COVID-19 has had an unprecedented global impact on human health. Understanding the Ab memory responses to infection is one tool needed to effectively control the pandemic. Among 173 outpatients who had virologically confirmed SARS-CoV-2 infection, we evaluated serum Ab concentrations, microneutralization activity, and enumerated SARS-CoV-2–specific B cells in convalescent human blood specimens. Serum Ab concentrations were variable, allowing for stratification of the cohort into high and low responders. Neither participant sex, the timing of blood sampling following the onset of illness, nor the number of SARS-CoV-2 spike protein–specific B cells correlated with serum Ab concentration. Serum Ab concentration was positively associated with microneutralization activity and participant age, with participants under the age of 30 showing the lowest Ab level. These data suggest that young adult outpatients did not generate as robust Ab memory, compared with older adults. Body mass index was also positively correlated with serum Ab levels. Multivariate analyses showed that participant age and body mass index were independently associated with Ab levels. These findings have direct implications for public health policy and current vaccine efforts. Knowledge gained regarding Ab memory following infection will inform the need for vaccination in those previously infected and allow for a better approximation of population-wide protective immunity.

The ongoing COVID-19 pandemic caused by SARS-CoV-2 has resulted in >250 million cases and 5 million deaths worldwide as of December 2021. With a case fatality rate near 2%, most of those infected survive and go on to generate immune memory against the virus. Numerous studies have shown that >95% of those infected generate Abs that recognize SARS-CoV-2 proteins in the months immediately following infection (as reviewed in Ref. 1). SARS-CoV-2 infection results in neutralizing Ab production in most of those infected, although the half-life of neutralizing Abs may be a relatively short 2 mo (1). Neutralizing Ab production is thought to be protective against reinfection in ∼90% of individuals (1). Many studies to date have focused on those hospitalized with severe COVID-19, although several groups have published investigations into the Ab response to SARS-CoV-2 infection in those with less severe disease (Refs. 27 and F. González, O. Zepeda, C. Toval-Ruiz, A. Matute, H. Vanegas, N. Munguia, E. Centeno, Y. Reyes, L. Svensson, J. Nordgren, et al., manuscript posted on medRxiv, DOI: 10.1101/2021.04.28.21256122). Comparison of the Ab response in patients with severe disease with those with mild/moderate disease or the response induced by vaccination have shown that severe illness results in robust and greater Ab levels than the other settings (8, 9).

Although SARS-CoV-2 infection has been shown to produce a significant Ab response that initially protects against reinfection, it has been shown to wane over the course of months. This waning immune memory suggests a need for immune boosting by vaccination (8). Furthermore, the impact of age, sex, and body mass index (BMI) on Ab response to SARS-CoV-2 is inconsistent in the literature and may play a key role in defining which patients are most in need of boosting for sustaining adequate immune memory (as reviewed in Ref. 10). The goal of the current study was to better understand the Ab response to SARS-CoV-2 infection in adults with mild to moderate illness who sought outpatient care using well-characterized and controlled assays for SARS-CoV-2 spike-specific (S1, receptor-binding domain [RBD], S2 domains) and nucleoprotein (N)-specific IgG Abs. SARS-CoV-2 Ab concentrations were correlated with in vitro virus neutralization activity to demonstrate the functional relevance of high and low Ab concentration responders. In addition, the presence of IgG-producing memory B cells was examined by flow cytometry in participants with high serum Ab responses to SARS-CoV-2.

All participant samples were collected as part of the U.S. Influenza Vaccine Effectiveness network (Pittsburgh site). This network is an annual prospective study of outpatients who seek care for acute respiratory illness. This study was approved by the University of Pittsburgh Institutional Review Board. All 173 participants were adults (≥18 y of age) and were SARS-CoV-2 positive by molecular testing. Relevant participant information was subsequently extracted from the medical record. One participant did not disclose sex, one participant did not have the timing of blood sampling relative to illness recorded, and two participants did not have BMI recorded. These participants were excluded for the relevant analyses. Participant age groupings were selected based on the observation of raw Ab data, which revealed lower levels in participants between 18 and 29 y of age; additional groups were then established using 15-y intervals.

Blood was collected in Becton Dickinson serum separator tubes (SSTs; for serum) or cell preparation tubes (CPTs; for PBMCs) and processed following the manufacturer’s instructions. Serum was aliquoted and stored at −80°C prior to analyses. PBMCs were aliquoted, frozen, and stored in liquid nitrogen prior to analyses.

SARS-CoV-2–specific Ab concentrations were determined using a commercially available Bio-Rad Bio-Plex kit (Bio-Plex Pro human IgG SARS-CoV-2 N/RBD/S1/S2 four-plex panel; catalog no. 12014634). Analyses of serum samples were performed per the manufacturer’s instructions. Banked pre–COVID-19 serum samples were used as negative controls and were all below the limit of detection (n = 5). SARS-CoV-2 Ab standards were provided by Bio-Rad, with concentrations of U/ml. To control for plate-to-plate variability, five samples were repeated on all three assay plates run. Concordance was very high between assay runs (r2 of 0.85–0.96), suggesting that combining data was appropriate (Supplemental Fig. 1A–C). In addition, World Health Organization SARS-CoV-2 Ab reference samples were assayed to compare with values obtained using the Bio-Plex kit. The COVID-19 convalescent plasma panel (NIBSC 20/118) was obtained from the National Institute for Biological Standards and Control, United Kingdom. Concentrations derived using the Bio-Plex kit correlated with the reported World Health Organization reference sample concentrations (Supplemental Fig. 1D, 1E).

