key: cord-0880429-7t22f7cq
authors: Natarajan, Harini; Crowley, Andrew R.; Butler, Savannah E.; Xu, Shiwei; Weiner, Joshua A.; Bloch, Evan M.; Littlefield, Kirsten; Wieland-Alter, Wendy; Connor, Ruth I.; Wright, Peter F.; Benner, Sarah E.; Bonny, Tania S.; Laeyendecker, Oliver; Sullivan, David; Shoham, Shmuel; Quinn, Thomas C.; Larman, H. Benjamin; Casadevall, Arturo; Pekosz, Andrew; Redd, Andrew D.; Tobian, Aaron A. R.; Ackerman, Margaret E.
title: Markers of Polyfunctional SARS-CoV-2 Antibodies in Convalescent Plasma
date: 2021-04-20
journal: mBio
DOI: 10.1128/mbio.00765-21
sha: 5f00666614d9159a71e1630afb3cbd4032f39899
doc_id: 880429
cord_uid: 7t22f7cq

Convalescent plasma is a promising therapy for coronavirus disease 2019 (COVID-19), but the antibody characteristics that contribute to efficacy remain poorly understood. This study analyzed plasma samples from 126 eligible convalescent blood donors in addition to 15 naive individuals, as well as an additional 20 convalescent individuals as a validation cohort. Multiplexed Fc Array binding assays and functional antibody response assays were utilized to evaluate severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antibody composition and activity. Donor convalescent plasma samples contained a range of antibody cell- and complement-mediated effector functions, indicating the diverse antiviral activity of humoral responses observed among recovered individuals. In addition to viral neutralization, convalescent plasma samples contained antibodies capable of mediating such Fc-dependent functions as complement activation, phagocytosis, and antibody-dependent cellular cytotoxicity against SARS-CoV-2. Plasma samples from a fraction of eligible donors exhibited high activity across all activities evaluated. These polyfunctional plasma samples could be identified with high accuracy with even single Fc Array features, whose correlation with polyfunctional activity was confirmed in the validation cohort. Collectively, these results expand understanding of the diversity of antibody-mediated antiviral functions associated with convalescent plasma, and the polyfunctional antiviral functions suggest that it could retain activity even when its neutralizing capacity is reduced by mutations in variant SARS-CoV-2.

ABSTRACT Convalescent plasma is a promising therapy for coronavirus disease 2019 (COVID- 19) , but the antibody characteristics that contribute to efficacy remain poorly understood. This study analyzed plasma samples from 126 eligible convalescent blood donors in addition to 15 naive individuals, as well as an additional 20 convalescent individuals as a validation cohort. Multiplexed Fc Array binding assays and functional antibody response assays were utilized to evaluate severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antibody composition and activity. Donor convalescent plasma samples contained a range of antibody cell-and complement-mediated effector functions, indicating the diverse antiviral activity of humoral responses observed among recovered individuals. In addition to viral neutralization, convalescent plasma samples contained antibodies capable of mediating such Fc-dependent functions as complement activation, phagocytosis, and antibody-dependent cellular cytotoxicity against SARS-CoV-2. Plasma samples from a fraction of eligible donors exhibited high activity across all activities evaluated. These polyfunctional plasma samples could be identified with high accuracy with even single Fc Array features, whose correlation with polyfunctional activity was confirmed in the validation cohort. Collectively, these results expand understanding of the diversity of antibody-mediated antiviral functions associated with convalescent plasma, and the polyfunctional antiviral functions suggest that it could retain activity even when its neutralizing capacity is reduced by mutations in variant SARS-CoV-2.

IMPORTANCE Convalescent plasma has been deployed globally as a treatment for COVID-19, but efficacy has been mixed. Better understanding of the antibody characteristics that may contribute to its antiviral effects is important for this intervention as well as offer insights into correlates of vaccine-mediated protection. Here, a survey of convalescent plasma activities, including antibody neutralization and diverse effector functions, was used to define plasma samples with broad activity profiles. These polyfunctional plasma samples could be reliably identified in multiple cohorts by multiplex assay, presenting a widely deployable screening test for plasma selection and investigation of vaccine-elicited responses.

S ince its emergence in 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has spread rapidly and infected over 115 million individuals worldwide. As the medical community has mobilized to identify effective therapies to combat the virus, treatment with convalescent plasma derived from individuals who have recovered from coronavirus disease 2019 (COVID-19) has emerged as a potential therapeutic intervention (1, 2) . Preliminary evidence suggests that patients treated early with convalescent plasma show improved survival and reduced viral load, but efficacy data have been mixed (1, (3) (4) (5) (6) (7) (8) (9) (10) . While the Expanded Access Program demonstrated the efficacy of convalescent plasma in a dose-response effect of neutralizing titers, units with low titers were also efficacious, suggesting that there are other contributing activities that have not been measured (11) . Antibody responses resulting from infection are highly variable in magnitude and character (12) (13) (14) . Thus, a better understanding of the breadth and spectrum of antiviral activities of the humoral immune response is critical to understanding convalescent plasma as well as offering insights into correlates of vaccine-mediated protection.

