Evolution of antibody immunity to SARS-CoV-2

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has infected 78 million individuals and is responsible for over 1.7 million deaths to date. Infection is associated with the development of variable levels of antibodies with neutralizing activity, which can protect against infection in animal models^1,2. Antibody levels decrease with time, but, to our knowledge, the nature and quality of the memory B cells that would be required to produce antibodies upon reinfection has not been examined. Here we report on the humoral memory response in a cohort of 87 individuals assessed at 1.3 and 6.2 months after infection with SARS-CoV-2. We find that titres of IgM and IgG antibodies against the receptor-binding domain (RBD) of the spike protein of SARS-CoV-2 decrease significantly over this time period, with IgA being less affected. Concurrently, neutralizing activity in plasma decreases by fivefold in pseudotype virus assays. By contrast, the number of RBD-specific memory B cells remains unchanged at 6.2 months after infection. Memory B cells display clonal turnover after 6.2 months, and the antibodies that they express have greater somatic hypermutation, resistance to RBD mutations and increased potency, indicative of continued evolution of the humoral response. Immunofluorescence and PCR analyses of intestinal biopsies obtained from asymptomatic individuals at 4 months after the onset of coronavirus disease 2019 (COVID-19) revealed the persistence of SARS-CoV-2 nucleic acids and immunoreactivity in the small bowel of 7 out of 14 individuals. We conclude that the memory B cell response to SARS-CoV-2 evolves between 1.3 and 6.2 months after infection in a manner that is consistent with antigen persistence.

Extended Data Fig. 1 |. Clinical correlates of plasma antibody titres.

a, Normalized AUC anti-RBD IgG titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. b, Normalized AUC anti-RBD IgG titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. c, Normalized AUC anti-RBD IgM titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. d, Normalized AUC anti-RBD IgM titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. e, Normalized AUC anti-RBD IgA titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. f, Normalized AUC anti-RBD IgA titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. g, Normalized AUC anti-N IgG titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. h, Normalized AUC anti-N IgG titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. i, Index values (IV) of anti-RBD IgG titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. j, Index values of anti-RBD IgG titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. k, Cut-off index (COI) values of anti-N total Ig titres at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. l, COI values of anti-N total Ig titres at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. m, NT₅₀ values at 1.3 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. n, NT₅₀ values at 6.2 months for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. o, Age in years for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. p, Severity of acute infection as assessed by the WHO ‘Ordinal Clinical Progression/Improvement Scale’ for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. q, Duration of symptoms during acute infection for participants with (n = 38) or without (n = 49) persistent post-acute symptoms. Horizontal bars indicate median values. Statistical significance was determined using two-tailed Mann–Whitney U-tests.

Extended Data Fig. 2 |. Correlations of plasma antibody measurements.

a, Normalized AUC for IgG anti-RBD plotted against Pylon IgG anti-RBD index values at 1.3 months. b, Normalized AUC for IgG anti-RBD plotted against Pylon IgG anti-RBD index values at 6.2 months. c, Normalized AUC for IgM anti-RBD plotted against Pylon IgM anti-RBD index values at 1.3 months. d, Normalized AUC for IgM anti-RBD plotted against Pylon IgM anti-RBD index values at 6.2 months. e, Normalized AUC for IgG anti-N plotted against Roche COI values for anti-N total Ig titres at 1.3 months. f, Normalized AUC for IgG anti-N plotted against Roche COI values for anti-N total Ig titres at 6.2 months. g, Anti-N total Ig COI values for 80 individuals at the initial 1.3- and 6.2-month follow-up visit. h, Relative change in anti-N total Ig levels between 1.3 and 6.2 months plotted against the anti-N total Ig levels at 1.3 months. i, Normalized AUC for IgM anti-RBD at 6.2 months plotted against IgM anti-RBD at 1.3 months. j, Normalized AUC for IgG anti-RBD at 6.2 months plotted against IgG anti-RBD at 1.3 months. k, Normalized AUC for IgA anti-RBD at 6.2 months plotted against IgA anti-RBD at 1.3 months. l, Normalized AUC for IgG anti-N at 6.2 months plotted against IgG anti-N at 1.3 months. m, COI values for anti-N total Ig titres at 6.2 months plotted against anti-N total Ig titres at 1.3 months. n, Anti-RBD IgM titres at 1.3 months plotted against anti-N IgG titres at 1.3 months. o, Anti-RBD IgG titres at 1.3 months plotted against anti-N IgG titres at 1.3 months. p, Anti-RBD IgA titres at 1.3 months plotted against anti-N IgG titres at 1.3 months. q, NT₅₀ values at 1.3 months plotted against anti-RBD IgG titres at 1.3 months. r, NT₅₀ values at 6.2 months plotted against anti-RBD IgG titres at 6.2 months. The r and P values were determined by two-tailed Spearman’s correlations.