Ab neutralization of SARS-CoV-2 was performed as published (11). Spots in Vero E6 cell cultures were imaged, counted, and processed using an ImmunoSpot Counter (CTL). Foci were counted in experimental wells and compared with control wells. The dilution of serum at which 80% of foci are neutralized is reported as the FRNT80.

Frozen PBMCs were rapidly thawed in a water bath at 37°C, rinsed with 1 ml of flow sorter buffer (1× PBS, 2% newborn calf serum, and 0.1% sodium azide) and centrifuged at 500 × g for 5 min, 4°C. Cells were resuspended in 100 μl of sorter buffer with 2.5 μg of human Fc Block (clone Fc.3216) for 10 min at room temperature. Next, cells were incubated with 1 mM PE-Cy5 decoy for 5 min at room temperature followed by 1 mM SARS-CoV-2 spike-PE B cell tetramer for 25 min on ice in the dark using decoy and tetramer created in the laboratory as previously described (12, 13). Cells were then washed and incubated with anti-PE magnetic beads (Miltenyi Biotec) and passed over a LS column on a QuadroMACS magnet. Eluted cells were stained for 30 min at 4°C in the dark with the following anti-human Abs (clone): CD19 (SJ25C1), CD3 (UCHT1), CD14 (HCD14), CD16 (B73.1), CD21 (Bu32), CD27 (M-T271), IgD (IA6-2), IgM (MHM-88), and IgG (G18-145). Next, cells were washed with 1 ml of flow sorter buffer, centrifuged at 500 × g for 5 min, and fixed with 200 μl of 2% paraformaldehyde at 4°C for 10 min. Cells were washed and resuspended in flow sorter buffer and collected on an LSRFortessa (Becton Dickson). Data were analyzed using FlowJo software (Tree Star, Ashland, OR). As a negative control, four banked prepandemic PBMC samples were analyzed.

Statistical analyses were performed using GraphPad Prism software (GraphPad Software, San Diego, CA). Correlation data were analyzed by simple linear regression with an α of 0.05 for significance. Differences between two means were analyzed by an unpaired t test with an α of 0.05 for significance. Differences between more than two means were analyzed by one-way ANOVA followed by a Tukey post hoc test with an α of 0.05 for significance. Cohort datasets are shown as violin plots with the median and quartiles indicated. Alternatively, datasets are displayed as dot plots with mean and SEM displayed. Multivariable linear regression models were run separately for each SARS-CoV-2 spike and nucleoprotein specific IgG Abs and assessed the association of age on the SARS-CoV-2 Ab response adjusting for BMI, and vice versa. Statistical significance of two-sided tests was set at type I error (α) equal to 0.05. These analytical procedures were performed using SAS 9.4 (SAS, Cary, NC).

To examine the Ab response to mild SARS-CoV-2 infection, we recruited 173 convalescent outpatients for blood sampling through the U.S. Influenza Vaccine Effectiveness network. The patient cohort included 105 females with a mean age of 38.3 y (range, 19–79 y) and 67 males with a mean age of 42.7 y (range, 20–78 y). One participant did not disclose his or her sex; this patient was excluded from birth sex analyses. The convalescent blood sample was drawn at an average of 44.5 d (range, 22–131 d) after symptom onset for females and 42.3 d (range, 21–89 d) for males. BMI was a mean 30.2 for females (range, 17.9–59.8) and a mean of 30.9 for males (range, 19.3–72.0). There were no significant differences in age, timing of sampling, or BMI by sex of participants.

SARS-CoV-2 Abs were detectable in most participant samples. Concentrations ranged from 0 to 2 × 106 U/ml. Low responders were defined as Ab concentration <5 × 104 U/ml. Approximately one-fifth of participants had low Ab levels against N, RBD, and S1 proteins (Table I). The incidence of low response to S2 was lower than the other Ags tested at 6.9% of participants. Approximately 6% of participants had very low Ab responses (<104 U/ml) to N, RBD, and S1, whereas only a few had very low Ab levels against S2. Of the 18 persons with the lowest Ab concentrations, 13 were females. Of these 18, low responses were noted against one of four Ags in 6 persons, two Ags in 7 persons, three Ags in 3 persons, and all four Ags in 2 persons.

Table I.

Outpatients with lower Ab concentration against SARS-CoV-2

AbFemale (n = 105)Male (n = 67)Total (n = 173)
   
 <104 7 (6.7%) 4 (6.0%) 11 (6.4%) 
 <5 × 104 25 (23.8%) 13 (19.4%) 39 (22.5%) 
RBD    
 <104 9 (8.6%) 2 (3.0%) 11 (6.4%) 
 <5 × 104 26 (24.8%) 12 (17.9%) 39 (22.5%) 
S1    
 <104 9 (8.6%) 3 (4.5%) 12 (6.9%) 
 <5 × 104 26 (24.8%) 12 (17.9%) 39 (22.5%) 
S2    
 <104 2 (1.9%) 1 (1.5%) 3 (1.7%) 
 <5 × 104 5 (4.8%) 6 (9.0%) 12 (6.9%) 
AbFemale (n = 105)Male (n = 67)Total (n = 173)
   
