Comparison Study of the Bio-Plex and Meso Scale Multiplexed SARS-CoV-2 Serology Assays Reveals Evidence of Diminished Host Antibody Responses to SARS-CoV-2 after Monoclonal Antibody Treatment

Abstract

Background: Assessing the breadth and duration of antigen-specific binding antibodies provides valuable information for evaluating interventions to treat or prevent SARS-CoV-2 infection. Multiplex immunoassays are a convenient method for rapid measurement of antibody responses but can sometimes provide discordant results, and antibody positive percent agreement for COVID-19 diagnosis can vary depending on assay type, disease severity, and population sampled. Therefore, we compared two assays marked for research applications, MSD and Bio-Plex Pro, to evaluate qualitative interpretation of serostatus and quantitative detection of antibodies of varying isotypes (IgG, IgM, and IgA) against receptor binding domain (RBD) and nucleocapsid (N) antigens.

Methods: Specimens from ACTIV-2/A5401, a placebo-controlled clinical trial of the SARSCoV-2 monoclonal antibody (mAb) bamlanivimab to prevent COVID-19 disease progression, were used to evaluate the concordance of the Bio-Rad Bio-Plex Pro Human SARS-CoV-2 Serology Assay and the Meso Scale Discovery (MSD) V-PLEX COVID-19 Panel 1 serology assay in detecting and quantifying IgG, IgA, and IgM binding anti-SARS-CoV-2 antibody responses against the RBD and N antigens. Data were disaggregated by study arm, bamlanivimab dose, days post-enrollment, and presence of emerging resistance.

Results: We observed 90.5% (412 of 455 tests) concordance for anti-RBD IgG and 87% (396 of 455) concordance for anti-N IgG in classifying samples as negative or positive based on assay-defined cutoffs. Antibody levels converted to the WHO standard BAU/mL were significantly correlated for all isotypes (IgG, IgM, and IgA) and SARS-CoV-2 antigen targets (RBD and N) tested that were common between the two assays (Spearman r 0.65 to 0.92, P < 0.0001). Both assays uncovered evidence of diminished host-derived IgG immune responses in participants treated with bamlanivimab compared to placebo. Assessment of immune responses in the four individuals treated with the 700 mg of bamlanivimab with emerging mAb resistance demonstrated a stronger anti-N IgG response (MSD) at day 28 (median 2.18 log BAU/mL) compared to participants treated with bamlanivimab who did not develop resistance (median 1.55 log BAU/mL).

Conclusions: These data demonstrate the utility in using multiplex immunoassays for characterizing the immune responses with and without treatment in a study population and provide evidence that monoclonal antibody treatment in acute COVID-19 may have a modest negative impact on development of host IgG responses.

Keywords: Bio-Plex; COVID; COVID-19; Meso Scale Discovery; SARS-CoV-2; bamlanivimab; monoclonal antibodies.

Figure 1.

Comparison of total IgG, IgM, and IgA antibody response to the SARS-CoV-2 RBD and N proteins using Bio-Plex and Meso Scale assays in ACTIV-2/A5401 participants at study entry (pre-treatment, within 10 days of COVID-19 symptom onset). For all graphs, log Binding Antibody Units per milliliter of serum (BAU/mL) were plotted for Bio-Plex (y-axis) against Meso Scale (x-axis). For both assays, BAU/mL was calculated by applying the conversion factor to the concentration values for each assay; only values with detectable MFI or signal in both assays are plotted on this graph. Paired values where one or both MFI or signal were undetectable and did not produce a quantifiable MFI or signal value are excluded from the correlation. The Spearman Rank Correlation was calculated for each Bio-Plex/MSD pair using GraphPad Prism version 9.5.0 with R value noted in each graph. The P-value for all xy pairs was < 0.0001, denoted by (****) on each graph. (A) Anti-RBD IgG, N = 222 pairs; (B) Anti-N IgG, N = 220 pairs; (C) Anti-RBD IgA, N = 267 pairs; (D) Anti-N IgA, N = 239 pairs; (E) Anti-RBD IgM, N = 268 pairs; and (F) anti-N IgM, N = 268 pairs.

Figure 2.

Total antibody response to various SARS-CoV-2 antigens over time in individuals receiving bamlanivimab and placebo. Data generated using the Meso Scale Discovery V-PLEX COVID-19 Panel 1 serology assay (MSD) are in panels (A), (C), (E), and (G). Data generated using the Bio-Rad Bio-Plex Pro Human SARS-CoV-2 serology assay are in panels (B), (D), (F), and (H). For all graphs, log₁₀ Binding Antibody Units per milliliter of serum (BAU/mL) were plotted for each A5401 study arm as follows: 700 mg bamlanivimab (light orange circles ); 700 mg placebo (light blue circles ); 7,000 mg bamlanivimab (dark orange diamonds ); 7,000 mg placebo (dark blue diamonds ). Each arm is plotted from samples collected at baseline (day 0, within 10 days of symptom onset), day 28, and week 12 post-enrollment. For Bio-Plex, seronegative samples whose BAU/mL was under the limit of quantitation could not be plotted thus were omitted from the graph. The median log BAU/mL for each data set is indicated by a solid line. The number of samples plotted is noted under the x-axis. The dashed horizontal line from the y-axis denotes the log BAU/mL value under which the sample is classified as IgG negative; these values are set at (A) 1.17 (anti-RBD IgG MSD); (B) 0.83 (anti-RBD IgG Bio-Plex); (C) 1.25 (anti-S IgG MSD); (D) 0.74 (anti-S1 IgG Bio-Plex); (E) 1.07 (anti-N IgG MSD); (F) 0.56 (anti-N IgG Bio-Plex); (G) no qualitative cutoff is available for anti-NTD IgG MSD; and (H) 0.72 (anti-S2 IgG Bio-Plex). Mann-Whitney test was used to compare unpaired ranked log BAU/mL for each arm (700 mg or 7,000 mg bamlanivimab) against its corresponding placebo arm for each time point of collection. Comparisons are denoted as follows: not significant (ns); P ≤ 0.05 (*); P ≤ 0.01 (**); P ≤ 0.001 (***) and P ≤ 0.0001 (****).

Figure 3.

Area Under the Curve (AUC) of Day 28 Anti-N IgG against SARS-CoV-2 Viral Load. Blue dots represent 700 mg placebo arm samples, and orange dots represent 700 mg bamlanivimab samples. Data generated using the Meso Scale Discovery V-PLEX COVID-19 Panel 1 serology assay (MSD) in (A) and the Bio-Rad Bio-Plex Pro Human SARS-CoV-2 serology assay in (B). X-axis shows nasoparyngeal (NP) log₁₀ SARS-CoV-2 RNA copies/ml area under curve (AUC). Only P-values < 0.05 are considered significant and are denoted by P, while Spearman's rho is indicated by ρ.