Associations of SARS-CoV-2 serum IgG with occupation and demographics of military personnel

Human subjects

Volunteers from the CT Army and Air National Guard were recruited to have their SARS-CoV-2 spike antigen-specific antibody levels tested. Subjects provided 3cc of blood by venipuncture using vacutainer tubes (Becton Dickinson; Franklin Lake, NJ); serum was separated and stored frozen at −80°C until tested by enzyme linked immunosorbent assay (ELISA). The studies were reviewed and ethical approval was given by the Yale University Institutional Review Board. All participants provided informed verbal consent.

ELISA methods

ELISAs were performed as previously described with minor modifications (11, 12). In short, Triton X-100 and RNase A were added to serum samples at final concentrations of 0.5% and 0.5mg/ml respectively and incubated at room temperature (RT) for 30 minutes before use to reduce risk from any potential virus in serum. 96-well MaxiSorp plates (ThermoFisher, Waltham, MA) were coated with 50 μL/well of recombinant SARS Cov-2 ectodomain (Sino Biologicals; Wayne, PA) at a concentration of 1 μg/ml in NaCO₃ buffer pH 9.6 and incubated overnight at 4°C. The coating buffer was removed, and plates were incubated for 1h at RT with 200 μL of blocking solution (PBS with 3% milk powder). Serum was diluted 1:100 in dilution solution (PBS with 0.05% Tween20, 1% milk powder) and 100 μL of diluted serum was added for one hours at RT. Plates were washed three times with PBS-T (PBS with 0.1% Tween-20) and 50 μL of HRP anti-Human IgG Antibody (Pharmingen/BD Biosiences, San Jose, CA) or HRP anti-human IgA (BioLegend, San Diego, CA) were added at 1:2000-fold dilution. After 1 h of incubation at RT, plates were washed three times with PBS-T. Plates were developed with 100 μL of TMB Substrate Reagent Set (BD Biosciences, San Jose, CA) and the reaction was stopped when an internal pooled serum positive control sample reaches an OD of 1.0 at 650 nm, by the addition of 2 N sulfuric acid. Plates were then read at a wavelength of 450 nm with reference wavelength calibration (650nm).

Cell lines and virus

VeroE6 kidney epithelial cells from ATTC were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 1% sodium pyruvate (NEAA) and 5% fetal bovine serum (FBS) at 37°C and 5% CO₂. SARS-CoV-2, strain USA-WA1/2020, was obtained from BEI Resources (#NR-52281) and amplified in VeroE6 cells. Cells were infected at a multiplicity of infection level of 0.01 for four three days to generate a working stock. Culture supernatant was clarified by centrifugation (450g × 5min) and filtered through a 0.45-micron filter. The pelleted virus was then resuspended in PBS then aliquoted for storage at −80°C. Viral titers were measured by standard plaque assay using Vero E6 cells. Briefly, 300 µL of serial fold virus dilutions were used to infect Vero E6 cells in MEM supplemented NaHCO₃, 4% FBS 0.6% Avicel RC-581. Plaques were resolved at 48h post infection by fixing in 10% formaldehyde for 1h followed by with 0.5% crystal violet in 20% ethanol staining. Plates were rinsed in water to plaques enumeration. All experiments were performed in a biosafety level 3 with the Yale Environmental Health and Safety office approval.

Neutralization assay

Study subject sera were heat treated for 30m at 56°C. Sixfold serially diluted plasma, from 1:3 to 1:2430 were incubated with SARS-CoV-2 for 1 h at 37°C. The mixture was subsequently incubated with Vero E6 cells in a 6-well plate for 1h. Then, cells were overlayed with MEM supplemented NaHCO₃, 4% FBS 0.6% Avicel mixture. Plaques were resolved at 40h post infection by fixing in 10% formaldehyde for 1h followed by staining in 0.5% crystal violet. All experiments were performed in parallel with negative controls sera, in a establish viral concentration to generate 120-150 plaques/well.

Statistical Methods

Descriptive statistics (means and standard deviations for continuous variables and percentages for categorical variables) of both predictor and outcome variables for the entire study population were calculated. Linear regression models were developed to model the loge of spike antigen IgG response level as the data had a log normal distribution. Unadjusted (bivariate) models were developed for each predictor. Additionally, a parsimonious model was determined utilizing multivariate regression modeling incorporating a backward selection strategy with significance level-to-stay specified as α=0.05. To evaluate whether the spike antigen IgG response level exceeded a specified threshold, logistic regression models we utilized. Again, unadjusted (bivariate) models were developed for each predictor. Additionally, a parsimonious model was determined utilizing multivariate regression modeling incorporating a backward selection strategy with significance level-to-stay specified as α=0.05. SAS v9.4® was utilized for all statistical analyses.