 <104 7 (6.7%) 4 (6.0%) 11 (6.4%) 
 <5 × 104 25 (23.8%) 13 (19.4%) 39 (22.5%) 
RBD    
 <104 9 (8.6%) 2 (3.0%) 11 (6.4%) 
 <5 × 104 26 (24.8%) 12 (17.9%) 39 (22.5%) 
S1    
 <104 9 (8.6%) 3 (4.5%) 12 (6.9%) 
 <5 × 104 26 (24.8%) 12 (17.9%) 39 (22.5%) 
S2    
 <104 2 (1.9%) 1 (1.5%) 3 (1.7%) 
 <5 × 104 5 (4.8%) 6 (9.0%) 12 (6.9%) 

The numbers of patients with Ab concentrations <104 U/ml or <5 × 104 U/ml are shown.

We examined correlations between Ab levels for pairs of SARS-CoV-2 proteins. The highest correlation (R2 = 0.74) in Ab levels was for S1 and RBD, which would be predicted since RBD is part of S1 (Fig. 1). Significant correlations between Ag Ab levels were found for all four Ags tested. Interestingly, N protein Ab levels correlated at the lowest levels in all comparisons. Despite statistically significant correlations, the relationship between individual Ag Ab levels were not overtly linear, as was the case for S1 and RBD.

FIGURE 1.

Correlation between SARS-CoV-2 Ag-specific IgG concentrations. Samples in the cohort were analyzed for four SARS-CoV-2 Ags (n = 173). Concentration values (U/ml) were compared using simple linear regression. R2 values are denoted, p ≤ 0.0001 for all comparisons. Red depicts the highest correlation (1.0), and green indicates the lowest correlation (0), with color gradients in between.

FIGURE 1.

Correlation between SARS-CoV-2 Ag-specific IgG concentrations. Samples in the cohort were analyzed for four SARS-CoV-2 Ags (n = 173). Concentration values (U/ml) were compared using simple linear regression. R2 values are denoted, p ≤ 0.0001 for all comparisons. Red depicts the highest correlation (1.0), and green indicates the lowest correlation (0), with color gradients in between.

Close modal

To determine whether measurable differences in the SARS-CoV-2 Ab concentrations were relevant to functional Ab, microneutralization assays were performed on a subset of participants with the most divergent Ab levels. This subset of 56 participants was selected to assess both high and low responders: based on high Ab response (>7 × 105 U/ml for at least one Ag) and low Ab response (<1 × 105 U/ml for at least one Ag). Evidence of neutralizing Ab activity (FRNT80) was uncommon in the low responder group with only 9 of 28 participants achieving a neutralizing titer ≥1:20 (Fig. 2). No high-titer neutralizing activity was seen in the low responder group (greatest observed 1:80). Conversely, all but one high responder displayed neutralizing activity, with 18 of 28 showing a neutralizing titer ≥1:160. Furthermore, the microneutralization titer significantly correlated with the concentration of Ab determined by Bio-Plex (Supplemental Fig. 2). Correlation was strongest between microneutralization titer and spike protein Ags (e.g., r2 of 0.36 for RBD).

FIGURE 2.

In vitro SARS-CoV-2 microneutralization titers in high and low responding participants. (A and B) A subset of participants was stratified by relative Ab response into two groups of lowest (A) and highest (B) Ab concentrations (n = 28 each). The dilution of serum at which 80% of viral foci are neutralized is reported as the FRNT80 (1:x titer). The limit of detection for the assay is 1:20.

FIGURE 2.

In vitro SARS-CoV-2 microneutralization titers in high and low responding participants. (A and B) A subset of participants was stratified by relative Ab response into two groups of lowest (A) and highest (B) Ab concentrations (n = 28 each). The dilution of serum at which 80% of viral foci are neutralized is reported as the FRNT80 (1:x titer). The limit of detection for the assay is 1:20.

Close modal

Flow cytometry was then used to determine whether SARS-CoV-2–specific B, memory B, and IgG-producing B cells were detectable in PBMCs from high responders. We employed a SARS-CoV-2 spike tetramer and a gating strategy to differentiate total spike-specific B cells, memory B cells, and class-switched IgG-producing B cells (Supplemental Fig. 3). We successfully detected spike-specific B cells in all three gates from all high responders. The number of spike-specific B cells did not correlate with microneutralization titer, nor did memory B cell or IgG-producing B cell numbers (Fig. 3). No differences were seen in the number of spike-specific B cells, spike-specific memory, or IgG-producing B cells by patient sex (Fig. 4). Participant age was then compared with the number of spike-specific B cells; again, no significant correlations were observed (Fig. 4). There was no correlation between B cell numbers and the timing of sampling after illness (spike-specific B cells R2 = 0.0406, p = 0.3814; spike-specific memory B cells R2 = 0.0054, p = 0.7518; IgG-producing B cells R2 = 0.0192, p = 0.5497).

FIGURE 3.

Correlation of spike-specific B cells and serum microneutralization activity. Flow cytometry was performed on high responder participant PBMCs using a spike-specific tetramer and B cell markers (n = 21). (A) Spike B cells were defined as CD19+tetramer+ cells. (B) Spike memory B cells were defined as CD19+tetramer+CD27+ cells. (C) Spike-specific IgG B cells were defined as CD19+tetramer+IgDIgMIgG+ cells. Data were compared using simple linear regression.