Antibody responses to SARS-CoV-2 have exhibited wide variation, not only in titer but also in the quality of the antibody response including neutralization potential (14) (15) (16) (17) (18) . Antibodies directed against the receptor-binding domain of the fusogenic spike (S) protein can block the interaction of the spike with the angiotensin-converting enzyme 2 (ACE2) receptor of airway epithelial cells. Such neutralizing antibodies have demonstrated the ability to inhibit infection in vitro (19, 20) and in vivo (21) (22) (23) . Accordingly, neutralizing responses have been a key target in development of vaccines to prevent SARS-CoV-2 and monoclonal antibodies to treat COVID-19 disease (18) . Recent data suggest that the frequency of neutralizing antibodies (nAbs) within the total humoral response could be quite low (24, 25) and that many antibodies are directed toward nonneutralizing epitopes within more conserved regions of the S protein (26) (27) (28) . In the absence of sufficient levels of direct antiviral activity via antibody (Ab)-mediated blocking, the burden for humoral protection falls to the extraneutralizing effector functions, which are initiated by the relatively constant domain (Fc) of virus-specific antibodies and executed by innate immune cells and the complement cascade. By engaging soluble and cell surface-expressed Fc receptors, antibodies can trigger a variety of functions such as phagocytosis, cellular cytotoxicity, and complement deposition, which play an important role in clearing diverse viral infections (29, 30) . In the context of SARS-CoV-1, antibody-mediated phagocytosis has been observed to play a critical role in clearing infection in vivo, even in the context of existing potent neutralization activity (31) .

Further, these activities, which in principle can be mediated by a greater diversity of epitope specificities, may gain in importance in the context of emerging neutralization-resistant viral variants (32) (33) (34) . Even in the context of potent neutralization activity, there is growing appreciation of the role that effector functions play in vivo. Multiple passive transfer experiments have shown that effector mechanisms contribute to the antiviral activity of monoclonal antibodies (35) (36) (37) (38) and polyclonal antibodies raised in the context of vaccination (39, 40) . This evidence extends beyond correlative observations (37, 39) to include mechanistic evidence of in vivo contributions via Fc sequence engineering to knock out or enhance these activities (36, 38) , as well as on the basis of depletion of effector cells (38) . Such findings indicate that understanding the ability of convalescent donor plasma to elicit these effector functions in diverse infected subjects may contribute to efficacy.

Biophysical characterization of SARS-CoV-2 convalescent plasma. Convalescent plasma samples from 126 eligible donors from the Baltimore/Washington, DC area (Johns Hopkins Medical Institutions [JHMI] cohort) (14) and serum samples from 15 naive controls and 20 convalescent subjects from New Hampshire (Dartmouth-Hitchcock Medical Center [DHMC] cohort) (41) serving as a validation cohort were collected (see Table S1 in the supplemental material). Antibody responses were evaluated using an Fc Array assay that assesses both variable fragment (Fv) and Fc domain characteristics of antibodies (42) . The assay was customized to assess responses across a panel of SARS-CoV-2 antigens. This panel consisted of the nucleocapsid (N) protein, stabilized trimeric spike protein (S-2P) (43) , spike subdomains, including S1 and S2, the receptorbinding domain (RBD), and the fusion peptide. Influenza hemagglutinin (HA) and herpes simplex virus gE antigens served as controls. Characterization extended beyond antigen specificity to include antibody isotype, subclass, and propensity to bind Fc receptors (FcRs).

Diverse SARS-CoV-2-specific immunoglobulin isotypes and subclasses, particularly IgG1 and IgG3, IgA, and IgM, were elevated in SARS-CoV-2 convalescent subjects across different epitope and antigen specificities ( Fig. 1A ; see also Fig. S1 to S3 in the supplemental material). Robust responses to stabilized spike (S-2P) and N were apparent, and lower magnitudes of responses were detected to functionally relevant RBD and fusion peptide domains. IgA responses were primarily driven by the IgA1 subclass, whereas IgG responses were dominated by IgG1 and IgG3. FcgR binding profiles, which capture the effects of other factors known to impact avid FcgR recognition, such as binding to multiple epitopes, distinct spatial recognition profiles, and differing antibody glycosylation, were distinct from measurements of the overall SARS-CoV-2-specific IgG response (titer), as well as compared to these innate immune activating IgG subclasses.

