Recurrent waning of anti-SARS-CoV-2 neutralizing antibodies despite multiple antigen encounters

Abstract

Background: SARS-CoV-2 neutralizing antibodies may protect against symptomatic infection in immunized individuals. However, vaccine-induced antibody levels wane over time, reducing vaccine efficacy. The definition of the waning kinetics of neutralizing SARS-CoV-2 responses and the potential impact of sequential antigen encounters are still poorly defined.

Methods: Plasma neutralizing activity was determined in longitudinally collected samples from SARS-CoV-2 infected, primo-vaccinated and boosted individuals. Neutralizing activity decay kinetics were modeled against time using Log-Log and biexponential models.

Results: Neutralizing antibody levels wane after an initial peak in all groups of vaccinated individuals with half-life ranging from 29 to 60 days. Exponential models showed a subsequent stabilization of neutralizing titers to a plateau. Both the peak response and the plateau values depended on vaccine type, infection status and severity of infection. Booster immunization by either vaccines or breakthrough infections did not modify peak, plateau or decay rate values.

Conclusions: Our results indicate that the waning of SARS-CoV-2 neutralizing antibody responses was recurrent even after several antigen encounters. Repeated immunizations would be required to maintain high levels of neutralizing antibodies and protect vulnerable individuals from symptomatic infection.

Keywords: Booster; Breakthrough infection; COVID-19; Humoral response; Hybrid immunity; Kinetics; Neutralization; Vaccine.

Fig. 1

Cohort description. a shows a schematic representation (violin plot) of the distribution of plasma sample collection over time in the different analyzed groups. b shows the demographic characteristics of the different groups. Sex is shown as %, while age, follow-up duration, and samples/individual are shown by the mean and standard deviation. The sample size shown on the right indicates the number of follow-up profiles analyzed in each group

Fig. 2

Longitudinal decay of neutralization activity. a–c show four-parameter logistic dose–response curves from an Infected Hospitalized participant (298–017), an uninfected vaccinated participant (298–445), and a Boosted participant (298–060), respectively. Each curve corresponds to a unique plasma sample collected at the indicated days after antigenic encounter. Data points represent the mean of duplicate measurements, with error bars indicating standard deviation at each dilution. The dotted line marks the 50% neutralization, and the accompanying tables report days since the last antigenic event and ID₅₀ values (reciprocal plasma dilution) for each time point

Fig. 3

Longitudinal analysis of neutralization levels. a Shows the different statistical modeling analyzed for the indicated groups: LOESS (solid lines), biexponential (dashed lines) and Log–Log (dotted lines). Dots represent the neutralization titers from each analyzed sample. b Shows Log–Log models of neutralizing activity decay across all the groups. The inset graph indicates the time (days) in a logarithmic scale, while the main graph indicates the time on a linear scale

Fig. 4

Kinetics of neutralization levels in the final defined groups. a shows single decay models with plateau for all groups (All Boost group include Uninfected Boosted, Infected Boosted and BTI subgroups). b–d Show the estimated values and the 95% confidence interval for half-life, peak and plateau parameters of the model. Welch’s t-test: p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (****). For b, test was performed using decay rate estimates, from which half-life is calculated. In b, the gray area indicates that half-life of Infected Mild group could not be properly estimated and was excluded from the comparisons

Fig. 5

Longitudinal analysis of neutralizing activity based on mRNA vaccine type and prior infection status. a Shows single decay model with plateau (solid curves for BNT162b2 and dashed curves for mRNA-1273) for all four vaccinated groups. Background lines show the individual trajectories of analyzed profiles. b–d Show the estimated values and the 95% confidence interval for half-life, peak and plateau parameters of the model. Welch’s t-test: p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (****). For b, test was performed using decay rate estimates, from which half-life is calculated

Fig. 6

Longitudinal analysis of neutralizing activity based on COVID-19 severity. a Shows single decay model with plateau for Infected-Vaccinated Mild or Hospitalized individuals (solid curves). Background lines show the individual trajectories of analyzed profiles. b–d Show the estimated values and the 95% confidence interval for half-life, peak and plateau parameters of the model. Welch’s t-test: p > 0.05 (ns), p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (****). For b, test was performed using decay rate estimates, from which half-life is calculated. e Shows Log–Log models for Infected-Vaccinated Mild and Hospitalized groups. Time (days) transformed to linear scale. f Shows the model estimation and observed values of neutralization after 1 year (365 ± 60 days) for Infected-Vaccinated Mild and Hospitalized groups, with dotted lines to facilitate comparison. Model estimations are shown on the left part, each group is represented by a central line indicating the estimated value, and whiskers denoting the 95% confidence interval. Wald test: p ≤ 0.0001 (****). In the observed values (right part), each group is represented by individual neutralization values, a central line indicating the median, and whiskers denoting the interquartile range. Mann–Whitney test: p ≤ 0.001 (***)