FIGURE 3.

Correlation of spike-specific B cells and serum microneutralization activity. Flow cytometry was performed on high responder participant PBMCs using a spike-specific tetramer and B cell markers (n = 21). (A) Spike B cells were defined as CD19+tetramer+ cells. (B) Spike memory B cells were defined as CD19+tetramer+CD27+ cells. (C) Spike-specific IgG B cells were defined as CD19+tetramer+IgDIgMIgG+ cells. Data were compared using simple linear regression.

Close modal
FIGURE 4.

Effect of biologic sex or participant age on SARS-CoV-2 spike-specific B memory and IgG-producing B cells. Flow cytometry was performed on high responder participant PBMCs using a spike-specific tetramer and B cell markers (n = 21). (A and D) Spike B cells were defined as CD19+tetramer+ cells. (B and E) Spike memory B cells were defined as CD19+tetramer+CD27+ cells. (C and F) Spike-specific IgG B cells were defined as CD19+tetramer+IgDIgMIgG+ cells. Data were compared using an unpaired t test or simple linear regression.

FIGURE 4.

Effect of biologic sex or participant age on SARS-CoV-2 spike-specific B memory and IgG-producing B cells. Flow cytometry was performed on high responder participant PBMCs using a spike-specific tetramer and B cell markers (n = 21). (A and D) Spike B cells were defined as CD19+tetramer+ cells. (B and E) Spike memory B cells were defined as CD19+tetramer+CD27+ cells. (C and F) Spike-specific IgG B cells were defined as CD19+tetramer+IgDIgMIgG+ cells. Data were compared using an unpaired t test or simple linear regression.

Close modal

We examined the impact of sample timing and/or biologic variables on SARS-CoV-2 Ab responses in the entire cohort. The mean sampling time was ∼6 wk after the onset of illness, which was used to divide the group into those sampled before and after 42 d post-illness onset. There was no significant difference in Ab levels by time of sampling for any of the four Ags tested (Fig. 5). Furthermore, there was no significant correlation between sample timing and Ab concentrations against N, RBD, and S1 Ags (Supplemental Fig. 4A–D). S2 Ab levels were positively correlated with longer timing of sampling after illness. There was no difference in Ab concentration against the SARS-CoV-2 Ags tested when analyzed by patient sex (Fig. 6).

FIGURE 5.

The impact of sample timing on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by when blood samples were collected after the onset of COVID-19 symptoms with a split at 42 d (n = 97 ≤42 d, 75 >42 d). Ab concentrations were determined by a Bio-Plex assay and compared using an unpaired t test.

FIGURE 5.

The impact of sample timing on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by when blood samples were collected after the onset of COVID-19 symptoms with a split at 42 d (n = 97 ≤42 d, 75 >42 d). Ab concentrations were determined by a Bio-Plex assay and compared using an unpaired t test.

Close modal
FIGURE 6.

The impact of biologic sex on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by sex (n = 105 female, 67 male). Ab concentrations were determined by a Bio-Plex assay and compared using an unpaired t test.

FIGURE 6.

The impact of biologic sex on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by sex (n = 105 female, 67 male). Ab concentrations were determined by a Bio-Plex assay and compared using an unpaired t test.

Close modal

We broke the cohort into four groups of 18–29, 30–45, 46–59, and ≥60 y. When examined by age groups (18–29, 30–45, 46–59, and over ≥60 y), Abs against SARS-CoV-2 N, RBD, and S1 were significantly lower in participants under the age of 30 y compared with participants over age 45 y (Fig. 7). Furthermore, patient age was significantly positively correlated with SARS-CoV-2 Ab concentration (Supplemental Fig. 4E–H). However, there was no correlation between older age and serum-neutralizing activity. Although all participants were recruited as outpatients, we attempted to characterize the severity of disease by measuring the number of days until recovery and whether the patient returned to normal activity at the time of convalescent sampling. The mean days to recovery was significantly longer in participants >45 y of age compared with those in the 18–29 and 30–45 groups (Table II). Similarly, a greater proportion of participants had not yet returned to normal activity in the >45 group compared with the two young participant groups.

FIGURE 7.

The impact of participant age on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by age (n = 58 for age 18–29 y, 55 for age 30–45 y, 36 for age 46–59, 24 for ≥60 y of age). Ab concentrations were determined by a Bio-Plex assay and compared using a one-way ANOVA followed by a Tukey post hoc test. (E) Serum neutralization activity and age were compared using simple linear regression.

FIGURE 7.

The impact of participant age on SARS-Cov-2 Ab concentration. (AD) Participants were stratified by age (n = 58 for age 18–29 y, 55 for age 30–45 y, 36 for age 46–59, 24 for ≥60 y of age). Ab concentrations were determined by a Bio-Plex assay and compared using a one-way ANOVA followed by a Tukey post hoc test. (E) Serum neutralization activity and age were compared using simple linear regression.

Close modal
Table II.