Relationship between antibody characteristics and clinical characteristics. Differences in the antibody response toward SARS-CoV-2 and endemic CoV associated with sex, age, and hospitalization status were evaluated (Fig. 2) . A chi-squared test was performed to examine whether there was an association between disease severity and gender. It was determined that there was no statistically significant link between these two traits (P = 0.7355). Moreover, a univariate logistic regression analysis was employed to assess the association between age and either gender of hospitalization status (sex: age, P = 0.711; hospitalization: age, P = 0.593). Therefore, the potential confounding effect on the association between biophysical measurements and each clinical trait were eliminated. SARS-CoV-2 IgG, IgA, and FcgR-binding antibodies were significantly elevated in older and male subjects, characteristics which are considered risk factors for more severe disease. Elevated SARS-CoV-2-specific IgG and FcgR-binding antibody features were also observed in hospitalized subjects, consistent with prior studies (14, 41, 44) , and it is possible that IgG responses may drive disease enhancement (45) (46) (47) . However, despite being associated with both age and sex risk factors, elevated IgA features were not observed in hospitalized subjects, consistent with the possibility that IgA responses may contribute to milder infection (41) .

Antibody effector functions. To explore the biological functions of antibodies in convalescent serum donors, both neutralizing and extraneutralizing activities were evaluated (Fig. 3A) . Consistent with the overall SARS-CoV-2 antibody response magnitude, neutralization activity against live virus was higher among donors who were hospitalized than those who were not hospitalized. While antibody-dependent cell-mediated phagocytosis (ADCP), FcgRIIIa activation as a surrogate for antibody-dependent cellular cytotoxicity (ADCC), and antibody-dependent complement deposition (ADCD) elicited by RBD-specific antibodies were likewise elevated among donors who had been hospitalized, correlative relationships between functions showed distinctions among these antiviral activities (Fig. 3B) . ADCP, which has previously been associated with clearance of SARS-CoV-1 in a mouse model (31) , and FcgRIIIa activation (ADCC) were highly correlated with each other (Pearson correlation coefficient [R P ] = 0.82), and moderately correlated with neutralization (R P = 0.64 and 0.57, respectively). Complement activation, which has been associated with increased inflammation and disease pathology in a mouse model of SARS-CoV-1 (48) and which may also contribute to COVID-19 disease (49, 50), was less well correlated with other activities. Antibody features associated with antiviral functions. Because a number of the effector functions were tested specifically against the RBD antigen, we measured the degree and direction of correlation between RBD-specific Ab biophysical features and other Ab functions (Fig. 3C ). ADCP and FcgRIIIa activation (ADCC) were most strongly correlated with FcgR-binding antibodies, and IgG1 and IgG3, which strongly ligate FcgR. Among FcR, correlations with activating FcgRIIa and FcgRIIIa were most strongly correlated, consistent with their known mechanistic relevance to ADCP and FcgRIIIa activation (ADCC).

As previously observed (41) , IgM also positively correlated with neutralization activity. Relationships between serum IgA responses and antibody functions were considerably weaker than those with IgG responses. Correlative relationships with ADCD tended to be weaker, consistent with the strong dependence of this function on spatial aspects of avid antibody binding and immune complex formation that are better captured by detection with the C1q, an initiator of the complement cascade that was not evaluated in this study.

Distinctions between subjects defined by humoral response profiles. To define similarities and differences among donors more globally, dimensionality reduction was performed on biophysical features using Uniform Manifold Approximation and Projection (UMAP) (51) . Subjects were distributed across the antibody biophysical profile landscape into a set of four distinct clusters (Fig. 4A ). Though hospitalized subjects were observed in multiple clusters, they were most prevalent in cluster 2 and adjacent regions of clusters 1 and 3. To understand aspects of the humoral response that distinguished each cluster, univariate testing was performed to determine and depict which Fc Array features were distinct for individual clusters (Fig. S4 ). Relative responses for these features among convalescent donors in each group reflect differences in the magnitude of the response, with cluster 1 having lower humoral responses to SARS-CoV-2 antigens in general, clusters 2 and 3 exhibiting intermediate responses, and cluster 4 typically showing globally elevated antibody responses. Clusters 2 and 3, which both presented with intermediate response magnitudes, were distinguished by relative differences in IgG1 and IgA responses (Fig. S4) .

Identification of polyfunctional plasma samples. Based on the potential clinical value of polyfunctional plasma, the neutralization and effector functions of each subject according to their UMAP cluster was defined (Fig. 4B ). Subjects in cluster 2 exhibited elevated activity across diverse functions, identifying this cluster of donors as possessing polyfunctional plasma. Because screening plasma for multiple individual functional activities is impractical, we sought to identify the ability of individual Fc Array measurements and classical clinical enzyme-linked immunosorbent assay (ELISA) tests to identify Markers of Polyfunctional SARS-CoV-2 Antibodies ® polyfunctional plasma samples. To determine which features were able to distinguish polyfunctional (cluster 2) from nonpolyfunctional (clusters 1, 3, and 4), receiver operator characteristic (ROC) curves were generated for each Fc Array feature and available clinical ELISA. This analysis identified SARS-CoV-2 S-2P-and S2-specific IgG and FcgR-binding antibodies as exhibiting good discriminatory capacity (Fig. 5A) . Area under the ROC curve for the top nine Fc Array features was superior to that of S1-based clinical ELISAs (Fig. 5B) .

