MVA-S (MVA-SARS-2-S) and MVA-ST (MVA-SARS-2-ST) are two vaccine candidates which are based on recombinant MVA vectors. MVA-S encodes the native full-length spike protein of SARS-CoV-2, while MVA-ST encodes a prefusion-stabilized spike protein with an inactivated S1/S2 furin cleavage site (10, 11, 15). The licensed ChAd vaccine ChAdOx1 nCoV-19 (Vaxzevria) is based on the modified chimpanzee adenovirus ChAdOx1 vector, encoding the full-length spike protein and a tissue plasminogen activator leader sequence (26). The licensed mRNA vaccines BNT162b2 (Comirnaty) and mRNA-1273 (Spikevax) are lipid-nanoparticle-formulated nucleoside-modified mRNA vaccines encoding the prefusion-stabilized spike protein (27). All four vaccines are monovalent, based on the spike protein of wild type SARS-CoV-2.

The phase 1 clinical trials investigating the vaccine candidates MVA-S and MVA-ST were reviewed and approved by the National Competent Authority (Paul-Ehrlich-Institute, EudraCT numbers 2020-003875-16 and 2021-000548-23) and the Ethics Committee of the Hamburg Medical Association (reference numbers 2020-10164-AMG-ff; 2021-100621-AMG-ff), conducted under the sponsorship of the University Medical Center Hamburg-Eppendorf (Hamburg, Germany) in accordance with ICH-GCP and the EU directives 2001/20/EC and 2001/83/EC, and are registered at ClinicalTrials.gov. (NCT04569383; NCT04895449). The monitoring of immune responses following vaccination with the licensed ChAd and mRNA vaccines was approved by the Ethics Committee of the Hamburg Medical Association (reference number: 2020-10376-BO-ff). Written informed consent was obtained from all participants.

NCT04569383 was a phase 1 clinical trial conducted between October 2020 and August 2021 to evaluate the MVA-S vaccine candidate in 30 seronegative individuals divided into two ascending dose groups. Participants were enrolled at the University Medical Center Hamburg-Eppendorf and received two single injections 28 days apart of either a low dose of 1 × 10⁷ ± 0.5 log IU (N = 15) or a high dose of 1 × 10⁸ ± 0.5 log IU (N = 15). In this study, we included a subgroup of participants (N = 12), of whom six participants received the low dose and six received the high dose vaccine.

NCT04895449 was a phase 1b clinical trial conducted between July 2021 and November 2022 to evaluate the MVA-ST vaccine candidate in seronegative individuals (Part A) and in individuals who had previously received two doses of the BNT162b2 vaccine (Part B). In Part A, participants received two single injections 28 days apart, either a low dose of 1 × 10⁷ ± 0.5 log IU (N = 8) or a middle dose of 5 × 10⁷ ± 0.5 log IU (N = 7) and were enrolled at the University Medical Center Hamburg-Eppendorf and the University of Cologne. All participants of Part A (N = 15) were included into this study, while Part B of the clinical trial was not included in the present manuscript.

The mRNA and ChAd/mRNA control cohorts received licensed vaccines known to be immunogenic and were used as a benchmark for the evaluation of the MVA-based vaccines. The participants received either two doses of mRNA vaccine 21 days apart (N = 10), or one dose ChAd plus one dose mRNA vaccine 84 days apart (N = 8), respectively. The participants were enrolled at the University Medical Center Hamburg-Eppendorf between December 2020 and March 2021.

Individuals with self-reported prior SARS-CoV-2 exposure, clinical evidence of COVID-19-like symptoms or positive SARS-CoV-2 antigen and/or PCR test were excluded. If participants acquired a SARS-CoV-2 infection during the study period, subsequent immunogenicity time points were excluded from the analyses in this manuscript.

An overview of all study cohorts is given in Figure 1 . Baseline characteristics, the interval between vaccinations, and the number of samples included in the different analyses according to cohort and time point, are reported in Supplementary Tables 1 - 6 .