Recovery from SARS-CoV-2 symptoms by patient age

18–29 y30–45 y46–59 y0+ y
Full recovery (d) 11.1 ± 0.6 (n = 51) 12.7 ± 0.7 (n = 43) 15.4 ± 1.9 (n = 19)* 16.2 ± 1.8 (n = 17)** 
Return to normal (no) 9/57 15/52 16/32** 12/24* 
18–29 y30–45 y46–59 y0+ y
Full recovery (d) 11.1 ± 0.6 (n = 51) 12.7 ± 0.7 (n = 43) 15.4 ± 1.9 (n = 19)* 16.2 ± 1.8 (n = 17)** 
Return to normal (no) 9/57 15/52 16/32** 12/24* 

The number of days to full recovery and whether the patient had returned to normal activity by the time of sampling are shown.

*

p < 0.05,

**

p < 0.01 when compared with the 18–29 group.

Finally, we compared participant BMI with serum Ab levels. BMI was significantly positively correlated with Ab concentrations for all four Ags tested (Fig. 8). We next examined the relationship between patient age and BMI since both parameters correlated with Ab levels. BMI was not significantly associated with patient age (R2 = 0.0115, p = 0.1624). To test for interaction between these variables, multivariable regression was performed to determine the association of age upon SARS-CoV-2 Ab response when adjusting for BMI. In adjusted models, age was significantly related to SARS-CoV-2 Ab response (Table III). BMI adjusted for age was also significantly associated with SARS-CoV-2 Abs. There was no evidence of an interaction between age and BMI for any SARS-CoV-2 Ab outcome.

FIGURE 8.

The association between BMI and SARS-CoV-2 Ab concentration. (AD) Simple linear regression was performed on participant BMI and Ab levels (n = 171). Ab concentrations were determined by a Bio-Plex assay.

FIGURE 8.

The association between BMI and SARS-CoV-2 Ab concentration. (AD) Simple linear regression was performed on participant BMI and Ab levels (n = 171). Ab concentrations were determined by a Bio-Plex assay.

Close modal
Table III.

Association of age and BMI on SARS-COV-2 Abs: multivariable linear regression

VariableβSEp Value
   
 Age 9,038.1094 2,691.2749 0.0010* 
 BMI 24,357.6913 4,870.0729 <0.0001* 
RBD    
 Age 8,318.1547 2,595.7786 0.0016* 
 BMI 11,804.7176 4,697.2647 0.0129* 
S1    
 Age 7,755.0474 2,480.7334 0.0021* 
 BMI 13,769.9620 4,489.0816 0.0025* 
S2    
 Age 6,922.26184 3,016.9755 0.0230* 
 BMI 10,904.20579 5,459.4536 0.0474* 
VariableβSEp Value
   
 Age 9,038.1094 2,691.2749 0.0010* 
 BMI 24,357.6913 4,870.0729 <0.0001* 
RBD    
 Age 8,318.1547 2,595.7786 0.0016* 
 BMI 11,804.7176 4,697.2647 0.0129* 
S1    
 Age 7,755.0474 2,480.7334 0.0021* 
 BMI 13,769.9620 4,489.0816 0.0025* 
S2    
 Age 6,922.26184 3,016.9755 0.0230* 
 BMI 10,904.20579 5,459.4536 0.0474* 

Linear regression equation: [Ag] = β0 + β1*age + β2*BMI + ε.

n = 171,

*

p < 0.05.

β0 is the intercept term and ε is the error term, β12 are the β terms for age and BMI.

This study focused on a cohort of COVID-19 outpatients with mild or moderate illness. Many published serology studies have focused on more severe disease, with mild illness limited to a comparator for the robust Ab response in hospitalized participants. More than 10 studies have shown that Ab response in severe COVID-19 is greater than in mild cases (as reviewed in Ref. 10). These findings have been extended by similar observations regarding the presence of neutralizing Abs in serum (10). Although the study of severe disease is important, most cases of COVID-19 do not require hospitalization and, thus, most individuals infected with SARS-CoV-2 would be expected to have Ab responses similar to those observed in our study. Our finding that 98.8% (171/173) of our outpatient cohort developed measurable Abs against at least one SARS-CoV-2 Ag is consistent with findings in many other studies that suggest that natural infection of variable severity induces humoral immunity against the virus (1).

Given our focus on COVID-19 outpatients, we were able to stratify these patients into groups of low and high responders to perform functional Ab assays and to examine the presence of SARS-CoV-2–specific memory B cells. Microneutralization assays revealed that relatively high and low Ab concentrations detected in the Bio-Plex assay correlated with the amount of functional Ab present. In the low responder group, the highest FRNT80 titer observed was 1:80, whereas nearly all high responders tested were equal to or greater than this level. Surprisingly, two participants had a high Ab concentration, but no or low microneutralizing activity, suggesting that Ab level alone may not be protective for some patients. Correlation of microneutralization titer with Ab concentration revealed a positive relationship between high spike protein Ab and functional ability, as one would predict given that spike is the target of neutralizing Abs. The correlation between neutralizing activity and N protein Ab was less strong. Others have shown that SARS-CoV-2–specific memory IgG-producing B cells are formed in both mild and severe patients (12, 14). In severe COVID-19 patients, memory B cells were found in all patients studied; however, only 80% had neutralizing Ab. In our study, all Ab high responders had detectable spike-specific memory B cells, although the number of memory B cells did not correlate with Ab concentration. All participants tested for memory B cells had serum-neutralizing activity of variable titer. These data suggest that high responding outpatients form memory B cells that may provide longer lasting protection against SARS-CoV-2 than serum Ab concentration would indicate.