Basic Demographics

This study included N=995 soldiers enlisted in the Connecticut Army and Air National Guard (Table 1). The racial composition was 7.7% African-Americans, 2.5% Asian, 3.9% mixed, and 72.5% white individuals. Race data was not available for N=120 (12.1%) individuals who provided only information on ethnicity (Hispanic). The study group’s age distribution was slightly shifted to the left compared to the working civilian population in CT (13), and contained predominately younger individuals (Supplemental Materials Figure S1). Similar to the overall US military population the study group was largely (79.2%) male and healthy (14). None of the participants were hospitalized with COVID-19 like illness since the onset of the SARS-CoV-2 pandemic.

SARS-CoV-2 Serology Results and Demographic Associations

A total of N=103 (10.3%) of the subjects tested positive for SARS-CoV-2 spike antigen serum IgG, a rate 2.5-fold higher than the regional population at that time (15). Unadjusted logistic regression models (Table 1) demonstrated significant (p < 0.024) associations of SARS-CoV-2 IgG positivity with race and ethnicity, with notably increased odd ratios for black and Hispanic individuals (OR 2.1, 95% CI 1.2-4.1, and OR 1.7; 95% CI 1.1-2.8). Significant association of SARS-CoV-2 IgG positivity was also observed with household size ≥6 (p < 0.005) and known exposure to COVID-19 (p < 0.009), but not with specific lifestyle habits, tobacco use, vaping or alcohol consumption/frequency. In a parsimonious adjusted logistic regression model (Supplemental materials Table S1) SARS-CoV-2 IgG positivity remained significantly associated with race, household size and known exposure to COVID-19 (p < 0.017, 0.012, and 0.008 respectively).

Linear regression analysis of the data (supplemental Table S2) with SARS-CoV-2 IgG ELISA OD values quantitated as a continuous variable yielded stronger associations vs. logistic analysis with race, ethnicity, household size, and known COVID-19 exposure. Linear regression analysis further revealed a significant association of SARS-CoV-2 IgG levels with gender (p < 0.0002). While males were not more likely to be SARS-CoV-2 IgG positive than females, when they were positive, they had significantly (p<0.001) higher SARS-CoV-2 IgG ELISA OD values, on average by 0.25 OD units (Figure 1). Race, household size, known COVID-19 exposure, and gender remained significantly associated with SARS-CoV-2 IgG levels after parsimonious adjustment of linear regression models (Supplemental Materials Table S3).

• Download figure
• Open in new tab

Figure 1.

Spike antigen-specific serum IgG levels in males vs. females. The amount of spike antigen -specific IgG was quantitated based on ELISA OD values (Y-axis) and positive test results (OD > 0.356) are compared for males and females. * indicates significant (p < 0.023) difference between genders.

SARS-CoV-2 Serology Results and Occupational Associations

We further evaluated potential associations of SARS-CoV-2 serum IgG with exposures that might be connected to work, using military operation specialty (MOS) codes to define job title and related duties. Greater than 2/3 of the troops belonged to MOS groups that contained N ≥ 7 people. Logistic and linear regression analysis demonstrated significant associations of SARS-CoV-2 IgG (p <0.016 and p < 0.002 respectively) with MOS codes 2T and 92F (Table 1 and Fig. 2). MOS 2T performs and manages activities related to transportation by ground and air, while MOS 92F is a fuel specialist for both ground and air transport. Troops with MOS codes 2T and 92T come in contact with individuals and cargo from around the US and the world via air transport. MOS 2T and 92F remained significantly associated with SARS-CoV-2 serum IgG levels after parsimonious adjustment of linear regression models, accounting for demographic associations (Supplemental Materials Table S3).

• Download figure
• Open in new tab

Figure 2.

Spike antigen-specific IgG in serum from subjects with different MOS codes. The mean and standard error of spike IgG ELISA OD values are shown for subjects that belong to different MOS groups with N ≥ 7 per MOS. * indicates significant (p < 0.017) difference vs. other all other MOS codes.