A total of 384 peripheral blood samples were obtained from 45 donors. Whole blood was collected in EDTA vacutainers at day 0 (D0, baseline before vaccination), day 1 (D1), day 3 (D3), day 7 (D7) and day 14 (D14) after the first and second vaccination (V1, V2). After centrifugation, plasma was removed and stored at −80°C. PBMCs were isolated by density-gradient centrifugation using Ficoll-Histopaque (Sigma) or SepMate™ (Stemcell), cryopreserved and stored in a gas phase liquid nitrogen tank. Serum was collected at V1D0 and V2D14 using S-Monovettes® containing gel with clotting activators and stored at −20°C. Additionally, whole blood was collected directly into PAXgene RNA tubes (PreAnalytiX, QIAGEN) for bulk RNA sequencing on D0, D1, D3 and D7 post V1 and V2. After keeping the samples for 3-4h at room temperature (RT), samples were stored at -80°C until RNA isolation.

Gene expression was analyzed on D0, D1, D3 and D7 after each vaccination. RNA isolation from thawed PAXgene RNA tubes was performed following the manufacturer’s instructions (PAXgene Blood RNA Kit Handbook, version 2, 2015) using the PAXgene blood RNA kit (PreAnalytiX, QIAGEN), including optional DNase digestion. Extracted RNA from all samples was stored at -80°C.

RNA-seq libraries were generated using the TruSeq stranded mRNA Kit according to the manufacturer’s guidelines. Library size and quality were assessed using a BioAnalyzer high-sensitivity chip (Agilent Technologies), and concentrations were measured with the Qubit 2.0 fluorometer (Life Technologies). Multiplex sequencing was performed on the NovaSeq S4 (2x100bp run). Samples with RNA integrity number (RIN) < 8, insufficient library concentration, or sequencing depth below 25 million reads were excluded from the analysis.

Quality control of raw reads was conducted using FastQC v0.11.8 and MultiQC v1.9. Removal of low-quality read ends and adapters was carried out using fastp v0.23.2 prior to read mapping and gene-wise counting with STAR v.2.7.10a. Ensembl genome GRCh38 and Annotation version 110 were utilized for this purpose. Normalization of raw counts and identification of differentially expressed genes (DEGs) were performed using DESeq2 v1.36.0, with all time points compared to V1D0. A gene was considered differentially expressed with an absolute log2 fold change > 1 and FDR < 0.1.

Innate immune cells were assessed on D0, D1, D3 and D7 after each vaccination. Cryopreserved PBMCs were thawed, distributed into a 96-well V-bottom plate at a number of 1x10⁶ cells per well and stained. For the staining of subsets and activation status of monocytes and dendritic cells (DCs), PBMCs were incubated for 15min at 37°C in 100 µl staining buffer containing PBS, 1% FBS, 1mM EDTA, anti-CD3-APC/Cy7 (1 µl), anti-CD19-APC/Cy7 (1 µl), anti-CD16-BV510 (3 µl), anti-CD14-BV711 (3 µl), anti-CD303-PerCP/Cy5.5 (3 µl), anti-CD11c-APC (1 µl), anti-HLADR-BV785 (3 µl), anti-CD197 (CCR7)-BV605 (3µL), anti-CD40-PE (1 µl), anti-CD1c-FITC (2 µl), anti-CD141-PE/Cy7 (3 µl), anti-CD11b-BUV395 (0.5 µl) and Zombie NIR™ fixable viability stain (0.2 µl).

cTFH cell subsets and activation status were analyzed on D0, D7 and D14 following each vaccination. Following thawing of cryopreserved PBMCs, the cells were rested for 90 min at 37°C and 5% CO₂ in R10 (RPMI containing 1% Penicillin/Streptomycin and 10% FBS) and subsequently distributed in a 96 well V-bottom plate (2x10⁶ cells/well) in duplicates. In order to analyze cTFH cells in their ex vivo state and after restimulation with SARS-CoV-2 spike overlapping peptide pools, one replicate of each sample was incubated with R10, and the other replicate was stimulated with overlapping peptide pools M1-M4, spanning the S1 and S2 subunit of the SARS-CoV-2 spike protein (final concentration: 1 µg/ml; for peptide sequences see Supplementary Table 7 ) in a final volume of 150 µl per well for 16 h at 37°C and 5% CO_2. Both replicates – stimulated and unstimulated – were incubated in the presence of 1 µl anti-CCR6-PE/Dazzle per well. Subsequently, cells were stained by incubating them for 15 min at 37°C in 100 µl of staining buffer (PBS, 1% FBS, 1mM EDTA) containing anti-CD4-PerCP/Cy5.5 (1 µl), anti-CD45RO-FITC (1 µl), anti-ICOS-PE/Cy7 (0.2 µl), anti-CD25-PE (1 µl), anti-CD19-APC/Cy7 (1 µl), anti-CD14-APC/Cy7 (1 µl), anti-CD3-AF700 (1 µl), anti-PD1-APC (1 µl), anti-CD127-BV785 (1 µl), anti-CXCR5-BV711 (1 µl), anti-CXCR3-BV650 (1 µl), anti-CD38-BUV605 (1 µl), anti-CCR7-BV421 (1 µl), anti-HLADR-BUV737 (1 µl), anti-CD8-BUV395 (1 µl), and Zombie NIR™ fixable viability stain (0.2 µl).