Extended Data Fig. 3 |. Persistent longitudinal changes in the phenotypic landscape of B cells in individuals recovered from COVID-19.

a, Global viSNE projection of pooled B cells for all participants pooled shown in background contour plots, with overlaid projections of concatenated controls, and convalescent participants at 1.3 and 6.2 months. b, viSNE projection of pooled B cells for all participants of B cell clusters identified by FlowSOM clustering. Column-scaled z-scores of median fluorescence intensity (MFI) as indicated by cluster and marker. c, Frequency of B cells from each group in FlowSOM clusters indicated. Each circle represents an individual control individual (n = 20) (grey), convalescent participant at 1.3-month post-infection (n = 41) (red) or convalescent participant at 6.2 months post-infection (n = 41) (green). Horizontal bars indicate mean values. Significance determined by two-tailed paired t-test for comparisons between time points within individuals and two-tailed unpaired t-test for comparison between controls and convalescent individuals. d, Individual viSNE projections of indicated protein expression.

Extended Data Fig. 4 |. Flow cytometry.

a, Gating strategy used for cell sorting. Gating was on singlets that were CD20⁺ and CD3⁻CD8⁻CD16⁻Ova⁻. Sorted cells were RBD–PE⁺ and RBD–AF647⁺. b, Flow cytometry showing the percentage of RBD-double-positive memory B cells from month 1.3 or month 6 post-infection in 21 randomly selected participants.

Extended Data Fig. 5 |. Frequency distributions of human V genes.

a, Two-sided binomial tests were used to compare the frequency distributions of human V genes of anti-SARS-CoV-2 antibodies from donors at 1.3 months and 6.2 months. *P < 0.05, **P < 0.01 ***P < 0.001, ****P < 0.0001. b, Two-sided binomial tests were used to compare the frequency distributions of human V genes of anti-SARS-CoV-2 antibodies from this study to sequences from ref. . *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. c, Sequences from all six individuals with clonal relationships depicted as circos plots as in Fig. 2d. Interconnecting lines indicate the relationship between antibodies that share V and J gene-segment sequences at both IGH and IGL. Purple, green and grey lines connect related clones, clones and singles, and singles to each other, respectively. d, For each participant, the number of IgG heavy-chain sequences (black) analysed at month 1.3 (left) or month 6.2 post-infection (right). The number in the inner circle indicates the number of cells that was sorted for each individual denoted above the circle. e, The same as d but showing combined data for all six participants. f, Comparison of the percentage of IgG-positive B cells from all six individuals at month 1.3 or month 6.2 post-infection. The horizontal bars indicate the mean. Statistical significance was determined using two-tailed t-test.

Extended Data Fig. 6 |. Analysis of antibody somatic hypermutation of persisting clones, CDR3 length and hydrophobicity.

a–f, Number of somatic nucleotide mutations in both the IGVH and IGVL of persisting clones found at month-1.3 and month-6.2 time points in six participants: COV21 (a), COV47 (b), COV57 (c), COV72 (d), COV96 (e) and COV107 (f). The VH and VL gene usage of each clonal expansion is indicated above the graphs, or are indicated as ‘singlets’ if the persisting sequence was isolated only once at both time points. Connecting line indicates the somatic hypermutation of the clonal pairs that were expressed as a recombinant monoclonal antibodies. g, For each individual, the amino acid length of the CDR3 at IGVH and IGVL is shown. The horizontal bars indicate the mean. The number of antibody sequences (IGVH and IGVL) evaluated for each participant are n = 90 (COV21), n = 78 (COV47), n = 53 (COV57), n = 87 (COV72), n = 104 (COV96), n = 120 (COV107). Right, all antibodies combined (n = 532 for both IGVH and IGVL). Statistical significance was determined using two-tailed Mann–Whitney U-tests and horizontal bars indicate median values. NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. h, Distribution of the hydrophobicity GRAVY scores at the IGH CDR3 in 532 antibody sequences from this study compared to a public database: statistical analysis is provided in Methods. The box limits are at the lower and upper quartiles, the centre line indicates the median, the whiskers are 1.5× interquartile range and the dots represent outliers. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. NS, not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

Extended Data Fig. 7 |. ELISA of wild-type or mutant RBD for monoclonal antibodies.