By measuring IgG against four SARS-CoV-2 Ags, we were able to characterize the relative immunodominance of the different Ags. Ab response to N protein only partially correlates with spike protein Ab responses. Our data showed that some participants responded to N better than to spike proteins and vice versa, with the relationship between N and S1 Abs being the strongest positive correlation. The S1 and S2 Ab response was positively correlated, but not entirely linear, suggesting differential immune recognition of these spike protein domains among the cohort. Interestingly, the incidence of a low Ab response was lowest for the S2 domain of the spike protein when compared with S1, RBD, or N proteins. In the rare occurrence of very low Ab response to one or more Ags tested, females were more likely than males to display this poor response phenotype.

Our study did not show an impact of sample timing or biologic sex on serum Ab concentrations. Numerous studies have shown a degree of Ab waning at 2–3 mo postinfection. Our cohort spanned 1–4 mo after the onset of symptoms with a median near 42 d. In outpatients, we did not observe a decrease in serum Ab in this time period, nor did we see a negative correlation between sample timing after illness and Ab concentration. Several studies have shown no effect of biologic sex on serum Ab titer; however, neutralizing Ab levels have been shown to correlate with male sex (as reviewed in Refs. 10, 15, 16). This is consistent with our findings of Ab concentration; however, we did not see a significant difference in neutralizing Ab titer and sex (data not shown). The effect of sex on SARS-CoV-2 Ab response does not appear to be a primary driver of response in outpatients in this study.

Importantly, a predictor of Ab response in our study was participant age. Participants aged 18–29 y had significantly lower Ab concentration for N, S1, and RBD proteins compared with older participants. We also found a significant positive correlation between patient age and serum Ab levels. This is consistent with some studies that have shown a correlation between IgG levels and age, with lower Ab concentration in younger adults (15, 1719). In those studies, conflicting results were found regarding neutralizing Ab titers, as one study found an increase in older adults, whereas the others showed higher levels in younger adults. Other research did not show a relationship between age and Ab concentration (as reviewed in Refs. 10, 16). One possible explanation for the age effect observed would be the severity of disease in our cohort. To address this issue, we examined the number of days until recovery and whether a return to normal was seen at the time of sampling. We did show that recovery and a lack of return to normal activities was significantly longer in participants >45 y of age, compared with younger participants. However, these parameters were not different in participants 18–29 or those 30–45 y of age, although Ab concentration was lower between these two groups. These data suggest that measures of outpatient severity do not likely explain the lower Ab response in young participants in our study. It is possible that younger participants had better controlled viral replication, and thus there was less viral Ag to stimulate ongoing immune responses.

Finally, our study showed a significant positive correlation between BMI and serum Ab levels. The impact of obesity on COVID-19 remains unclear, as some have found a positive correlation between BMI and severity, whereas others have not (20, 21). Similarly, the relationship between BMI or metabolic syndrome and Ab levels has varied. Poor metabolic health (defined by a low adiponectin/leptin ratio) was shown to correlate with increased cross-reactive SARS-CoV-2 strain Abs (21). Furthermore, BMI and metabolic syndrome were shown to correlate with higher Ab levels (22, 23). However, others have shown either no or the opposite relationship (24, 25). In our study, multivariate analyses of age and BMI revealed that they are independent predictors of Ab response to SARS-CoV-2 in outpatients. Again, it is unclear why BMI may be associated with Ab levels, but differences in viral load, comorbidities, or subtle variability in disease severity may be involved.

In summary, we characterized serum Ab and B cell responses to SARS-CoV-2 in a cohort of COVID-19 outpatients. We did so using well-validated assays to generate reliable Ab concentration, microneutralization, and flow cytometry data. The primary findings were that participant age and BMI independently affected serum Ab concentration, with younger and lower BMI participants producing lower Ab responses. We also showed that Ab concentration was predictive of microneutralization activity, but not the number of spike-specific B cells in blood. In whole, these data inform current vaccination policy supporting that young and low BMI previously infected people likely require vaccine boosting for long-term protection. Furthermore, Ab concentration may not directly relate to the presence of SARS-CoV-2–specific memory B cells capable of providing protection against infection and/or severity of disease. As serum Ab levels wane following infection, memory B cells may persist longer to provide sustained protection. These questions will need to be answered at later time points following infection. In addition, our study was conducted prior to SARS-CoV-2 vaccine availability or the SARS-CoV-2 delta and omicron variant waves, which may differentially impact Ab responses.

This work was supported by Centers for Disease Control and Prevention (CDC) Grant 5U01IP001035 and National Institutes of Health (NIH) Grant UL1TR001857; a Commonwealth of Pennsylvania, Department of Human Services COVID grant; the DSF Charitable Foundation; and by the University of Pittsburgh Clinical and Translational Science Institute, National Institutes of Health Grant U01 AI124289. This work represents the views of the authors and not the CDC or NIH.

The online version of this article contains supplemental material.

K.C. enrolled participants and collected blood samples; B.Z. and L.J.S. processed blood samples for analyses; S.K. performed Bio-Plex assays; D.L.B. made B cell spike tetramers and performed flow cytometry assays; A.K.M. performed microneutralization assays; R.K.Z. and J.F.A. conceived the study; R.K.Z., J.F.A., A.K.M., and J.B.M. obtained funding; D.L.B., K.K.M.G., A.K.M., J.B.M., and J.F.A. analyzed the data; and M.P.N., K.K.M.G., A.K.M., J.B.M., R.K.Z., and J.F.A. drafted and edited the manuscript.