To validate this observation, the strength and significance of correlations between these features and a polyfunctionality score for the serum samples from the validation cohort was evaluated (Fig. 5B ). Strong and significant correlations were observed for both FcgR-binding antibodies specific to stabilized spike (S-2P) and the S2 domain, as well as the magnitude of total IgG responses to these proteins. Together, discovery and validation cohorts support the ability of individual high-throughput multiplex assay measurements to predict the polyfunctionality of donor plasma samples. Identification of highly and broadly antiviral donor plasma samples has the potential to improve treatment outcomes or prevent reinfection.

Convalescent plasma is one of the leading treatments of hospitalized patients for COVID-19. Following transfusion of more than 100,000 individuals in the United States with convalescent plasma, the FDA issued an expanded use authorization. In general, convalescent plasma appears to be most effective when high-titer units are provided early, but efficacy data have been mixed (5, 10) . Thus, it is important to establish the specific, measurable qualities of convalescent plasma that may help us to understand the mechanism of action and resistance of convalescent donors to reinfection.

SARS-CoV-2-specific antibodies can elicit diverse antiviral functions beyond neutralization. These less well characterized functions were measured and related to biophysical antibody profiles. Effector functions were most strongly correlated with FcgR-binding antibodies, IgG1, and IgG3. Beyond IgG, neutralization was also correlated with IgM responses, which may suggest the development of novel responses, as opposed to reactivation of responses to endemic CoV. SARS-CoV-2-specific IgM has also attracted interest because of its association with lower risk of death from COVID-19 (12) , and recent evidence of direct neutralization activity (52, 53) . Evidence in support of the relevance of nonneutralizing mechanisms of antibody-mediated protection against SARS-CoV-2 is accruing (35) (36) (37) (38) (39) (40) , and there is evidence that both ADCC and phagocytosis can contribute antiviral effects against other coronaviruses (54) (55) (56) . Thus, activities measured by these and other in vitro assays have demonstrated the ability to play important roles in antibody-mediated defense against SARS-CoV-2, and as observed in a number of studies in other infectious disease settings (57) (58) (59) (60) (61) (62) .

The antibody responses measured in convalescent subjects in this study were highly diverse, both in the SARS-CoV-2 antigens recognized and the magnitude of the responses; the latter observation is largely characteristic of the humoral responses measured to date (12, 13) . Some of these differences are likely associated with the different manifestations of disease, but others may relate to the cross-sectional nature of the cohort and factors such as differences in time since infection. Other limitations of this study range from cohort composition to the experimental and analytical approaches employed. Individuals in the naive control cohort were generally younger and from a different geographic location. Recombinant antigen and lab-adapted cell lines were employed for several of the functional assays, and the substitution of surrogate measurements such as FcgRIIIa activation was made in place of infected cell death. Thus, while the value of such assays has been established in a diversity of other infectious disease settings (57) (58) (59) (60) (61) (62) , as well as in SARS-CoV-2 in animal models (35) (36) (37) (38) (39) (40) , in vitro function may differ substantially from the in vivo processes these assays are meant to mimic. Though performed on dilute plasma, these assay results could also include 

® effects of plasma components other than antibodies, such as cytokines and complement factors, which cannot be disambiguated from the data collected, and may or may not be relevant to the effects of convalescent plasma therapy (63) . Additionally, the convalescent and naive subjects enrolled in the DHMC cohort provided serum samples, whereas the convalescent subjects in the JHMI cohort contributed plasma samples, which could result in differences in antibody detection and functional activity. Nevertheless, the Fc Array feature that best identified polyfunctional convalescentphase plasma samples was also able to accurately identify polyfunctional convalescent-phase serum samples.

In summary, this study establishes three Fc-dependent activities in convalescent plasma beyond viral neutralization that could have antiviral effects against SARS-CoV-2, namely, ADCC, phagocytosis, and complement activation. These activities could explain therapeutic effects of plasma samples with low neutralizing capacity (11) . With this study we also provide a proof of principle for the correlation of diverse antiviral activities against SARS-CoV-2 using biophysical inputs more amenable to high-throughput measurement. The recent futility-based termination of the ACTIV-3 clinical trial of neutralizing antibody bamlanivimab in hospitalized subjects points to potential challenges to therapy based on individual monoclonal antibodies. Taken together, it is possible that optimally effective antibody therapies for SARS-CoV-2 may require a polyclonal antibody response, which can control virus through binding to multiple epitopes and eliciting a range of effector functions. This work begins to define the specificities and Fc domain characteristics of antibodies associated with potent neutralization as well as effector function and to distill a means to identify polyfunctional plasma samples from a single high-throughput multiplex assay readout.