Following the staining of surface markers for either innate or cTFH cells, the PBMCs were fixed in 4% PFA for 15 min at RT, washed, resuspended in 100 µl of staining buffer and acquired using an LSR Fortessa™ (BD). Data analysis was performed using the FlowJo software (v.10.7.2, Becton Dickinson, Franklin Lakes). Gating strategies are shown in Supplementary Figures 1 , 2 . Fluorescence minus one controls were used to set the gates for activation markers.

A virus neutralization test (VNT₁₀₀) was used to assess the serum neutralization capacity against wild type SARS-CoV-2 as described and reported previously (28). Following heat-inactivation for 30 min at 56°C and sample dilution in a two-fold dilution series (1:4–1:512), 100 plaque-forming units (PFU) of SARS-CoV-2 (German isolate BavPat1/2020; European Virus Archive Global #026 V-03883 (Genbank: MZ558051.1)) were added and samples were incubated for 1 h at 37°C. Subsequently, 2 × 10⁴ Vero C1008 cells (ATCC, Cat. No. CRL-1586, RRID: CVCL_0574) were added. At day 4 post infection, cytopathic effects were evaluated. Neutralization was defined as the absence of cytopathic effects. The reciprocal neutralization titer was calculated from the highest serum dilution without cytopathic effects (geometric mean based on three replicates). The lower limit of detection (LLOD) was defined as a reciprocal titer of 8, corresponding to the first dilution of the respective serum.

DEGs were identified by calculating p-values using Wald test, with adjustments for multiple comparisons using the Benjamini & Hochberg correction. Subsequently, gene expression data were analyzed using Ingenuity Pathway Analysis (v.111725566, QIAGEN). Data analysis and visualization were conducted with GraphPad Prism (v.9.5.1, Dotmatics, Boston, USA) and Rstudio (2023.06.1, R v.4.2.0, The R Foundation, Vienna, Austria). Apart from gene expression analysis, statistical significance tests were not conducted due to the limited statistical power resulting from the small sample size and the high number of parameters assessed. Correlation analyses were performed in R using Spearman’s non-parametric correlation. Descriptive statistics of all reported parameters are reported in Supplementary Tables 8 - 30 .

To gain insight into the early immune responses induced by the MVA-S and MVA-ST vaccines, as well as by the licensed mRNA and ChAd vaccines, we collected blood from the study participants of all cohorts at time points D0 (baseline), D1, D3, D7 and D14 following two vaccinations (V1, V2) against SARS-CoV-2. Whole blood transcriptome, as well as innate and cTFH cell activation were analyzed longitudinally. To identify early immune parameters potentially related to the differential immunogenicity of the MVA-S and MVA-ST vaccines, based on the same viral vector but encoding different conformations of the SARS-CoV-2 spike protein, we performed a correlation analysis between early immune parameters and adaptive immune outcomes after completion of the vaccination series (V2D14). The study design and cohorts are depicted in Figure 1 , including an overview of the different vaccination schedules, vaccine platforms and antigens, sampling time points and analyzed parameters. Baseline characteristics of study participants are listed in Supplementary Table 1 .

To profile vaccine-induced changes in the gene expression, we performed a whole blood transcriptome analysis based on PaxGene tubes collected at baseline (D0), as well as on D1, D3 and D7 following each vaccination (V1, V2). Overall, we integrated transcriptional data of 317 samples from 43 adults, vaccinated with one of four different SARS-CoV-2 vaccination regimens (MVA-S, MVA-ST, mRNA and ChAd/mRNA).

