Antiviral responses induced by Tdap-IPV vaccination are associated with persistent humoral immunity to Bordetella … – Nature.com
March 8, 2024
Study design
This human clinical study was designed and conducted in accordance with the provisions of the Declaration of Helsinki (1996) and the International Conference on Harmonisation Guidelines for Good Clinical Practice. The trial is registered at the EU Clinical Trial database (EudraCT number 2016-003678-42) and was approved by the Medical Research Ethics Committee United (MEC-U, NL60807.100.17-R17.039) in the Netherlands and the South Central - Hampshire B Research Ethics Committee (REC, 19/SC/0368) in the UK. Written informed consent was provided by participants and/or legal guardians at the start of the study, after the nature and possible consequences of the studies were explained. Clinical data of participants was recorded using OpenClinica, electronic case record form software that enables compliance with regulatory guidelines such as 21 CFR Part 11. The same online database system was used across the Dutch and UK sites. In the Netherlands, participants/parents/legal guardians were asked to keep their vaccination booklets at hand for the first visit. In case participants/parents/legal guardians did not have a vaccination booklet anymore, permission was asked to contact DVP (Vaccine Supply and Prevention Programs Service) to check their vaccination status according to NIP. In the UK, vaccination history was obtained either from the childs Red Book, or alternatively by means of the participant seeking confirmation from the GP recorded on a standard document.
The aim of this study was to identify early molecular and cellular correlates in peripheral blood that are associated with the humoral response that is mounted in response to Tdap-IPV vaccination in healthy adolescents. This study was conducted as part of a larger multi-center phase IV clinical study that evaluated immune responses to Tdap-IPV vaccination in participants from the United Kingdom, Finland, and the Netherlands7. Participants in the Dutch arm of the study were enrolled between October 2017 and March 2018. Participants were recruited by mail-outs in the Hoofddorp region, facilitated by the Municipal Administration, and the study was conducted by the Spaarne Academy (Spaarne Hospital, Hoofddorp, the Netherlands). Participants in the UK cohort were enrolled between April 2018 and January 2020 via mail out to eligible participants within postal areas. Participants in the NL or UK cohorts were included if they were in good health and had received all regular vaccines for their age group according to the Dutch or UK national immunization program (NIP), including an aP booster vaccination in preschool. Sex was not explicitly considered in the study design and all children between ages 11 and 15 years old were eligible for inclusion in the study. Male and female adolescents with either aP or wP vaccination priming backgrounds were included in the present study (NL cohort: N=14, 8 males and 6 females; UK cohort: N=12, 6 males and 6 females), TableS1. Assignment of aP or wP background in this study was basedon the participant date of birth and the available pertussis vaccine in the Netherlands at that time7. The number of participants included in the present study was informed on the basis of feasibility for carrying out a study for systems vaccinology. All participants received a dose of Tdap-IPV vaccine (BoostrixTM-IPV, GlaxoSmithKline (GSK), Wavre, Belgium) administered by intramuscular injection in the deltoid muscle of the upper arm at baseline (D0). The primary outcome was pertussis toxin-specific IgG antibody concentration at D28 post-vaccination. Secondary outcomes include but are not limited to IgG responses against the other Tdap vaccine antigens TT, Dt, Prn, FHA at D0, D28, and Y1, and PT at D0 and Y1. Exploratory outcomes relevant for this manuscript were measured at D0 and D1 post-vaccination, whole blood transcriptomic analysis, measurements of immune cell abundance and cytokine production by mass cytometry, and single-cell gene expression profiling of innate immune cells.
Whole blood was collected in tubes containing sodium heparin (Greiner vacuette in 4ml, 6ml, and 9ml volumes, catalog numbers 454030, 456028, 455051) to prevent coagulation. Complete blood counts were obtained using a Sysmex XN-450 hematology analyzer and sera were stored at 20oC until analysis. Whole blood was processed for single-cell RNA sequencing or mass cytometry analysis (detailed below). Whole blood was also collected directly in PAXgene Blood RNA Kit tubes (PreAnalytix) for transcriptome analysis and frozen at 80oC until processing.
Sera were analyzed for PT-, FHA-, Prn-, TT-, and Dt-specific IgG antibody concentrations using a fluorescent-bead-based multiplex immunoassay29. Antigens were covalently coupled to distinct color-coded activated carboxylated MicroPlex Microspheres (beads) (Luminex, Austin, Texas, USA) according to the manufacturers instructions. The following antigens were used for coupling: PT (GSK, Belgium), FHA (Sanofi, France), Prn (GSK, Belgium), diphtheria toxoid (Netherlands Vaccine Institute) and TT (T3194, Sigma-Aldrich, Saint Louis, Missouri, USA). After a wash step in PBS, 12.5106 carboxylated beads/mL were activated in PBS containing 2.5mg of 1-ethyl-3-(-3-dimethylaminopropyl)-carbodiimide hydrochloride (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and 2.5mg of N-hydroxy-sulfosuccinimide (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The antigens for coupling were diluted in PBS to a concentration of 10ug of PT, FHA or Prn, 100ug of Dt or 25ug of Tt per 12.5106 activated beads and incubated for 2h at room temperature in the dark under constant rotation. After 3 wash steps the antigen-coupled beads were stored in the dark in PBS containing 0.03% (wt/vol) sodium azide and 1% (wt/vol) bovine serum albumin at 4C until use.