Among those subjects that held MOS codes with N ≤6 persons it was notable that 2/2 subjects with MOS code 155E (C-12 fighter pilots) were SARS-CoV-2 IgG positive, as were several subjects with MOS codes for nursing subspecialties, MOS 66-68 (Supplemental Materials Fig S2). However, the low number of subjects with the corresponding MOS codes lacked statistical power to attribute significance to the observed associations with MOS 155E or 66-68.

SARS-CoV-2 Serology Results and Geolocation

The distribution of SARS-CoV-2 IgG positivity throughout CT state was visualized based on zip code as shown in Figure 3. The notable hot spots were in Stamford, the 1^st major city north of New York City (NYC), the Northern part of the state near the international airport, and the zip codes for Bristol, Bridgeport and Waterbury; areas in CT’s upper quartile for population density and diversity, but lower quartile for median household and per capita income. The data are consistent with NYCs concomitant outbreak and the socio-economics of COVID-19, and further suggest association of infection with travel, given the number SARS-CoV-2 IgG positive cases near the airport/interstate highway, and the association with MOS codes 2T and 92F.

• Download figure
• Open in new tab

Figure 3.

Geolocation of SARS-CoV-2 IgG+ subjects according to home zip code. The number of SARS-CoV-2 IgG+ subjects per zip code was mapped based on home zip code. Note highlighted hot spots (*) along the major (I-95 and I-91) interstate highways, Stamford, the town just north of NYC, Bristol, Bridgeport, Waterbury, and the north-central part of the state near the Hartford-Springfield International Airport.

SARS-CoV-2 neutralization capacity of serum

SARS-CoV-2 serum neutralizing capacity of non-hospitalized individuals with mild or asymptomatic COVID-19 is relatively understudied vs. that of hospitalized patients with more severe disease (16). We measured the neutralizing ability of serum from a subgroup of subjects (N=12) with positive SARS-CoV-2 IgG test results in a live viral plaque assay. As shown in Fig. 4, all twelve subjects’ serum had significant neutralizing capacity with IC_50s ranging from 1:15 to 1:280. On average the subjects serum neutralizing capacity was comparable, albeit lower, than that of hospitalized patients as we recently published (17). Notably, serum spike IgG ELISA levels (OD) were closely related (r=.75) to the observed variability in IC₅₀.

• Download figure
• Open in new tab

Figure 4.

Neutralizing capacity of serum from individuals with mild/asymptomatic COVID-19, not requiring hospitalization. (A) Serum was titrated in a live viral plaque assay to measure neutralization capacity (% inhibition of plaque formation) expressed on the Y axis vs. serum concentration (X-axis). Different color line/symbol represents data from a single subject with ID# shown in key in upper right, and negative control (blue triangles). (B) Correlation between the IC₅₀ (Y-axis) and the spike IgG ELISA OD values (X-axis). Each symbol represents the average of replicate data points from a single individual in key above. r²=.567 for the best fit trend line, log(y) = .495x + 0.991. (C) Comparison of serum neutralizing activity, IC₅₀ (Y-axis) of the present National Guard troops (NG) vs. hospitalized study subjects recently published (17). Each symbol represents data from a single subject’s serum titration. The mean and standard error for the different groups of study subjects is shown.

Longevity of SARS-CoV-2 serum IgG

Persistence of SARS-CoV-2 IgG in serum was evaluated in a limited number of subjects that returned for repeat testing. Follow-up samples obtained 6 weeks to 5 months after the initial test showed no significant difference in average SARS-CoV-2 spike IgG serum levels vs. the initial time point (Fig. 5). However, the lack of change may be obscured by differences in time of sample collection relative to infection, as some subjects SARS-CoV-2 spike IgG serum levels decreased, while others increased. Perhaps most notably, four subjects remained SARS-CoV-2 spike IgG positive for >7 months after the initial test, with sustained elevated ELISA signal.

• Download figure
• Open in new tab

Figure 5.

Longevity of serum SARS-CoV-2 spike IgG. The amount of spike antigen-specific IgG was measured based on the ELISA OD (Y-axis) over time (X-axis). Time point 1 = baseline, time point 2 = 6 weeks to 5 months, time point 3 = 7 months. Each different color symbol/line represents the mean of replicate data from a single time point for a single subject. The mean OD at time point 2 is not significantly different (p > 0.05) from time point 1.