Study Design
Wake Forest University School of Medicine IRB authorized the study and consent forms under IRB#71181. Participants were recruited from a single retirement facility staffed by academic geriatricians. All retirement facility residents were eligible for participation. Exclusion criteria included lacking capacity to consent, and clinically significant changes in health status including vital sign instability and/or unstable health conditions. Older adults living in either a nursing home or assisted living were assessed by a physician member of the research team for the capacity to consent at study enrollment prior to volunteering written informed consent. Capacity to participate was assessed at each subsequent sample collection.
A cohort of older adults in a retirement facility were followed from the first Moderna mRNA-1273 vaccine dose in February 2021 with blood sample collections at baseline, second dose (4 weeks post first dose), and 2 weeks, 3 months, and 6 months post second dose. A subset of participants was followed for a final blood collection 2 weeks following a third Moderna mRNA-1273 vaccine dose (booster dose). Each study visit blood draw occurred in participant domiciles, which was accompanied by a comprehensive geriatrics clinical assessment with frailty status characterized using the 9-point Clinical Frailty Scale developed by Rockwood.35 Health conditions, including prior COVID-19 infection, were verified through review of participants’ medical information in the electronic health record (EHR) [EPIC Systems, Madison, WI USA]. For our primary outcomes we measured antibody titers, avidity, and cellular activation to the first 2 COVID-19 mRNA vaccine doses. For our secondary outcomes we measured antibody titers and avidity to the booster vaccine dose. We conducted exploratory analysis of the association cellular immune phenotype with participant frailty status.
Human Sample Collection and Storage
Sample collections and timeframe were modeled from the Moderna BNT162b2 vaccine study.55 Plasma was collected using Heparin vacutainer tubes (BD Biosciences). Peripheral blood mononuclear cells (PBMC) were separated from fresh plasma using Ficoll (Cytiva) density gradient centrifugation in Leucosep tubes (Greiner Bio-One) and were cryopreserved in 10% dimethylsulfoxide (DMSO from Sigma) supplemented with fetal bovine serum (FBS from Atlanta Biologicals). Serum was drawn into SST vacutainer tubes containing clotting activator (BD Biosciences) and left at room temperature for 30–60 min, before centrifuging for 10 min at room temperature. Serum and plasma following PBMC separation were aliquoted and frozen at − 80°C.
Assessment of Humoral Responses
Elisa
We performed enzyme-linked immunosorbent assays (ELISA) to quantify anti-spike and anti-RBD IgG antibodies from serum and plasma with previously established assays.74 Antibodies were validated by the manufacturer and titrated for ELISA by serial dilution.
Reagents included phosphate buffered saline (PBS from BioWhittaker, Lonza ), Tween 20 (Fisher), TMB for chromogenic development (Sigma-Aldrich), and milk (BD Sciences). Ninety-six half well-plates (Greiner Bio-One) were coated with antigen or PBS for a negative control overnight at 4°C, then washed with PBS-0.1% Tween and blocked with PBS-3% milk at room temperature for 1 h. Antigens used in ELISA were SARS-CoV-2 Washington-1 spike protein at 12.5 ng/ml or RBD protein at 25 ng/ml (BEI Resources, NIAID, NIH). Aliquoted serum samples were titrated in eight 2-fold serial dilutions in PBS-1% milk starting at a minimum dilution of 1:100, and given high titers in some subjects, starting dilution was increased to a maximum of 1:12,800. Plates were incubated with diluted serum for 2 h at room temperature, washed with PBS-0.1% Tween and incubated with goat anti-Human IgG HRP (1:4,000) (Southern Biotech) detection antibody in PBS-0.1% milk for 1 h at room temperature. Plates were washed with PBS-0.1% Tween and developed with TMB for 30 minutes at room temperature in the dark. The reaction was stopped with 2N H2SO4 and the plates were read at 450 nm immediately after stopping. The limit of detection was defined as 1:100.