Sera diluted 1/200 and 1/4000 in PBS containing 0.1% (vol/vol) Tween 20 and 3% (wt/vol) BSA were incubated with antigen-coupled beads in a 96-well filter plate for 45min at room temperature at 750rpm in the dark. Reference sera in a dilution series, quality control sera and blanks were included on each plate. The in-house reference standard for pertussis was calibrated against WHO 1st IS Pert 06/140 and serially diluted 4-fold over 6 wells (1/200 to 1/204800). The in-house reference standard for Dt and Tt was calibrated against WHO NIBSC DI-3 and TE-3 and serially diluted 4-fold over 8 wells (1/50 to 1/819200). Following incubation, wells were washed 3 times with PBS, incubated with R-phycoerythrin-labeled goat anti-human IgG antibody (Jackson Immunoresearch Laboratories, West-Grove, PA, USA, catalog number, 109-115-098, 1:200 dilution) for 30min and washed. Beads were included in PBS and median fluorescence intensity (MFI) was acquired on a Bio-Plex LX200. MFI was converted to IU/mL by interpolation from a 5-parameter logistic standard curve using Bioplex Manager 6.2 software (Bio-Rad Laboratories, Hercules, California, USA) and exported to Microsoft Excel.
Sera in the NL cohort were also analyzed for Polio.I, Polio.II, and Polio.III-specific IgG antibody concentrations using a fluorescent-bead-based multiplex immunoassay28. Type-specific capture monoclonal antibodies (MAb), respectively antipoliovirus type 1 clone 9B4 (HYB 295-17-02 ThermoFischer scientific, Waltham, MA USA), type 2 clone 24E2 (HYB 294-06-02 ThermoFischer scientific) and type 3 clone 4D5 (HYB 300-06-02 ThermoFischer scientific) were conjugated to three distinct activated carboxylated microspheres. Briefly, 800ul of carboxylated microspheres (12.5106 beads/ml; Bio-Rad Laboratories, Hercules, CA USA) were activated by adding 100l of 50mg/ml N-hydroxy-sulfosuccinimide (sulfo-NHS, ThermoFischer scientific) and 100l 50mg/ml 1-ethyl-3-(-3-dimethylaminopropyl)-carbodiimide hydrochloride (EDC, ThermoFischer scientific) in PBS. The microspheres were activated at room temperature for 20min in the dark under constant rotation, washed once with 1ml PBS and resuspended in 1ml PBS containing 50g/ml of antipoliovirus monoclonal antibody. The beads were incubated for 2h at RT in the dark under constant rotation. Subsequently, the beads were washed three times with PBS and stored in the dark in PBS containing 0.05% (wt/vol) sodium azide and 1% (wt/vol) bovine serum albumin at 4C until use.
An in-house reference standard serum (RIVM MIA reference standard serum, Bilthoven, the Netherlands) was calibrated against the 3rd International Standard antiPoliovirus serum Types 1, 2 and 3 (NIBSC code: 82/585 assigned potency 11/32/3IU/ml) and used as the reference standard in the polio MIA assay. After calibration, the international standard was used as a control serum.
