount Sinai, New York, NY, USA. The RBD sequence is based on the genomic sequence of the first virus isolate, Wuhan-Hu-1, which was released on January 10th, 2020, ⁹. The plasmids for expression of recombinant human ACE2 (hACE2) was kindly provided by Dr. Ronit Rosenfeld from the Israel Institute for Biological Research (IIBR). The cloned region encodes amino acids 1-740 of hACE2 followed by 8xHis tag and a Strep Tag at the 3’ end, cloned in a pCDNA3.1 backbone. Recombinant RBD and hACE2 were produced in Expi293F cells (Thermo Fisher Scientific) by transfections of these cells with purified mammalian expression vector using an ExpiFectamine 293 Transfection Kit (Thermo Fisher Scientific), according to the manufacturer’s protocol, and as described previously ⁹. Supernatants from transfected cells were purified on HisTrap affinity column (GE Healthcare) using 2-step elution protocol with 5CV of elution buffer supplemented with 50mM imidazole in PBS, pH 7.4 followed by 250mM imidazole in PBS, pH 7.4. Elution fractions containing clean recombinant proteins were merged and dialyzed using Amicon Ultra (Mercury) cutoff 10K against PBS (pH 7.4). Dialysis products were analyzed by 12% SDS–PAGE for purity and concentration was measured using Take-5 (BioTek Instruments).

Purified recombinant proteins were biotinylated using the EZ-Link Micro-NHS-PEG4-Biotinylation kit (Thermo Scientific), according to the manufacturer’s protocol.

Serum anti-SARS-CoV-2 RBD antibodies assay

RBD⁺ Ig levels in sera were determined using 96 well ELISA plates that were coated overnight at 4°C with 2μg/ml RBD in PBS (pH 7.4). Next, coating solution was discarded, and ELISA plates were blocked with 300µl of 3% w/v skim milk in PBS for 1 hour at 37°C. Following the discarding of the blocking solution, duplicates of serum diluted 1:300 in 3% w/v skim milk in PBS were added to the microwells. Negative control serum samples were also added in single wells of 1:300 dilutions, with a range of 8-24 control in each plate. ELISA plates were then washed three times with PBST and 50μl of horseradish peroxidase (HRP) conjugated goat anti-human IgG (Jackson ImmunoResearch, #109035003) / anti-human IgM (Jackson ImmunoResearch, #109035129) / anti-human IgA (Jackson ImmunoResearch, #109035011) secondary antibodies were added to each plate at the detection phase (50μl, 1:5000 ratio in 3% w/v skim milk in PBS) and incubated for 1 hour at room temperature (RT), followed by three washing cycles with 0.05% PBST. Development was carried out by adding 50µl of TMB for 5 minutes of exposure, immediately followed with reaction quenching by adding 0.1M sulfuric acid. Plates were read using the Epoch Microplate Spectrophotometer ELISA plate reader using wavelengths of 450 and 620 nm. Reads from the 620 nm wavelength were reduced from those received by the 450 nm wavelength, resulting in a final signal value for each well, and therefore 2 signal duplicates for each tested serum and one for each negative control serum. A total “Negative control value” for each plate was computed as the average of all negative control signals, in addition to one standard division value. A “signal over negative” (S/N) value for each well was then produced. All sera were evaluated in a second independent technical repeat, resulting in 4 S/N values for each serum, that were averaged to a single RBD⁺ Ig titer level.

Computational and statistical analysis