Fig 1.
The B cell profiles of convalescent plasma donors were diverse.
(A) Representative flow plots of live, singlet, CD19+ lymphocytes demonstrating 5 primary B cell subsets in one healthy and 4 convalescent subjects. (B-I) Histograms of healthy and convalescent donor frequencies of (B) total CD19+ B cells, and (C) naïve & transitional type 3, (D) memory, (E) transitional type 1 and 2, (F) plasmablast (G) IgM+ memory, (H) switched Memory, and (I) activated naïve/memory subsets among viable CD19+ lymphocytes. Bars represent mean +/- SD. Statistical analysis between each donor subgroup was done with non-parametric Kruskal-Wallis test with Dunn’s correction for multiple comparisons. Adjusted p value was used to determine family-wise significance at alpha = 0.05. Healthy control and total convalescent groups also compared by Mann-Whitney test with two-tailed p value, alpha = 0.05. Healthy control n = 24, conv. total n = 40, asymp. n = 5, conv. early n = 12, conv. mid n = 6, conv. late n = 17, except for (E) and (F), where 1 and 2 healthy control statistical outliers were omitted, respectively.
Fig 2.
The frequency of memory B cells in the peripheral blood was correlated with shorter symptom duration following infection with SARS-CoV-2, and stable over time.
(A-C) Scatterplot correlation of symptom duration in days, vs. the frequency among viable CD19+ B cells of (A) total memory, (B) IgM+ memory, and (C) switched memory. (D-F) Scatterplot correlation of days since last symptom vs. (D) IgM+ memory, (E) switched memory, and (F) plasmablast frequency among viable CD19+ lymphocytes. (G-I) Scatterplot correlation of days since symptom onset vs. (G) IgM+ memory, (H) switched memory, and (I) plasmablast frequency among viable CD19+ B cells. (J-L) Scatterplot correlation of symptom duration in days, vs. (J) naïve and transitional type 3, (K) transitional type 1 and 2, and (L) plasmablast frequency among viable CD19+ B cells. Pearson’s correlation r value and 95% confidence intervals shown with two-tailed p values, alpha = 0.05. n = 35 (all symptomatic subjects) for all.
Fig 3.
Anti-spike RBD antibody levels correlated with memory B cell frequency in most seropositive convalescent plasma donors.
(A-F) Correlation of area under the curve for anti-RBD plasma absorbance vs. memory B cell frequency from convalescent subjects seropositive for each isotype (IgG1 & IgG3 n = 15, IgM n = 16). Subjects whose data points are circled in (A) were isolated from apparent outliers for statistical analysis of IgG1 in (B, G, and J; n = 12). (G-I) Scatterplot correlation of area under the curve for anti-RBD plasma absorbance vs. IgM+ memory B cell frequency among viable CD19+ lymphocytes. (J-L) Correlation of area under the curve for anti-RBD plasma absorbance vs. switched memory B cell frequency among viable CD19+ lymphocytes. Spearman’s correlation rs value and 95% confidence intervals shown with two-tailed p value, alpha = 0.05 for all analyses. Regression lines shown to demonstrate trend only. Sample number distribution was the same between total, IgM+ and switched memory B cell analyses.
Fig 4.
Convalescent SARS-CoV-2 subjects displayed a contraction of plasmablasts over time, but maintained memory B cells.
(A-K) Frequency of (A and B) plasmablasts without and with SD, respectively, (C) total memory, (D) switched memory, (E) IgM+ memory, (F) naïve + transitional 3, (G) transitional 1 + 2, (H) T-bet+, (I) CD11c+, (J) DN, and (K) activated cells among CD19+ viable lymphocytes from 15 convalescent plasma donors at initial draw and 3-month follow-up visit. Follow-up donors were not selected, but the first available convalescent plasma donors to consent to a follow-up visit. Each pair of connected points (and color) represents an individual subject. Symbol shapes indicate convalescent subset. Bars represent mean, SD omitted for clarity in all except (B). Initial vs. 3-month intra-individual convalescent samples were compared by Wilcoxon matched-pairs signed rank test with two-tailed p value, alpha = 0.05.