For quantification of Polio.I, Polio.II, and Polio.III antibodies in sample sera, reference standard serum, and control serum, sera were diluted and pre-incubated with monovalent IPV type 1, 2, 3 (National Vaccine Institute NVI/Bilthoven Biologicals, the Netherlands28) The reference standard (RIVM MIA reference standard serum) was diluted in 10 steps of 1.5-fold dilutions (1/181/692). Sample sera were diluted 1/2 and 1/25 and the control serum diluted 1/20 in dilution buffer (PBS containing 0.25% (v/v) Tween-20, 1% (w/v) BSA and 0.5M NaCl). Subsequently all dilutions were mixed 1:1 (v/v) with IPV in dilution buffer containing IPV type 1:56 DU/ml, IPV type 2:16 DU/ml and IPV type 3:80 DU/ml, ergo resulting in a final reference serum dilution of 1/361/1384, in serum dilutions of 1/4 and 1/50, in a control serum dilution of 1/40 and a IPV concentration of 0.7/0.2/1.0 DU/well (25ul) for IPV type 1, 2 and 3. After 2h incubation on a shaking platform (1000 RPM) at RT, 25ul of the samples, reference standard and controls dilutions were transferred to a pre-wetted 96-well Multiscreen HTS filter plate (Millipore Corporation, Billerica, MA) containing a 25ul per well mix of antipoliovirus MAb type 1, 2 and 3 conjugated microspheres (4000 beads/region/well) in dilution buffer. The plates were incubated for 1h at room temperature in the dark on a plate shaker at 600rpm. The beads were collected by filtration using a vacuum manifold and washed three times with 100ul PBS. A recombinant Human CD155/FC chimera (rhCD155/Fc 100gr, Sino Biological Inc. Beijing, China) was labeled with R-phycoerythrin (RPE) using the SiteClick RPE antibody labeling kit (ThermoFischer/life technologies) following manufacturers instructions and used for detection of IPV type 1, 2 and 3. To each well 50ul of a 1/2000 dilution of RPE conjugated CD155-Fc (1mg/ml) in PBS was added and the plate was incubated for 30min with continued shaking (600rpm). Plates were washed and beads were included in PBS and median fluorescence intensity (MFI) was acquired on a Bio-Plex LX200. MFI was converted to IU/mL by interpolation from a 5-parameter logistic standard curve using Bioplex Manager 6.2 software (Bio-Rad Laboratories, Hercules, California, USA) and exported to Microsoft Excel.
RNA extraction from whole blood, library preparation, sequencing, data processing, and differential gene expression analysis is described in the Supplementary Information.
Differential gene expression analysis of pre- (D0) and post-vaccination (D1) blood samples was performed with DESeq2 (v1.38.3), which takes a raw read count x sample matrix as input and includes library size normalization, estimation of dispersion, model fitting and gene filtering steps. We analyzed the effect of priming background (aP or wP vaccine) and sex on the vaccine response (D1 vs D0) with a model that included the background or sex variable, as well as its interaction with vaccination day and subject ID. A gene was considered differentially expressed if the FDR adjusted p-value was <0.05 and |log2 fold change|>0.5. As such, we analyzed gene expression with a final model that specifies vaccination day as the covariate of interest and controls for subject-level variations. The output from the DESeq2 pipeline includes a gene matrix with log2 fold change (D1/D0) and FDR-corrected p-values, which was used for downstream analysis.
Processing of whole blood samples for mass cytometry is described in the Supplementary Information.
FCS files were imported into Cytobank data analysis software (v6.1.2) and the arcsinh transform of marker expression values was calculated for downstream analysis. Major immune cell lineages were manually gated (Fig. S10A). For subpopulation analysis of APCs, manually gated monocytes, pDCs, and mDCs were exported per participant per timepoint, for a total of 24 FCS files (one pre- and one post-vaccination sample per participant for 12 participants). These files were imported into R (v3.6.1) for unsupervised clustering and subpopulation analysis according to a predefined workflow67. In total 285,780 APCs were identified across all 24 samples and were pooled for clustering, which was performed using the following markers: HLA-DR, CD69, CD31, CD86, CD16, CD123, CD33, CD14, CD1c, CD11c, CD62L, CD38, CD141, CD11b. Cluster markers were selected on the basis of their importance for describing heterogeneity and activation states among monocytes and dendritic cells. In total, 14 clusters were detected and manually annotated based on their phenotypic marker expression (Fig. S11A). Cluster abundance was similar across all samples (Fig. S11B). We derived a high-level stratification of cells for downstream analysis by manually merging clusters of APCs based on their similarity, which yielded eight subpopulations (phenotypic marker expression and the mapping of clusters to subpopulations is provided in Fig. S11C). The frequency of each cell type and APC subpopulation in our mass cytometry data was calculated as a fraction of CD45+ cells for each sample. These frequency values were multiplied by the total leukocyte counts (Fig. S10D) per ml of blood per sample to derive cell type (Fig. S10E) and APC subpopulation (Fig.4CE) abundance values. Expression values for analysis of intracellular cytokine responses were calculated as the mean signal intensity of each subpopulation for each sample. In order to identify co-expression of IL-6 and IFN-a in classical monocytes (Fig. S12), per donor, classical monocytes from D0 and D1 samples were pooled and the 95% quantile was calculated for each cytokine. Cells that were above the 95% quantile were classified as positive, and cells below this threshold were classified as negative. Thereafter, the frequency of single-positive or double-positive cells was calculated among all cells within a D1 or D0 sample.
Single-cell index sorting of immune cells, RNA extraction, library preparation, pre-processing of raw RNA sequencing data, filtering of low-quality cells, as well as pre-processing of single-cell FACS data is provided in the Supplementary Information.