a, EC₅₀ values for binding to wild-type RBD of shared singlets and shared clones of monoclonal antibodies obtained at the initial 1.3- and 6.2-month follow-up visit, divided by participant (n = 6 (COV21), n = 13 (COV47), n = 3 (COV57), n = 6 (COV72), n = 15 (COV96), n = 9 (COV107)). Lines connect shared singlets or clones. Monoclonal antibodies with improved EC₅₀ at the 6.2-month follow-up visit are highlighted in green; remaining monoclonal antibodies are shown in black. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. b–j, Graphs show ELISA binding curves for different antibodies obtained at 1.3 months (dashed lines) and their clonal relatives obtained after 6.2 months (solid lines) binding to wild type, R346S, E484K, Q493R, N439K, N440K, A475V, S477N, V483A and V367F RBDs (colours as indicated). Antibody identifiers of pairs are as indicated on top of panels (1.3 months/6.2 months). k, Heat map shows log₂-transformed relative fold change in EC₅₀ against the indicated RBD mutants for 52 antibody clonal pairs obtained at 1.3 (black) and 6.2 months (red). The clonal and participant origin for each antibody pair is indicated above.

Extended Data Fig. 8 |. Neutralization of wild-type and mutant RBDs, C51 alignment and binding projection.

a, IC₅₀ values of shared singlets and shared clones of monoclonal antibodies obtained at the initial 1.3- and 6.2-months follow-up visit, divided by participant (n = 6 (COV21), n = 13 (COV47), n = 3 (COV57), n = 6 (COV72), n = 15 (COV96), n = 9 (COV107)). Lines connect shared singlets or clones. Monoclonal antibodies with undetectable IC₅₀ at 1.3 months are plotted at 10 μg ml⁻¹ and are highlighted in red; monoclobal antibodies with improved IC₅₀ at the 6.2-month follow-up visit are highlighted in green; remaining monoclonal antibodies are shown in black. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. b–f, The normalized relative luminescence values for cell lysates of 293T_ACE2 cells 48 h after infection with SARS-CoV-2 pseudovirus containing wild-type RBD or RBD mutants (wild-type, Q493R, E484G and R346S RBD shown in black, red, green and blue, respectively) in the presence of increasing concentrations of monoclonal antibodies obtained at the 1.3-month initial visit (1.3m, dashed lines) and their shared clones or singlets at the 6.2-month follow-up visit (6.2m, continuous lines). Antibody identifiers are as indicated. g, VH and VL amino acid sequence alignment of C144 and derivative antibodies C051, C052, C053 and C054. Germline-encoded residues are highlighted in green. Residues in the proximity of RBD-binding C144 paratope are highlighted in red. h–j, Surface representation of two adjacent ‘down’ RBDs (RBD_A and RBD_B) on a spike trimer with the C144 epitope on the RBDs highlighted in cyan and positions of amino acid mutations that accumulated in C052 (h), C053 (i) and C054 (j), compared to the parent antibody C144, highlighted as stick side chains on a Cα atom representation. The C052, C053 and C054 interactions with two RBDs was modelled on the basis of a cryo-electron microscopy structure of C144 Fab bound to spike trimer.

Extended Data Fig. 9 |. SARS-CoV-2 antigen in human enterocytes in the gastrointestinal tract at three months after COVID-19 diagnosis, and pre-COVID-19 control individuals without detectable SARS-CoV-2 antigen.

a–j, Immunofluorescence images of human gastrointestinal tissue are shown. Staining is for EPCAM (red), DAPI (blue) and SARS-CoV-2 nucleocapsid (green) Samples are derived from intestinal biopsies in the gastrointestinal tract as indicated. a–h, Biopsies from participant CGI-088 (Supplementary Table 7) taken 92 d after the onset of COVID-19 symptoms. i, j, Biopsy taken 27 months before the onset of COVID-19 symptoms from participant CGI-088. Arrows indicate enterocytes with detectable SARS-CoV-2 antigen. Isotype and no-primary controls for each tissue are shown in the right two columns. Scale bars, 100 μm. k, l, Immunofluorescence images of biopsy samples in the gastrointestinal tract (ileal (k) and duodenal (l)) obtained from ten preCOVID-19 control individuals between January 2018 and October 2019 are shown. Staining is for EPCAM (red), DAPI (blue) and SARS-CoV-2 nucleocapsid (green). Scale bars, 100 μm. All experiments were repeated independently at least twice with similar results.