Abbreviations used in this article:

BMI

body mass index

FRNT80

dilution of serum at which 80% of foci are neutralized

N

nucleoprotein

RBD

receptor-binding domain

S

spike

1.
Cromer
D.
,
J. A.
Juno
,
D.
Khoury
,
A.
Reynaldi
,
A. K.
Wheatley
,
S. J.
Kent
,
M. P.
Davenport
.
2021
.
Prospects for durable immune control of SARS-CoV-2 and prevention of reinfection.
Nat. Rev. Immunol.
21
:
395
404
.
2.
Grossberg
A. N.
,
L. A.
Koza
,
A.
Ledreux
,
C.
Prusmack
,
H. K.
Krishnamurthy
,
V.
Jayaraman
,
A. C.
Granholm
,
D. A.
Linseman
.
2021
.
A multiplex chemiluminescent immunoassay for serological profiling of COVID-19-positive symptomatic and asymptomatic patients.
Nat. Commun.
12
:
740
.
3.
Peghin
M.
,
M.
De Martino
,
M.
Fabris
,
A.
Palese
,
E.
Visintini
,
E.
Graziano
,
V.
Gerussi
,
G.
Bontempo
,
F.
D’Aurizio
,
A.
Biasotto
, et al
2021
.
The fall in antibody response to SARS-CoV-2: a longitudinal study of asymptomatic to critically ill patients up to 10 months after recovery.
J. Clin. Microbiol.
59
:
e0113821
.
4.
Röltgen
K.
,
A. E.
Powell
,
O. F.
Wirz
,
B. A.
Stevens
,
C. A.
Hogan
,
J.
Najeeb
,
M.
Hunter
,
H.
Wang
,
M. K.
Sahoo
,
C.
Huang
, et al
2020
.
Defining the features and duration of antibody responses to SARS-CoV-2 infection associated with disease severity and outcome.
Sci. Immunol.
5
:
eabe0240
.
5.
Sasson
J. M.
,
J. J.
Campo
,
R. M.
Carpenter
,
M. K.
Young
,
A. Z.
Randall
,
K.
Trappl-Kimmons
,
A.
Oberai
,
C.
Hung
,
J.
Edgar
,
A. A.
Teng
, et al
2021
.
Diverse humoral immune responses in younger and older adult COVID-19 patients.
MBio
12
:
e0122921
.
6.
Strömer
A.
,
R.
Rose
,
O.
Grobe
,
F.
Neumann
,
H.
Fickenscher
,
T.
Lorentz
,
A.
Krumbholz
.
2020
.
Kinetics of nucleo- and spike protein-specific immunoglobulin G and of virus-neutralizing antibodies after SARS-CoV-2 infection.
Microorganisms
8
:
1572
.
7.
Wellinghausen
N.
,
M.
Voss
,
R.
Ivanova
,
S.
Deininger
.
2020
.
Evaluation of the SARS-CoV-2-IgG response in outpatients by five commercial immunoassays.
GMS Infect. Dis.
8
:
Doc22
.
8.
Anichini
G.
,
C.
Terrosi
,
C.
Gandolfo
,
G.
Gori Savellini
,
S.
Fabrizi
,
G. B.
Miceli
,
M. G.
Cusi
.
2021
.
SARS-CoV-2 antibody response in persons with past natural infection.
N. Engl. J. Med.
385
:
90
92
.
9.
Long
Q. X.
,
B. Z.
Liu
,
H. J.
Deng
,
G. C.
Wu
,
K.
Deng
,
Y. K.
Chen
,
P.
Liao
,
J. F.
Qiu
,
Y.
Lin
,
X. F.
Cai
, et al
2020
.
Antibody responses to SARS-CoV-2 in patients with COVID-19.
Nat. Med.
26
:
845
848
.
10.
Post
N.
,
D.
Eddy
,
C.
Huntley
,
M. C. I.
van Schalkwyk
,
M.
Shrotri
,
D.
Leeman
,
S.
Rigby
,
S. V.
Williams
,
W. H.
Bermingham
,
P.
Kellam
, et al
2020
.
Antibody response to SARS-CoV-2 infection in humans: a systematic review.
PLoS One
15
:
e0244126
.
11.
Xu
L.
,
J.
Doyle
,
D. J.
Barbeau
,
V.
Le Sage
,
A.
Wells
,
W. P.
Duprex
,
M. R.
Shurin
,
S. E.
Wheeler
,
A. K.
McElroy
.
2021
.
A cross-sectional study of SARS-CoV-2 seroprevalence between Fall 2020 and February 2021 in Allegheny County, Western Pennsylvania, USA.
Pathogens
10
:
710
.
12.
Rodda
L. B.
,
J.
Netland
,
L.
Shehata
,
K. B.
Pruner
,
P. A.
Morawski
,
C. D.
Thouvenel
,
K. K.
Takehara
,
J.
Eggenberger
,
E. A.
Hemann
,
H. R.
Waterman
, et al
2021
.
Functional SARS-CoV-2-specific immune memory persists after mild COVID-19.
Cell
184
:
169
183.e17
.
13.
Taylor
J. J.
,
R. J.
Martinez
,
P. J.
Titcombe
,
L. O.
Barsness
,
S. R.
Thomas
,
N.
Zhang
,
S. D.
Katzman
,
M. K.
Jenkins
,
D. L.