Further insights into the antibody response to COVID-19 among convalescent plasma donors may be of special interest due to the emergence of novel variants of SARS-CoV-2. Whereas convalescent plasma has shown compromised neutralization capability against novel variants (33, 34) , other effector functions that are typically effectively driven by a broad diversity of epitope specificities may prove less sensitive to neutralization escape mutants or other forms of antigenic drift. However, therapeutically desirable plasma antibody functions have yet to be determined in humans. A strong evidence base exists for the role of antibodies in protection based on animal models and in the setting of human immune responses against other CoV (64) (65) (66) and SARS-CoV-2 (67, 68) . This evidence now includes observations of both beneficial as well as detrimental potential consequences of diverse effector functions (35-40, 69, 70) , suggesting that continued analysis of the associations between passively transferred plasma characteristics and patient outcomes will be key to identifying the recipients who are most likely to benefit and the donors most likely to provide that benefit in the context of the COVID-19 pandemic.

Human subjects. The principal cohort of the study has been previously described (14) . Briefly, it comprised 126 adult subjects (mean age, 43 years; range, 19 to 78 years) previously diagnosed with SARS-CoV-2 infection by PCR-positive (PCR1) nasal swab who met the standard eligibility criteria for blood donation and were collected in the Baltimore, MD and Washington, DC area (Johns Hopkins Medical Institutions [JHMI] cohort). The cohort was composed of 68 males (54.0%) and 58 females (46.0%). Eleven cases (8.7%) were severe enough to require hospitalization (mean duration of stay, 4 days; range, 1 to 8 days). The validation cohort comprised 20 SARS-CoV-2 convalescent individuals from the Hanover, New Hampshire area (Dartmouth Hitchcock Medical Center [DHMC] cohort) (mean age, 40 years; range, 18 to 77 years) and comprised 10 males and 10 females among which 4 subjects (20%) were hospitalized. Infection with SARS-CoV-2 was confirmed in all convalescent subjects by realtime reverse transcriptase polymerase chain reaction of a nasopharyngeal swab. Plasma (JHMI) or serum (DHMC) was collected from each donor approximately 1 month after symptom onset or first positive PCR test in the case of mild or asymptomatic disease (see Table S1 in the supplemental material).

Human subject research was approved by both the Johns Hopkins University School of Medicine's Institutional Review Board and the Dartmouth-Hitchcock Medical Center Committee for the Protection of Human Subjects. All participants provided informed written consent.

Antigen and Fc receptor expression and purification. Prefusion-stabilized, trimer-forming spike protomers (S-2P) of SARS-CoV-2 and a fusion of the receptor-binding domain of SARS-CoV-2 N terminally to a monomeric human IgG4 Fc domain were transiently expressed in either Expi 293 or Freestyle 293-F cells and purified via affinity chromatography, all according to the manufacturers' protocols, as previously described (41) . Human FcgR were expressed and purified as described previously (71) .

Fc Array assay. Coronavirus antigens-including S trimers, S subdomains (i.e., S1 and S2), and other viral proteins from SARS-CoV-2 (Table S2 )-and the control antigens influenza HA and herpes simplex virus (HSV) gE proteins were covalently coupled to Luminex Magplex magnetic microspheres using twostep carbodiimide chemistry as previously described (70) . Biotinylated SARS-CoV-2 fusion peptide was immobilized on neutravidin-coupled microspheres. Pooled polyclonal serum IgG (intravenous immunoglobulin [IVIG]), the SARS CoV-1-specific monoclonal Ab CR3022 that cross-reacts with SARS-CoV-2 S (72), and VRC01, an HIV-specific monoclonal Ab, were used as controls to define bead antigenicity profiles. Pilot experiments were used to determine the optimal dilution of plasma for titrations. Test concentrations for plasma ranged from 1:250 to 1:5,000 and were varied per detection reagent (Table S3) . Isotypes and subclasses of antigen-specific Abs were detected using R-phycoerthrin (PE)-conjugated secondary Abs and by FcRs tetramers as previously described (73) . A FlexMap three-dimensional (3D) array reader detected the beads and measured PE fluorescence in order to calculate the median fluorescence intensity (MFI).