We defined subpopulations of APCs by integrating FACS marker and single-cell gene expression to define populations using multi-omic factor analysis (MOFA68), which enables variance decomposition of the datasets and factors using the coefficient of determination (R2). The MOFA model was constructed using the single-cell FACS marker data and the top 30 principal components of the single-cell RNA sequencing data. We followed the developers directions for model selection and downstream analysis (https://biofam.github.io/MOFA2/tutorials.html). Both FACS and single-cell gene expression contributed substantially to the variation in the final model. Furthermore, several factors were identified with shared contributions from each dataset, thereby highlighting the utility of a joint analysis (Fig. S13D). To define subpopulations, we used the single-cell x MOFA factor graph for unsupervised clustering. Similar to the mass cytometry analysis, single-cell data was clustered using a two-step approach. The clusters that were identified are presented in Fig.6A, and their identity was verified by examining both single-cell surface marker and gene expression profiles. In total, we identified classical monocytes (cMo), intermediate monocytes (iMo), non-classical monocytes (ncMo), classical DC1 and DC2, as well as pDCs, which correspond to the same subpopulations previously defined in blood using integrated single-cell FACS and gene expression69.
We created pseudobulk RNA libraries by computationally pooling the gene expression of cells in a given subpopulation for a given pre- or post-vaccination sample. Cells from different single-cell RNA sequencing batches were also pooled. The total number of cells in each pseudobulk sample is shown in Fig. S6E, where each colored segment for each subpopulation represents the number of cells in a pseudobulk sample. The classical DC1 subpopulation was removed from downstream analysis since no sample had >10 cells. Next, raw gene counts of cells within a pseudobulk sample were summed together. We used the DESeq270 bulk RNAseq pipeline to perform differential gene expression analysis based on the developers directions (http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html). We analyzed gene expression of each subpopulation separately with a model that specifies the number of days after vaccination as the covariate of interest and controls for subject-level variations. The output from the DESeq2 pipeline includes a gene matrix for each subpopulation with log2 fold change (D1/D0) and FDR-corrected p-values. These gene matrices were used for downstream analysis of differentially expressed genes (Fig.5C), GSEA (Fig.5D), and correlations with whole blood gene expression (Fig.5E).
GSEA of whole blood and single-cell RNA sequencing data, as well as the comparison with external Tdap vaccination data from in ref. 26. is provided in the Supplementary Information.
Details of the stimulation of peripheral blood mononuclear cells (PBMCs) from healthy donors with vaccines and mass cytometry analysis are provided in the Supplementary Information.
Antibody concentration values were log10-transformed to account for their skewed distribution. Antibody responses were calculated as log10-fold change (LFC) over baseline (D28 or Y1/D0). In order to account for the cross-correlation of LFC antibody responses with D0 antibody levels (Fig. S6), we calculated the adjusted LFC response as described previously in refs. 32,33. Briefly, for each antibody response, at each timepoint, and within each cohort, we fit a linear model with the LFC as the response and the log10 D0 antibody concentration as the predictor variable. Residuals from this model were extracted and summed with the intercept to obtain the adjLFC for each sample of every antibody response. Statistical comparisons of antibody concentrations, antibody responses, and phospho-signaling responses were calculated using a paired Wilcoxon test. Statistical parameters are reported directly in the figures and figure legends. Based on the similarity of antibody responses and transcriptional responses, participants were not separated based on sex or priming background in subsequent analyses (Mass cytometry and single-cell RNA sequencing analysis). Examination of the effects of Tdap-IPV vaccination with mass cytometry consisted of two parts: (i) differential expression (DE) of intracellular cytokines, and (ii) differential abundance (DA) analysis for each subpopulation of cells (TableS4). DE analysis was calculated using the mean signal intensity values for the indicated cytokines, for each subpopulation in each sample. DA analysis was performed on cell counts per unit of whole blood. For both DE and DA analyses, the effects of Tdap-IPV vaccination were determined using a linear mixed-effects model, fit with the lme4 (v1.133) R package71, with cell counts or mean signal intensity as the response, and vaccination day as the fixed effect. P values were generated using the lmerTest (v3.13) package72. We accounted for sample pairs (before and after vaccination) by introducing the subject ID as a random effect. P-values were extracted and corrected for multiple testing using Benjamini-Hochberg method73. Scatterplots displaying the correlation of LFC values with D0 antibody levels (Fig. S6, Fig. S17C) present Spearman correlation coefficient and associated p.value with linear regression trendline. Other statistical analyses (differential gene expression, GSEA) are reported in their respective section of the methods.
Further information on research design is available in theNature Portfolio Reporting Summary linked to this article.
Excerpt from:
Antiviral responses induced by Tdap-IPV vaccination are associated with persistent humoral immunity to Bordetella ... - Nature.com