Setting
The study was conducted at the University Hospital Fundación Santa Fe de Bogotá (UHFSFB), a tertiary care hospital in Bogotá, the capital and most populated city of Colombia 20,21. During the COVID-19 pandemic, UHFSFB served as a referral hospital for COVID-19 medical attention in the city 22.
In February 2021, Colombia gradually started the COVID-19 vaccination process in two phases. The first phase sought to reduce mortality and incidence of severe disease and protect HCWs, while the second sought to reduce infectivity to reach herd immunity 23. HCWs were among the first to complete their primary vaccination schedule. Thereafter, by November 2021, the government approved a booster dose for adults, that was administered at least 6 months after completing the initial schedule 24. The vaccines available in the country included BNT162b2 (Pfizer), mRNA-1273 (Moderna), Ad26.COV2.S (Janssen), AZD1222 (AstraZeneca), and CoronaVac (Sinovac) 23. For the booster dose, HCWs could access a homologous booster or a heterologous booster (using an mRNA or a viral vector-based vaccine), according to their preferences and vaccine availability 24.
Study design and participants
In 2020, the UHFSFB, in collaboration with Universidad de los Andes, conducted the CoVIDA-FSFB study 25,26. The prospective cohort study enrolled a cohort of 420 voluntary adult hospital workers recruited between June 25 and October 30, 2020, who underwent routine SARS-CoV-2 RT-PCR and serological testing over six months, until April 30, 2021 (Fig. 1). During the follow-up by March 2021, a subgroup of participants received a two-dose BNT162b2 schedule. Collected serum samples were stored at -70°C until further analysis. To estimate the infection-induced, vaccine-induced, and hybrid humoral immunity against SARS-CoV-2, this analysis focused on a subgroup of the CoVIDA-FSFB participants who met any of the following inclusion criteria (n = 110): (i) RT-PCR-confirmed SARS-CoV-2 infection before study recruitment (n = 29), (ii) RT-PCR-confirmed SARS-CoV-2 infection during the study follow-up (n = 57), (iii) received a primary vaccination schedule with BNT162b2 (n = 24). For this analysis, participants with re-infections, contraindications for phlebotomy, and characteristics that hindered follow-up were excluded (e.g., change of residence, planned long-term travel outside the city). Stored samples of eligible participants were subsequently sent to the Center for Autoimmune Diseases Research (CREA) for analysis.
Subsequently, the participants received the first vaccine booster between November 26, 2021, and January 4, 2022. Those who received a booster dose of mRNA-1273 after a two-dose primary schedule of BNT162b2 (2BNT162b2 + 1mRNA-1273 schedule) (n = 36) were invited to participate in an ancillary component to study the humoral immunogenicity of this vaccine combination (Fig. 1). The cellular immune response was evaluated in a subset without underlying comorbidities, acute infections, and chronic or acute use of medication (n = 16). Individuals were scheduled a new visit between March 24 and April 11, 2022, to assess eligibility for this study component, and to collect additional information and blood samples to assess humoral and cellular immunogenicity in those who were eligible. Subsequently, they were followed for six additional months, which concluded on October 25, 2022. During this period, blood samples for humoral immunity assessment were collected at 6 and 9 months after the booster, and participants were contacted monthly to identify laboratory-confirmed COVID-19 cases (Fig. 1).
Outcomes
Measurement of IgG, IgA, and neutralizing antibodies
The Euroimmun anti-SARS-CoV-2 ELISA (Euroimmun, Luebeck, Germany) was used for serological detection of human IgG and IgA antibodies against the SARS-CoV-2 wild-type S1 structural protein, following the manufacturer’s instructions, as previously described 27. To evaluate results, a ratio of the OD of the patient sample over the OD of the calibrator was calculated. Ratios < 0.8 were deemed negative, ≥ 0.8 to < 1.1 were considered borderline, and ≥ 1.1 were classified as positive. Antibody positivity was determined using a 1:100 dilution.
The anti-S SARS-CoV-2 IgG II Quant assay (S-IgG) (Abbott, Sligo, Ireland) was used to assess the IgG response to the 2BNT162b2 + 1mRNA-1273 schedule. The assay was conducted on the Abbott ARCHITECT i2000SR system according to the manufacturer’s instructions 28. The assay allows for qualitative and quantitative determination of IgG antibodies against the SARS-CoV-2 glycoprotein receptor binding domain (RBD) in human serum and plasma 29. The units of the quantitative Abbott anti-S assay, Arbitrary Units per milliliter (AU/mL), were converted to the World Health Organization (WHO) units, Binding Antibody Units per milliliter (BAU/mL), by multiplying by a factor of 0.142, according to the manufacturer’s instructions.