Neutralization assays. Plasma samples from SARS-CoV-2 convalescent donors were tested in microneutralization assays using SARS-CoV-2/WA-1/2020 virus (14, 74) obtained from BEI Resources. VeroE6-TMPRSS2 cells were used to propagate the virus and to determined infectious virus titers using a 50% tissue culture infectious dose (TCID 50 ) assay as previously described for SARS-CoV (14, 74) using Institutional Biosafety Committee-approved protocols in biosafety level 3 containment. Twofold dilutions of plasma were incubated with 100 TCID 50 s for 1 h at room temperature in a volume of 100 ml. The virus-plasma solution was then added to one well of VeroE6-TMPRSS2 cells in a 96-well plate and incubated for 6 h before being replaced with medium. After incubation at 37°C for 2 days, the cells were fixed with 150 ml of 4% formaldehyde followed by staining with Naptho blue black (Sigma-Aldrich) and scoring for wells protected from infection. The assay was performed in hextuplicate, and the area under the curve was calculated from the neutralizing antibody curve. Neutralization of the serum samples was tested using a vesicular stomatitis virus (VSV)-SARS-CoV pseudovirus system as previously described (41, 75) , and neutralization was expressed as 60% inhibitory concentration (IC 60 ) values.

Phagocytosis assay. An assay of Ab-dependent phagocytosis by monocytes (ADCP) was performed essentially as described previously (76, 77) . Briefly, 1-mm yellow-green fluorescent microspheres (Thermo catalog no. F8813) covalently conjugated with recombinant RBD were incubated for 3 h with dilute plasma specimens and the human monocytic THP-1 cell line (ATCC, TIB-202). After pelleting, washing, and fixing, phagocytic scores were calculated as the product of the percentage of cells that phagocytosed one or more fluorescent beads and the median fluorescent intensity of this population was measured by flow cytometry with a MACSQuant analyzer (Miltenyi Biotec). CR3022 and VRC01 were Markers of Polyfunctional SARS-CoV-2 Antibodies ® used as positive and negative controls, respectively. Antibody-independent phagocytosis was measured from wells containing cells and beads, but no antibody.

FccRIIIa activation reporter assay. A Jurkat Lucia NFAT reporter cell line (Invivogen, jktl-nfat-cd16) was used to measure the ADCC potential, represented by the extent of FcgRIIIa activation, of each sample. Engagement of the cell surface receptor leads to the secretion of luciferase into the cell culture supernatant. The cells were cultured according to the manufacturer's recommendations. One day prior to performing the assay, a high binding 96-well plate was coated with 1 mg/ml SARS-CoV-2 RBD and incubated at 4°C overnight. The plates were then washed with phosphate-buffered saline (PBS) plus 0.1% Tween 20 and blocked at room temperature for 1 h with PBS plus 2.5% bovine serum albumin (BSA). After washing, dilute plasma and 100,000 cells/well in growth medium lacking antibiotics (with a total volume of 200 ml) were cultured at 37°C for 24 h. The following day, 25 ml of supernatant was drawn from each well and transferred to an opaque, white 96-well plate and 75 ml of QuantiLuc substrate was added. Luminescence was read immediately on a SpectraMax Paradigm plate reader (Molecular Devices) using 1 s of integration time. The reported values are the means of three kinetic reads taken at 0, 2.5, and 5 min. Negative-control wells contained assay medium instead of antibody sample, while cell stimulation cocktail (Thermo catalog no. 00-4970-93) plus an additional 2 mg/ml ionomycin were used to induce expression of the luciferase transgene as a positive control.

Complement deposition assay. Antibody-dependent complement deposition (ADCD) was quantified essentially as previously described (78) . In brief, plasma samples were heat inactivated at 56°C for 30 min prior to a 2-h incubation with multiplex assay microspheres at room temperature. After washing, each sample was incubated for 1 h at room temperature with human complement serum (Sigma catalog no. S1764) at a concentration of 1:50. Samples were washed, sonicated, and incubated for 1 h at room temperature with murine anti-C3b (Cedarlane catalog no. CL7636AP) followed by anti-mouse IgG1-PE secondary Ab (Southern Biotech catalog no. 1070-09) at room temperature for 30 min. After a final wash and sonication, samples were resuspended in Luminex Sheath Fluid and complement deposition in the form of the median fluorescent intensity of the PE measured on a MAGPIX (Luminex Corp.) instrument. Wells lacking Ab and but still containing heat-inactivated human complement serum served as negative controls.

Data analysis and visualization. Basic analysis and visualization were performed using GraphPad Prism. Heatmaps, correlation plots, and boxplots were generated in R version 3.6.1 (supported by R packages pheatmap [79] , corrplot [80] , and ggplot2 [81] ). Hierarchical clustering was used to cluster and visualize data using the Manhattan and Euclidean metrics. Fc Array features were filtered by elimination of features for which the samples exhibited signal within 10 standard deviations (SD) of the technical blank. Fc Array features were log transformed and then scaled and centered by their standard deviation from the mean (z-score). A Student's two-tailed t test with Welch's correction with a cutoff of P = 0.05 was used to define features different between groups. Pearson correlation coefficients were calculated for the correlation matrices. Spearman correlation coefficients were used to relate antibody features to polyfunctionality scores. The Benjamini-Hochberg method was used to adjust determinations of statistical confidence due to multiple hypothesis testing.