Avidity
Quality of antibody binding was assessed with an avidity assay following a previously established procedure.75,76 The ELISA assay was performed with spike protein as described above, with modifications as follows. Participant serum was used at a dilution that resulted in half-maximal peak ELISA titer values. Prior to incubation with detection antibody, sodium thiocyanate (NaSCN from Acros Organics) was added at a starting concentration of 5 M with 2-fold dilutions to 0.195 M in PBS per well. PBS alone was added to a negative control well. After a 15 min incubation at ambient temperature, the plate was washed again with PBS-Tween, detection antibody was added, and the ELISA was continued as previously described. The calculation of the avidity index will be described in more detail in the statistical analysis subsection.
Stimulation and Staining of Human Peripheral Blood Mononuclear Cells
Peptides from Washington-1 SARS-CoV-2 spike protein were obtained from BEI Resources (NR-52402). The entire 181 individual peptides were combined in a peptide mega-pool. Each peptide was resuspended in 70% acetonitrile, pooled, aliquoted, and lyophilized before storing at -80 C.
PBMC samples were thawed, washed, and resuspended in RPMI media with 5% human albumin and L-Glutamine with between 5x105 to 1x106 cells per well in 96-well U bottom plates. Cells were cultured for 24 hours in the presence of SARS-CoV-2 spike pooled peptides [0.8 µg/ml of each peptide] or 10 µg/mL PHA (Sigma) at 37°C in a humidified atmosphere containing 5% CO2. Incubation with an equimolar amount of DMSO (Sigma) was performed as negative control.
Surface staining was performed on PBMCs following 24h stimulation culture. Cells were resuspended in 100 ml PBS with 2% FBS (FACS buffer), then underwent wash and centrifugation between steps, including Live/Dead stain (Biolegend), Fc block (Biolegend), and antibody cocktail stain (antibodies from Biolegend) for 30 minutes at 4°C in the dark. Following surface staining, cells were washed twice with FACS buffer. After the final wash, cells were resuspended in 200ul PFA fixation buffer.
Flow Cytometric Activation Induced Cell Marker (AIM) Assay
We used activation induced surface markers (AIM), a cytokine independent approach, for functional measurement of T cell activation. This technique has previously been reported to be highly sensitive for detection of rare cell populations including circulating Tfh cells.77 We defined a 13-color flow cytometry panel of lymphocyte lineage markers (Supplementary Table 1) to assess AIM and the adaptive cellular immune-phenotype of vaccine respondents.
All samples were acquired on a Fortessa X20 cytometer configured with five excitation lasers (355, 405, 488, 561, 640 nm) and 20 detectable parameters (BD Biosciences, San Diego, CA). Data in .fcs format were exported from the FACSDIVA software of the cytometer and processed directly using FlowJo (version 10.1, FlowJo, LLC).
Statistical Analysis
Data was analyzed using R, version 4.0.3 (R Foundation, Vienna Austria). Figures were created using Prism 9.1.0 (GraphPad Software, San Diego California). A two-sided significance of P < 0.05 was considered significant. Results from Bayes statistics yielded a posterior probability distribution (PD), which was converted to a P value for general interpretability.
Using linear mixed regression models we examined the association of log-transformed antibody titers with fixed effects including clinical and demographic variables (age, sex, frailty status, and prior COVID-19 infection) for the secondary vaccine dose response with each study participant representing a random intercept. For the analysis of the response to the booster vaccine dose we used Bayesian linear mixed models to avoid singular model fits. Uninformative (flat) priors were used for all coefficients and intercept. Avidity assays used Prism software to calculate IC50 values for inhibitor response curves to estimate the concentration of inhibitor required for half-maximal antibody binding.
Cellular analysis used linear regression models to report results of AIM + T cells following stimulation with spike pooled peptides with subtraction of DMSO negative control values. We analyzed the relationship of antibody and cellular response with geometric mean titers (GMT) of individuals’ antibody titers following 2 vaccine doses in association with AIM + T cells. Next, we reported the associations of clinical variables and the cellular immune-phenotype with AIM + T cells in multivariable models using stepwise leap forward selection techniques from the Caret R package. Parameters were selected using 10-fold cross validation to minimize overall model root mean squared error. Final models were fit using predictors with the highest explanatory power measured by the lowest RMSE. Additionally, we conducted exploratory analysis with correlation plots to describe the associations of participant frailty status with their immune-phenotype.