To evaluate the neutralizing capacity of anti-SARS-CoV-2 antibodies, the semi-quantitative assay NeutraLISA kits (EUROIMMUN, Lübeck, Germany) was used. This kit detects IgG antibodies capable of neutralizing the S1 subunit where RBD of the SARS-CoV-2 spike protein is located. Results were reported as percent inhibition (%Inhibition) following the manufacturer's instructions, as previously described 30. Samples were classified as negative (< 20% inhibition), positive (≥ 35% inhibition), or inconclusive (20–34% inhibition).
PBMC Isolation and cryopreservation
Blood collected in EDTA-anticoagulated tubes was used to isolate peripheral blood mononuclear cells (PBMCs) through a density gradient centrifugation method using Ficoll-Histopaque 1077 (Sigma-Aldrich, St Louis, USA) following the manufacturer's instructions. For cryopreservation, the isolated PBMCs were washed twice with complete RPMI-1640 media (Gibco, NY, USA) and then frozen and stored in fetal bovine serum (FBS) (BioWest, Riverside, USA) containing 10% dimethyl sulfoxide (DMSO) (Sigma-Aldrich, St Louis, USA). Cryovials containing the PBMCs were initially stored at -70°C to allow for a gradual temperature decrease. After 24 hours, these cryovials were transferred to a liquid nitrogen tank, where they were stored until further use.
Measurement of SARS-CoV-2 Specific T-cell response
To explore the SARS-CoV2-specific T-cell response to the 2BNT162b2 + 1mRNA-1273 schedule, three different peptide pools of SARS-CoV-2 wild type (Mabtech AB, Nacka Strand, Sweden) were used:
-
SARS-CoV-2 S1 scanning pool, which contains 166 peptides from the human SARS-CoV-2 virus; the peptides are 15-mers overlapping with 11 amino acids, covering the S1 domain of the spike protein (amino acid 13–685).
-
SARS-CoV-2 SNMO defined peptide pool that contains 47 synthetic peptides from the human SARS-CoV-2 virus; the peptides are derived from the spike, nucleoprotein, membrane protein, ORF3a and ORF7a.
-
S2 N defined peptide pool, which contains 41 peptides from the human SARS-CoV-2.
According to the manufacturer's instructions, each pool of peptides was individually resuspended in DMSO and PBS, resulting in a final stock concentration of 200 µg/mL. Then, cryopreserved PBMCs were thawed in a 37°C water bath, washed twice with RPMI-1640 media pre-warmed to 37°C, and centrifuged at 350g for 10 min. Afterward, cells were analyzed for viability using trypan blue, and seeded at a density of 1 × 106 cells per well in a 96-well plate in RPMI-1640 supplemented with 10% FBS, 100 U/mL Penicillin, 100 µg/mL Streptomycin and 2mM L-glutamine (Gibco, NY, USA). After, cells were rested for 2h and then stimulated with each SARS-CoV-2 peptide pool independently at a final concentration of 2 µg/mL overnight (~ 18h) at 37°C and 5% CO2. As positive control, cells were stimulated with 5 ug/mL of phytohemagglutinin (Sigma Aldrich, St Louis, USA), and as negative control, cells were left unstimulated. All conditions were seeded with Brefeldin A at 10 µg/mL (Sigma Aldrich, St Louis, USA) to inhibit protein transport. The percentage of SARS-CoV-2-specific IFN-γ, IL-2, IL-4 and granzyme B-producing cells were evaluated by flow cytometry. After stimulation with the SARS-CoV-2 peptide pools, cells were harvested and stained with 7AAD-PERCP, anti-CD3-APCH7, anti-CD4-V500 and anti-CD8-APC antibodies (BD Biosciences, CA, USA) at room temperature for 30 min. For intracellular cytokine staining, cells were fixed and permeabilized with BD Cytofix/Cytoperm™ (BD Biosciences, CA, USA), followed by staining with anti-IFN-γ-FITC, anti-IL-2-V450, anti-IL-4-PECy7, and anti-Granzyme B-PE antibodies (BD Biosciences, CA, USA) for 30 min at 4°C in the darkness. Controls for these assays included single-staining and unstained cells, which were used for gating and compensation. Cells were acquired on a FACSCanto II flow cytometer (BD Biosciences) and data were analyzed with FlowJo software version 9 (BD Biosciences).