UMAP was employed in the R package "umap" version 0.2.6.0 (82) to enable dimensionality reduction of the JHMI Fc Array data set. Upon log transformation, default UMAP parameters were used with the following exceptions: random_state = 45, min_dist = 1E29, knn_repeats: 21, set_op_mix_ratio = 1. DBSCAN (83) was employed in the R package "dbscan" version 01.1-5 to identify clusters within the UMAP reduced dimensions, using settings of eps = 0.65 and MinPts = 5 to define the clusters. To identify features associated with each cluster, individual clusters were compared to the other three clusters using a Student's two-tailed t test with Welch's and Bonferroni's corrections and a cutoff of P = 0.05.

A receiver operating characteristic (ROC) curve was applied to evaluate the performance of biophysical features in discriminating polyfunctional plasma donors from nonpolyfunctional plasma donors. The binary labels were assigned among the subjects in cluster 2 (n = 27) versus the subjects in the rest of the clusters (n = 99). All biophysical features in the JHMI cohort were ranked by the area under the receiver operating characteristic (AUROC). R packages "ROCR" (84) and "pROC" (85) were employed for ROC curve generation and AUROC calculation.

A polyfunctionality score was defined for each validation (DHMC) cohort serum sample as follows. Minimum activities were assigned a value of zero, and maximum activity was assigned a value of one. Relative activity for each function was defined as (observed 2 minimum)/(maximum 2 minimum), and polyfunctionality was defined as the sum of these relative activity scores across functional assays.

Availability of data. Cohort characteristic, Fc Array, and functional assay data are available at https:// github.com/AckermanLab/Natarajan_et_al_COVID_2021.

Supplemental material is available online only. 

We thank all participants who enrolled in this study and the study and laboratory staff who helped collect and process the samples. VSV pseudovirus expression plasmids were provided by Michael Letko We declare that we have no competing interests.

Convalescent plasma treatment of severe COVID-19: a propensity score-matched control study

Deployment of convalescent plasma for the prevention and treatment of COVID-19

Effectiveness of convalescent plasma therapy in severe COVID-19 patients

The evidence for high-titer convalescent plasma in SARS-CoV-2

Early high-titer plasma therapy to prevent severe Covid-19 in older adults

Treatment of coronavirus disease 2019 patients with convalescent plasma reveals a signal of significantly decreased mortality

Significantly decreased mortality in a large cohort of coronavirus disease 2019 (COVID-19) patients transfused early with convalescent plasma containing high-titer anti-severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein IgG

Treatment of 5 critically ill patients with COVID-19 with convalescent plasma

Earlier the better: convalescent plasma

COVID-19 convalescent plasma: interim recommendations from the AABB

Effect of convalescent plasma on mortality among hospitalized patients with COVID-19: initial three-month experience

Distinct early serological signatures track with SARS-CoV-2 survival

China Medical Treatment Expert Group for Covid-19. 2020. Clinical characteristics of coronavirus disease 2019 in China

Sex, age, and hospitalization drive antibody responses in a COVID-19 convalescent plasma donor population

SARS-CoV-2 antibody avidity responses in COVID-19 patients and convalescent plasma donors

SARS-CoV-2-specific CD81 T cell responses in convalescent COVID-19 individuals

Comparative performance of five commercially available serologic assays to detect antibodies to SARS-CoV-2 and identify individuals with high neutralizing titers

The receptor binding domain of the viral spike protein is an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

Human monoclonal antibodies block the binding of SARS-CoV-2 spike protein to angiotensin converting enzyme 2 receptor

A human monoclonal antibody blocking SARS-CoV-2 infection

Evaluation of the mRNA-1273 vaccine against SARS-CoV-2 in nonhuman primates

A SARS-CoV-2 infection model in mice demonstrates protection by neutralizing antibodies

A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2

Characterization of neutralizing antibodies from a SARS-CoV-2 infected individual

Broad neutralization of SARS-related viruses by human monoclonal antibodies

Cross-reactive antibody response between SARS-CoV-2 and SARS-CoV infections

Characterization of spike glycoprotein of SARS-CoV-2 on virus entry and its immune crossreactivity with SARS-CoV

Viral epitope profiling of COVID-19 patients reveals cross-reactivity and correlates of severity

Antibody potency, effector function, and combinations in protection and therapy for SARS-CoV-2 infection in vivo