Mueller
.
2012
.
Deletion and anergy of polyclonal B cells specific for ubiquitous membrane-bound self-antigen.
J. Exp. Med.
209
:
2065
2077
.
14.
Varnaitė
R.
,
M.
García
,
H.
Glans
,
K. T.
Maleki
,
J. T.
Sandberg
,
J.
Tynell
,
W.
Christ
,
N.
Lagerqvist
,
H.
Asgeirsson
,
H. G.
Ljunggren
, et al
2020
.
Expansion of SARS-CoV-2-specific antibody-secreting cells and generation of neutralizing antibodies in hospitalized COVID-19 patients.
J. Immunol.
205
:
2437
2446
.
15.
Chvatal-Medina
M.
,
Y.
Mendez-Cortina
,
P. J.
Patiño
,
P. A.
Velilla
,
M. T.
Rugeles
.
2021
.
Antibody responses in COVID-19: a review.
Front. Immunol.
12
:
633184
.
16.
Li
L.
,
X.
Tong
,
H.
Chen
,
R.
He
,
Q.
Lv
,
R.
Yang
,
L.
Zhao
,
J.
Wang
,
H.
Xu
,
C.
Liu
, et al
2020
.
Characteristics and serological patterns of COVID-19 convalescent plasma donors: optimal donors and timing of donation.
Transfusion
60
:
1765
1772
.
17.
Yang
H. S.
,
V.
Costa
,
S. E.
Racine-Brzostek
,
K. P.
Acker
,
J.
Yee
,
Z.
Chen
,
M.
Karbaschi
,
R.
Zuk
,
S.
Rand
,
A.
Sukhu
, et al
2021
.
Association of age with SARS-CoV-2 antibody response.
JAMA Netw. Open
4
:
e214302
.
18.
Klein
S. L.
,
A.
Pekosz
,
H. S.
Park
,
R. L.
Ursin
,
J. R.
Shapiro
,
S. E.
Benner
,
K.
Littlefield
,
S.
Kumar
,
H. M.
Naik
,
M. J.
Betenbaugh
, et al
2020
.
Sex, age, and hospitalization drive antibody responses in a COVID-19 convalescent plasma donor population.
J. Clin. Invest.
130
:
6141
6150
.
19.
Wu
F.
,
M.
Liu
,
A.
Wang
,
L.
Lu
,
Q.
Wang
,
C.
Gu
,
J.
Chen
,
Y.
Wu
,
S.
Xia
,
Y.
Ling
, et al
2020
.
Evaluating the association of clinical characteristics with neutralizing antibody levels in patients who have recovered from mild COVID-19 in Shanghai, China.
JAMA Intern. Med.
180
:
1356
1362
.
20.
Nilles
E. J.
,
S. M.
Siddiqui
,
S.
Fischinger
,
Y. C.
Bartsch
,
M.
de St. Aubin
,
G.
Zhou
,
M. J.
Gluck
,
S.
Berger
,
J.
Rhee
,
E.
Petersen
, et al
2021
.
Epidemiological and immunological features of obesity and SARS-CoV-2.
Viruses
13
:
2235
.
21.
Tang
L.
,
S.
Cherry
,
E. I.
Tuomanen
,
E. K.
Roubidoux
,
C. Y.
Lin
,
K. J.
Allison
,
A.
Gowen
,
P.
Freiden
,
E. K.
Allen
,
Y.
Su
, et al
St. Jude Investigative Team
.
2021
.
Host predictors of broadly cross-reactive antibodies against SARS-CoV-2 variants of concern differ between infection and vaccination.
Clin. Infect. Dis.
DOI: 10.1093/cid/ciab996.
22.
Racine-Brzostek
S. E.
,
H. S.
Yang
,
G. A.
Jack
,
Z.
Chen
,
A.
Chadburn
,
T. J.
Ketas
,
E.
Francomano
,
P. J.
Klasse
,
J. P.
Moore
,
K. A.
McDonough
, et al
2021
.
Postconvalescent SARS-CoV-2 IgG and neutralizing antibodies are elevated in individuals with poor metabolic health.
J. Clin. Endocrinol. Metab.
106
:
e2025
e2034
.
23.
Soffer
S.
,
B. S.
Glicksberg
,
E.
Zimlichman
,
O.
Efros
,
M. A.
Levin
,
R.
Freeman
,
D. L.
Reich
,
E.
Klang
.
2021
.
The association between obesity and peak antibody titer response in COVID-19 infection.
Obesity (Silver Spring)
29
:
1547
1553
.
24.
Frasca
D.
,
L.
Reidy
,
C.
Cray
,
A.
Diaz
,
M.
Romero
,
K.
Kahl
,
B. B.
Blomberg
.
2021
.
Influence of obesity on serum levels of SARS-CoV-2-specific antibodies in COVID-19 patients.
PLoS One
16
:
e0245424
.
25.
Ozgocer
T.
,
S. N.
Dagli
,
M. R.
Ceylan
,
F.
Disli
,
C.
Ucar
,
S.
Yildiz
.
2021
.
Analysis of long-term antibody response in COVID-19 patients by symptoms grade, gender, age, BMI, and medication.
J. Med. Virol.
2021
:
jmv.27452
.

The authors have no financial conflicts of interest.