DEGs were identified after vaccination with both MVA-based vaccine candidates. However, we observed a stronger transcriptional response after the first vaccination with MVA-ST compared to MVA-S, with 75 compared to 31 DEGs on V1D1 and 42 compared to 8 DEGs on V1D7, respectively ( Figure 2A ). Also, the transcriptional response after the second vaccination with MVA-ST was stronger compared to MVA-S, with 201 compared to 2 DEGs on V2D1 and 74 compared to 0 DEGs on V2D7, respectively. In order to better understand the immune responses induced by different vaccine platforms, we compared transcriptional responses of both MVA-based vaccine candidates with the licensed ChAd and mRNA vaccines. We observed a strong transcriptional response after ChAd vaccination with 1964, 51 and 272 DEGs at V1D1, V1D3 and V1D7, respectively. In the mRNA cohort, we detected only 2 DEGs on V1D3 and 1 DEG on V1D7 following the first vaccination. These findings are similar to a study by Ryan et al., which did not detect any DEGs on D6 following mRNA vaccination (29). Following the second vaccination, numerous upregulated DEGs were detected in both the mRNA cohort (V2D1 = 653, V2D3 = 23, V2D7 = 15) and the ChAd/mRNA cohort (V2D1 = 890, V2D3 = 84, V2D7 = 39) ( Figure 2A ).

In order to identify differentially upregulated pathways, we next performed an ingenuity pathway analysis ( Figure 2B ). At V1D1 we observed an upregulation of the Interferon alpha/beta signaling and Interferon gamma signaling for both MVA-S and MVA-ST. For MVA-ST, an additional upregulation of multiple pathways of the innate immune response was observed, including the classical activation of macrophages, activation of IRF by cytosolic pattern-recognition receptors (PRRs), the cGAS-STING pathway and the human oligoadenylate synthetase (OAS) antiviral response, as well as MHC class I mediated antigen presentation. The same pathways were also upregulated at V1D1 after ChAd vaccination. In addition, ChAd vaccination induced an upregulation of the complement cascade and several pathways mediating adaptive immunity (interactions of lymphoid and non-lymphoid cells, B cell receptor (BCR) signaling and PI3K signaling in B lymphocytes). In contrast, a single mRNA vaccination did not lead to an upregulation of any of these immune pathways. While there were no pathways upregulated on V1D7 after vaccination with MVA-S or MVA-ST, an upregulation of BCR signaling was observed on V1D7 after ChAd vaccination. Following the second vaccination, we did not detect any differentially regulated pathways in the MVA-S cohort, while there was a number of upregulated pathways in the MVA-ST cohort. Besides the pathways identified at V1D1, the pathways upregulated at V2D1 in the MVA-ST cohort included the interactions between lymphoid and non-lymphoid cells, phagosome formation, neutrophil degranulation, the role of PRRs in recognition of bacteria and viruses and IL27 signaling. In addition to the upregulated pathways detected after the second MVA-ST vaccination, the pathways upregulated after the second vaccination in the ChAd/mRNA and mRNA cohorts (V2D1) included the complement cascade, PI3K signaling in B lymphocytes, IL10 and IL15 signaling, and FcγR-mediated phagocytosis. The number of DEGs involved in each of the pathways shown in Figure 2B are reported in Supplementary Tables 31 - 34 for each vaccine cohort. The shared upregulated pathways between the different cohorts and time points are additionally highlighted in the Circos plots ( Figure 2C ). The connecting lines underline that after the first vaccination pathways were mostly shared between MVA-ST and ChAd vaccination, whereas the second vaccination activated an overall higher number of pathways, which were shared between the MVA-ST, ChAd/mRNA and mRNA cohort.

To identify upstream cytokines that could potentially regulate the differential immune responses, we performed an upstream regulator analysis and filtered for predicted upstream cytokines using DEGs with FDR <0.1 and an absolute log fold change ≥1 compared to V1D0 ( Figure 2D ). We identified several members of the Interferon family as predicted upregulated upstream regulators of the observed DEG signature after the first vaccination with both MVA-S and MVA-ST, but only after the second vaccination with MVA-ST. In addition, the upstream regulators predicted for both the first and the second MVA-ST vaccination included cytokines produced by monocytes and macrophages (IL1α, IL1β, IL6, IL18), and molecules involved in the interaction of TFH cells with B cells (IL21, CD40L). For MVA-S, these upstream regulators were only predicted for the response to the first vaccination (IL1α, IL1β, IL21) or not predicted at all (IL6, IL18, CD40L) ( Figure 2E ).