Extended Data Fig. 10 |. SARS-CoV-2 antigen and RNA is detectable in different intestinal segments in several individuals convalescent from COVID-19.

a, Immunofluorescence images of biopsy samples in the gastrointestinal tract in different individuals are shown. Staining is for EPCAM (red), DAPI (blue) and SARS-CoV-2 nucleocapsid (green). Samples are derived from intestinal biopsies from four participants (CGI-089, CGI-092, CGI-100 and CGI-106) taken at least three months after COVID-19 infection. Arrows indicate enterocytes with detectable SARS-CoV-2 antigen. Scale bars, 100 μm. The experiments were repeated independently at least twice with similar results. b, Quantification of SARS-CoV-2-positive cells by immunofluorescence. The number of cells staining positive for the N protein of SARS-CoV-2 per mm² of intestinal epithelium is shown. The graphs show biopsy samples from the indicated individuals of the duodenum (left) and terminal ileum (right). Black dots represent the number of available biopsy specimen for each individual from the respective intestinal segment (CGI-088, n = 4 duodenal and n = 2 ileal; CGI-089, n = 2 duodenal and n = 2 ileal; CGI-092, n = 6 duodenal and n = 3 ileal; CGI-106, n = 4; CGI-100, n = 4). Boxes represent median values and whiskers the 95% confidence interval. c, d, SARS-CoV-2 viral RNA was visualized in intestinal biopsies of participant CGI-089 (c) and CGI-088 (d) using in situ hybridization. SARS-CoV-2 genomic RNA (black) and haematoxylin and eosin staining by smFISH–immunohistochemistry technique in the duodenum (left) or terminal ileum (right). Arrows indicate enterocytes with detectable SARS-CoV-2 RNA. e, Pre-COVID-19 control individuals show no detectable SARS-CoV-2 viral RNA in duodenal (left) or ileal (right) biopsies. Scale bars, 100 μm. The experiments were repeated independently three times with similar results.

Fig. 1 |. Plasma antibody dynamics against SARS-CoV-2.

a–d, Results of ELISAs measuring plasma reactivity to RBD (a, b, c) and N protein (d) at the initial 1.3- and 6.2-month follow-up visit, respectively. a, Anti-RBD IgM. b, Anti-RBD IgG. c, Anti-RBD IgA d, Anti-N IgG. The normalized area under the curve (AUC) values for 87 individuals are shown in a–d for both time points. Positive and negative controls were included for validation. e, Relative change in plasma antibody levels between 1.3 and 6.2 months for anti-RBD IgM, IgG, IgA and anti-N IgG in all 87 individuals. f–i, Relative change in antibody levels between 1.3 and 6.2 months plotted against the corresponding antibody levels at 1.3 months. f, Anti-RBD IgM. r = −0.83, P < 0.0001. g, Anti-RBD IgG. r = −0.76, 1.3 months 6.2 months 100 101 102 103 104 105 NT₅₀ at 1.3 months P < 0.0001. h, Anti-RBD IgA. r = −0.67, P < 0.0001. i, Anti-N IgG. r = −0.87, P < 0.0001. j, Ranked average NT₅₀ at 1.3 months (blue) and 6.2 months (red) for the 87 individuals studied. k, Graph shows NT₅₀ for plasma from all 87 individuals collected at 1.3 and 6.2 months. P < 0.0001. l, Relative change in plasma neutralizing titres between 1.3 and 6.2 months plotted against the corresponding titres at 1.3 months. For a–e, k plotted values and horizontal bars indicate geometric mean. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test in a–d, k, and Friedman with Dunn’s multiple comparison test in e. The r and P values in f–i, l were determined by two-tailed Spearman’s correlations.

Fig. 2 |. Sequences of anti-SARS-CoV-2 RBD antibodies.

a, Representative flow cytometry plots showing dual AlexaFluor-647–RBD- and PE–RBD-binding B cells for six study individuals (designated COV21, COV47, COV57, COV72, COV96 and COV107) (the gating strategy is shown in Extended Data Fig. 5). The percentage of antigen-specific B cells is indicated. b, As in a. Graph summarizes the percentage of RBD-binding memory B cells in samples obtained at 1.3 and 6.2 months from 21 randomly selected individuals. Red horizontal bars indicate geometric mean values. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. c, Number of somatic nucleotide mutations in the IGVH (top) and IGVL (bottom) genes in antibodies obtained after 1.3 or 6.2 months from the indicated individual (left) or all six donors (right). Statistical significance was determined using two-tailed Mann–Whitney U-tests. Horizontal bars indicate median values. d, Pie charts show the distribution of antibody sequences from six individuals after 1.3 (top) or 6.2 months (bottom). The number in the inner circle indicates the number of sequences analysed for the individual denoted above the circle. Pie-slice size is proportional to the number of clonally related sequences. The black outline indicates the frequency of clonally expanded sequences detected in each participant. Coloured slices indicate persisting clones (same IGHV and IGLV genes and highly similar CDR3) found at both time points in the same participant. Grey slices indicate clones unique to the time point. White slices indicate singlets found at both time points, and the remaining white area indicates sequences that were isolated once. e, Graph shows relative clonality at both time points for all six donors. Red horizontal bars indicate mean values. Statistical significance was determined using two-tailed t-test.