Data Sources
During the cohort’s first visit, researchers asked participants about their sociodemographic information and medical history and recorded this information in a medical record and an electronic questionnaire. The medical record was used to document information regarding comorbidities, flu vaccination, and previous viral infections. To gather data about previous viral infections, researchers asked participants whether they had ever been diagnosed with any of the following: dengue, chickenpox, zika, chikungunya, influenza, measles, or hepatitis. These infections were self-reported by the participants, and no specific diagnostic tests conducted to confirm them.
For the 2BNT162b + 1mRNA-1273 immunogenicity subgroup, additional data were collected through an electronic questionnaire implemented in REDCap (S2 File in Spanish) 31. This included data on sociodemographic characteristics (e.g., age, sex, socioeconomic status, city in which they live, and profession), clinical characteristics (e.g., height, weight, Body Mass Index (BMI), blood type, comorbidities, medications, and previous COVID-19 infection), and habits (e.g., physical activity, alcohol, and smoking cigarettes). To avoid inter-interviewer bias, the questionnaire was administered by the same investigator. Information regarding COVID-19 vaccination was obtained from the participant’s vaccination card, which contains information about the vaccination schedule, including the vaccine batch, laboratory, dosage, administration dates, and the health provider institution that administered the vaccine. During the follow-up period, the administration of a second booster dose was approved in Colombia. Participants were asked about receiving this dose, and verification was conducted through the vaccination certificate by the end of the follow-up.
Statistical analysis
For descriptive analysis qualitative variables were presented as frequencies and proportions, and quantitative variables as means or medians with standard deviations (SD) or interquartile ranges (IQR) depending on their distribution, according to the Shapiro Wilk test. There were no missing data on the independent variables. Missing data on the humoral immunogenicity outcome corresponded to 4.86% (7/144), which were not included in the analysis.
IgG and IgA titers were compared before and after a two-dose BNT162b2 primary schedule using the Wilcoxon signed rank test. To compare IgG, IgA, and neutralizing antibodies after the mRNA-1273 booster the Skillings Mack test was used. Differences in IgG, IgA, and neutralizing antibodies according to sociodemographic, clinical variables, and habits were graphically explored.
The CMIA kit used to measure anti-spike IgG, for 2BNT162b2 + 1mRNA-1273 immunogenicity assessment, provides values up to 5680 BAU/mL. Values above that threshold were set as equal to the threshold; thus, the data for this variable were right censored. To address this, and given the longitudinal nature of our data, a random effects Tobit model was used to determine factors related to anti-S-RBD IgG antibodies. Regression models were constructed using anti-S-RBD IgG post-booster as the dependent variable. All clinically relevant variables with biological plausibility previously identified through literature search were included as independent variables. Two models were constructed, a bivariate model and a multivariate reduced model with the minimum number of independent variables that best suited the data, using a 0.2 significance level for variable removal from the model 32. The multivariate model was used to adjust for confounders and detect effect modifiers. All possible interactions between the variables of interest were explored; however, these were not included in the final model as they were not statistically significant. Multicollinearity was assessed using the variance inflation factor (VIF), with a 5.0 cut-off point. Bootstrapping was used to provide more reliable standard errors 33. The random effects used account for autocorrelation that may arise from within cluster dependencies. The quadrature approximation used in the random-effect estimators was checked, with no relative differences in the coefficients larger than 0.01%. The normal distribution of raw residuals was confirmed.
T-cell responses in participants with 2BNT162b2 + 1mRNA-1273, were compared according to whether they previously had COVID-19. Specifically, CD4 + cells producing IFN-γ, IL-2, and IL-4, and CD8 + cells expressing Granzyme B, IFN-γ, IL-2, and IL-4. Responses were evaluated post-stimulation with three distinct peptide pools (S1, SNMO, and S2 N), and the Mann-Whitney U test was used to assess statistical significance. A P value < 0.05 was considered to indicate statistical significance for all statistical tests. Analysis was performed using Stata SE 17.0 34 and visualized in GraphPad Prism version 9 35.
Ethics statement
The study protocol was approved by the Fundación Santa Fe de Bogotá Ethics Committee (CCEI-12183-2020, and CCEI-13882-2022). This study was conducted in compliance with Act 008430 − 1993 of the Ministry of Health of Colombia, and classifies as minimal-risk research 36. All patients provided their written informed consent and were informed about the Colombian data protection law (1581 of 2012). All research was performed in accordance with relevant guidelines and regulations, and in accordance with the Declaration of Helsinki.