Since MVA-S and MVA-ST are based on the same vector but encode different conformations of the spike protein, one explanation for the observed differences in gene expression could be the different spike protein conformations. Taking a closer look at the DEGs induced by MVA-S and MVA-ST vaccination, we identified 25 DEGs that are shared between the two cohorts at V1D1, and no shared DEGs at V2D1 ( Figure 2E ). The majority of the detected DEGs (V1D1 = 50 DEGs, V2D1 = 201 DEGs) was unique for MVA-ST vaccination. The normalized counts of six selected DEGs, that are involved in the identified upregulated pathways and showed the strongest differences between the MVA-S and MVA-ST cohort at the peak time point, are shown in Figure 2F . We observed a stronger upregulation of FCGR1A (Interactions of lymphoid and non-lymphoid cells pathway), STAT1 (CGAS STING pathway and protein kinase R (PKR) in Interferon Induction and Antiviral Response pathway), OASL (OAS antiviral response pathway), IL27 (IL27 signaling pathway), IRF7 (Interferon gamma signaling) and SOCS1 (Macrophage classical activation signaling pathway), among other genes, on V1D1 and V2D1 for MVA-ST compared to MVA-S vaccination.

Building on the observed transcriptional upregulation of innate immune pathways such as the activation of classical macrophages after MVA-ST but not after MVA-S vaccination, we investigated the monocyte-macrophage system in more detail using PBMCs derived from vaccinees at D0, D1, D3, and D7 after each vaccination. Using flow cytometry, different monocyte and DC subsets were analyzed ( Figure 3A ). While a strong response of monocytes was observed after ChAd/mRNA vaccination, including an increase in the frequency of classical monocytes at D1 (median fold change compared to V1D0 (mfc) V1D1 = 1.7, V2D1 = 1.3) and an activation of all investigated monocyte subsets at D1 and D3 as measured by the expression of CD40 [mfc up to 2.5 (V1) and 2.7 (V2)], we only detected a slight increase in the frequency of classical monocytes after the second vaccination with the MVA-based vaccine candidates (mfc MVA-S, V2D1 = 1.1; MVA-ST, V2D1 = 1.2) ( Figure 3B ). Furthermore, an activation of DCs, in particular CD16+ DCs, was observed at D3 after vaccination with ChAd/mRNA (mfc V1D3 = 2.2, mfc V2D3 = 2.4), whereas no activation was observed after vaccination with MVA-S or MVA-ST (mfc < 1.1). Consistent with the absence of transcriptional responses after the first mRNA vaccination, we also did not observe any innate responses after the first vaccination on the PBMC level. However, after the second vaccination with mRNA, an increase in the frequency of classical monocytes (mfc V2D1 = 1.9), as well as an activation of all monocyte subsets [mfc up to 1.1 (V1) and 2.1 (V2)] and CD16+ DCs (mfc V1D3 = 1.2, V2D3 = 2.3) was detected ( Figure 3B ).

Differential stimuli of the innate immune response may drive the polarization of the TFH response, which in turn impacts B cell activation, differentiation, and antibody responses. We therefore investigated whether cTFH cells influence the distinct immunogenicity outcomes of MVA-S and MVA-ST vaccination. cTFH cell responses were measured based on PBMCs collected at D0, D7 and D14 after each vaccination, following restimulation with overlapping peptide pools M1-M4, spanning the SARS-CoV-2 spike protein. The analysis included different cTFH cell subsets (cTFH1, cTFH2, cTFH17) and activation markers (ICOS, CD38) ( Figures 4A, B ). We detected a strong cTFH cell response after the first vaccination with ChAd. We further characterized this response as a cTFH1 cell response (mfc ICOS = 2.8, CD38 = 3.0), whereas no increase in frequency and activation of cTFH2 or cTFH17 cells was observed (mfc < 1.1). A single mRNA vaccination only led to a slight increase in the subset of activated cTFH1 cells in some study participants (mfc ICOS = 1.3, mfc CD38 = 1.1). Similarly, a single vaccination with MVA-S or MVA-ST only induced a small increase in activated cTFH1 cells (mfc MVA-S, ICOS = 1.3, MVA-S, CD38 = 1.1, MVA-ST, ICOS = 1.3, MVA-ST, CD38 = 1.2) ( Figures 4B, C ). Following V2 of the MVA-S, MVA-ST, mRNA and ChAd/mRNA cohort, the frequency of activated (ICOS- or CD38-expressing) cTFH1 cells was slightly increased compared to baseline (mfc MVA-S, ICOS = 1.3, MVA-S, CD38 = 1.1, MVA-ST, ICOS = 1.3, MVA-ST, CD38 = 1.4, mfc mRNA, ICOS = 1.7, mRNA, CD38 = 1.4, ChAd/mRNA, ICOS = 1.7, ChAd/mRNA, CD38 = 1.6) ( Figure 4C ).