Fig. 3 |. Reactivity of anti-SARS-CoV-2 RBD monoclonal antibodies.

a, Graph shows anti-SARS-CoV-2 RBD antibody reactivity. ELISA EC₅₀ values for all antibodies measured at 1.3 months, and 122 selected monoclonal antibodies measured at 6.2 months. Horizontal bars indicate geometric mean. Statistical significance was determined using two-tailed Mann–Whitney U-test. b, EC₅₀ values for all 52 antibodies that appear at 1.3 and 6.2 months. Average of two or more experiments. Horizontal bars indicate geometric mean. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. c, Surface representation of the RBD with the ACE2-binding footprint indicated as a dotted line and selected residues found in circulating strains (grey) and residues that mediate resistance to class-2 (red) (C144) and −3 (green) (C135) antibodies highlighted as sticks. d, Graphs show ELISA binding curves for C144 (black dashed line) and its clonal relatives obtained after 6.2 months (C050, C051, C052, C053 and C054) (solid lines) binding to wild-type (WT), Q493R, R346S and E484K RBDs. e. Heat map shows log₂-transformed relative fold change in EC₅₀ against indicated RBD mutants for 26 antibody clonal pairs obtained at 1.3 and 6.2 months with the most pronounced changes in reactivity. The participant of origin for each antibody pair is indicated above. All experiments were performed at least twice.

Fig. 4 |. Neutralizing activity of anti-SARS-CoV-2 RBD monoclonal antibodies.

a, SARS-CoV-2 pseudovirus neutralization assay. IC₅₀ values for all antibodies measured at 1.3 months, and 122 selected antibodies measured at 6.2 months. Antibodies with IC₅₀ values above 1 μg ml⁻¹ were plotted at 1 μg ml⁻¹. Mean of two independent experiments. Red bar indicates geometric mean. Statistical significance was determined using two-tailed Mann–Whitney U-test. b, IC₅₀ values for 52 antibodies appearing at 1.3 and 6.2 months. Red bar indicates geometric mean. Statistical significance was determined using two-tailed Wilcoxon matched-pairs signed-rank test. c, IC₅₀ values for 5 pairs of monoclonal antibody clonal relatives obtained after 1.3 or 6.2 months for neutralization of wild-type and mutant SARS-CoV-2 pseudovirus. Antibody identifiers of the 1.3-month–6.2-month monoclonal antibody pairs as indicated. Bold styling denotes antibody pairs with substantial increase in neutralizing activity after 6.2 months. d, Graph shows the normalized relative luminescence units (RLU) for cell lysates of 293T cells expressing ACE2, 48 h after infection with SARS-CoV-2 pseudovirus containing wild-type RBD or one of three mutant RBDs (Q493R, E484G and R346S) in the presence of increasing concentrations of one of two monoclonal antibodies C144 (1.3 months) (dashed lines) or C051 (6.2 months) (solid lines). Experiments were performed at least twice. e, C051 binding model. Surface representation of two adjacent ‘down’ RBDs (RBD_A and RBD_B) on a spike trimer, with the C144 epitope on the RBDs highlighted in cyan and positions of amino acid mutations that accumulated in C051 compared to the parent antibody C144 highlighted as stick side chains on a Cα atom representation of C051 V_HV_L binding to adjacent RBDs. The C051 interaction with two RBDs was modelled on the basis of a cryo-electron microscopy structure of C144 Fab bound to spike trimer. HC, heavy chain; LC, light chain.

Fig. 5 |. Immunofluorescence imaging of intestinal biopsies.

a, Immunofluorescence images of human enterocytes stained for EPCAM (red), DAPI (blue) and either ACE2 (green in a, c) or SARS-CoV-2 N (green in b, d) in intestinal biopsies taken 92 d after onset of COVID-19 symptoms in participant CGI-088, in the terminal ileum (a, b) or duodenum (c, d). Regions in white boxes in the right panels of a, c are shown expanded in b, d, respectively. Arrows indicate enterocytes with detectable SARS-CoV-2 antigen. Scale bars, 100 μm. The experiments were repeated independently at least twice with similar results.