Dissecting antibody-mediated protection against SARS-CoV-2

Phagocytic cells contribute to the antibody-mediated elimination of pulmonary-infected SARS coronavirus

SARS-CoV-2 variants show resistance to neutralization by many monoclonal and serum-derived polyclonal antibodies

Increased resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7 to antibody neutralization

Resistance of SARS-CoV-2 variants to neutralization by monoclonal and serum-derived polyclonal antibodies

Dissecting strategies to tune the therapeutic potential of SARS-CoV-2-specific monoclonal antibody CR3022

Neutralizing and protective human monoclonal antibodies recognizing the N-terminal domain of the SARS-CoV-2 spike protein

Ultrapotent human antibodies protect against SARS-CoV-2 challenge via multiple mechanisms

Human neutralizing antibodies against SARS-CoV-2 require intact Fc effector functions for optimal therapeutic protection

Collaboration between the Fab and Fc contribute to maximal protection against SARS-CoV-2 following NVX-CoV2373 subunit vaccine with Matrix-M vaccination

Vaccination with SARS-CoV-2 spike protein and AS03 adjuvant induces rapid anamnestic antibodies in the lung and protects against virus challenge in nonhuman primates

Distinct features and functions of systemic and mucosal humoral immunity among SARS-CoV-2 convalescent individuals

Optimization and qualification of an Fc Array assay for assessments of antibodies against HIV-1/SIV

Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation

Antibody responses to SARS-CoV-2 in patients with COVID-19

A perspective on potential antibody-dependent enhancement of SARS-CoV-2

Antibody dependent enhancement due to original antigenic sin and the development of SARS

The role of SARS-CoV-2 antibodies in COVID-19: healing in most, harm at times

Complement activation contributes to severe acute respiratory syndrome coronavirus pathogenesis

Complement associated microvascular injury and thrombosis in the pathogenesis of severe COVID-19 infection: a report of five cases

The case of complement activation in COVID-19 multiorgan impact

UMAP: Uniform Manifold Approximation and Projection for dimension reduction

Role of IgM and IgA antibodies in the neutralization of SARS-CoV-2

IgA dominates the early neutralizing antibody response to SARS-CoV-2

Antibody-dependent cell-mediated cytotoxicity and spontaneous cell-mediated cytotoxicity against cells infected with porcine transmissible gastroenteritis virus

Antibody dependent cellular cytotoxicity against coronavirus 229E-infected cells

SARS-coronavirus replication in human peripheral monocytes/macrophages

Neutralizing antibodies against Mayaro virus require Fc effector functions for protective activity

Optimal therapeutic activity of monoclonal antibodies against chikungunya virus requires Fc-FcgammaR interaction on monocytes

Survivors of Ebola virus disease develop polyfunctional antibody responses

A role for Fc function in therapeutic monoclonal antibody-mediated protection against Ebola virus

Both neutralizing and non-neutralizing human H7N9 influenza vaccineinduced monoclonal antibodies confer protection

Non-neutralizing antibodies from a Marburg infection survivor mediate protection by Fc-effector functions and by enhancing efficacy of other antibodies

Cytokine and chemokine levels in coronavirus disease 2019 convalescent plasma

Purified coronavirus spike protein nanoparticles induce coronavirus neutralizing antibodies in mice

Neutralizing monoclonal antibodies as promising therapeutics against Middle East respiratory syndrome coronavirus infection

A humanized neutralizing antibody against MERS-CoV targeting the receptor-binding domain of the spike protein

COVA1-18 neutralizing antibody protects against SARS-CoV-2 in three preclinical models

Broad and potent activity against SARS-like viruses by an engineered human monoclonal antibody

Afucosylated IgG characterizes enveloped viral responses and correlates with COVID-19 severity

The functions of SARS-CoV-2 neutralizing and infection-enhancing antibodies in vitro and in mice and nonhuman primates

Highly parallel characterization of IgG Fc binding interactions

A highly conserved cryptic epitope in the receptor binding domains of SARS-CoV-2 and SARS-CoV

Multiplexed Fc array for evaluation of antigen-specific antibody effector profiles

An immunosuppressed Syrian golden hamster model for SARS-CoV infection

Functional assessment of cell entry and receptor usage for SARS-CoV-2 and other lineage B betacoronaviruses

Determining the phagocytic activity of clinical antibody samples

A robust, high-throughput assay to determine the phagocytic activity of clinical antibody samples

A versatile high-throughput assay to characterize antibody-mediated neutrophil phagocytosis

pheatmap: pretty heatmaps, vR package version 1

R package 'corrplot': visualization of a correlation matrix

ggplot2: elegant graphics for data analysis

umap: Uniform Manifold Approximation and Projection

Dbscan: fast density-based clustering with R

Visualizing the performance of scoring classifiers

Display and analyze ROC curves