Neutralizing antibodies, IgG and IgA against the spike protein as well as long-lasting antigen-specific B and T cell responses are crucial for protection against SARS-CoV-2 infection and disease (30–32). Vaccination with MVA-ST, encoding the prefusion-stabilized spike protein, induced increased neutralizing antibody titers compared to MVA-S, encoding the native spike protein ( Figure 5A ). Furthermore, we previously reported that vaccination with MVA-ST leads to an increased antibody, B and T cell response against the S1 subunit of the SARS-CoV-2 spike protein (11). To date, the underlying mechanisms of the differential immunogenicity outcomes following vaccination with the two MVA-based vaccines, encoding different conformations of the spike protein, are not fully understood.

In order to better understand which genes and early immune signatures are potential mediators of the increased immunogenicity of MVA-ST, we analyzed correlations of transcriptional, innate and cTFH cell responses with the neutralizing antibody responses, as well as S1/S2-specific IgG, B and T cell responses ( Figure 5B ). The correlogram includes the adaptive immune parameters after completion of the vaccination series (V2D14), and the peak time point of the previously selected DEGs (FCGR1A, STAT1, OASL, IL27, IRF7 and SOCS1), as well as the classical monocyte and cTFH1 cell frequency and activation, following the first (left panel) and second vaccination (right panel).

Following the first vaccination, the selected DEGs correlated moderately to strongly with each other (r = 0.34 to 0.84). A positive correlation was also observed between the gene expression of FCGR1A and the CD40 expression of classical monocytes (r = 0.48), while moderate negative correlations were identified between a number of DEGs and adaptive immune outcomes (IL27 vs. S2-specific IgG: r = -0.56; SOCS1 vs. nAB: r = -0.45; SOCS1 vs. S2-specific B cells: r = -0.42). cTFH1 cell activation positively correlated with spike-specific B cells (ICOS/cTFH1vs. S1-specific B cells: r = 0.51; ICOS/cTFH1vs. S2-specific B cells: r = 0.43).

Following the second vaccination, moderate to strong positive correlations (r = 0.54 to 0.94) were observed between all selected DEGs, as well as between the DEGs and the frequency of classical monocytes. A number of DEGs also correlated with activated CD38+ or ICOS+ cTFH1 cells (FCGR1A vs. CD38+/cTFH1: r = 0.41; STAT1 vs. CD38+/cTFH1: r = 0.45; FCGR1A vs. ICOS+/cTFH1: r = 0.51; IRF7 vs. ICOS+/cTFH1: r = 0.45; SOCS1 vs. ICOS+/cTFH1: r = 0.43). Furthermore, the frequency and activation of classical monocytes correlated with ICOS+ cTFH1 cells (r = 0.7 and 0.46, respectively). In turn, CD38+ and ICOS+ cTFH1 cells strongly correlated with S1-specific B cells (r = 0.82 and 0.87, respectively) and to a lower extend with S2-specific B cells (r = 0.5 and 0.76, respectively). ICOS+ cTFH1 cells also correlated with neutralizing antibodies (r = 0.67) and S1/S2-specific IgG (r = 0.51). In addition, the frequency of classical monocytes correlated moderately to strongly with neutralizing antibodies (r = 0.54), S2-specific IgG (r = 0.43), S1/S2-specific B cells (r = 0.56 and 0.71, respectively) as well as S1-specific T cells (r = 0.49). Finally, the expression of FCGR1A, IL27, STAT1, SOCS1 moderately correlated with S1-specific T cells (r = 0.68, 0.45, 0.42 and 0.59, respectively). SOCS1 gene expression correlated moderately with S1-, and FCGR1A gene expression with S2-specific B cells (r = 0.41 and 0.43, respectively). Correlations with r < |0.4| are not further discussed in this section.