Hepatitis B Vaccine Non-Responders Show Higher Frequencies of CD24highCD38high Regulatory B Cells and Lower Levels of IL-10 Expression Compared to Responders

Background The cellular mechanisms involved in the lack of protective antibody response after hepatitis B vaccination are still rather unclear. Regulatory B cells (Breg) known as modulators of B-and T-cell responses may contribute to poor vaccine responsiveness. The current study aimed to investigate the role of regulatory B cells (Breg) in hepatitis B vaccine non-responsiveness after immunization with second- or third-generation hepatitis B vaccines. Method We performed comparative phenotypic and frequency analysis of Breg subsets (CD24+CD27+ and CD24highCD38high Breg) in second-generation hepatitis B vaccine non-responders (2nd HBvac NR, n = 11) and responders (2nd HBvac R, n = 8) before (d0), on day 7 (d7), and 28 (d28) after booster vaccination. Cryopreserved peripheral blood mononuclear cells were stimulated ex vivo with a combination of CpG, PMA, and Ionomycin (CpG+P/I) and analyzed for numbers and IL-10 expression levels of Breg by flow cytometry-based analyses. Results Flow cytometry-based analyses revealed elevated frequencies of CD24+CD27+ Breg at all time points and significantly higher frequencies of CD24highCD38high Breg on d0 (p = 0.004) and 28 (p = 0.012) in 2nd HBvac NR compared to 2nd HBvac R. In parallel, we observed significantly lower levels of CpG+P/I-induced IL-10 expression levels of CD24+CD27+ and CD24highCD38high Breg (d0: p < 0.0001; d7: p = 0.0004; d28: p = 0.0003 and d0: p = 0.016; d7: p = 0.016, respectively) in 2nd HBvac NR compared to 2nd HBvac R before and after booster immunization. Frequencies of CD24+CD27+ and CD24highCD38high Breg significantly decreased after third-generation hepatitis B booster vaccination (d7: p = 0.014; d28: p = 0.032 and d7: p = 0.045, respectively), whereas IL-10 expression levels of both Breg subsets remained stable. Conclusion Here we report significantly higher frequencies of CD24highCD38high Breg in parallel with significantly lower IL-10 expression levels of CD24+CD27+ and CD24highCD38high Breg in 2nd HBvac NR compared to 2nd HBvac R. Anti-HBs seroconversion accompanied by a decrease of Breg numbers after booster immunization with a third-generation hepatitis B vaccine could indicate a positive effect of third-generation hepatitis B vaccines on Breg-mediated immunomodulation in hepatitis B vaccine non-responders.

Background: The cellular mechanisms involved in the lack of protective antibody response after hepatitis B vaccination are still rather unclear. Regulatory B cells (Breg) known as modulators of B-and T-cell responses may contribute to poor vaccine responsiveness. The current study aimed to investigate the role of regulatory B cells (Breg) in hepatitis B vaccine non-responsiveness after immunization with second-or thirdgeneration hepatitis B vaccines.
Method: We performed comparative phenotypic and frequency analysis of Breg subsets (CD24 + CD27 + and CD24 high CD38 high Breg) in second-generation hepatitis B vaccine non-responders (2 nd HBvac NR, n = 11) and responders (2 nd HBvac R, n = 8) before (d0), on day 7 (d7), and 28 (d28) after booster vaccination. Cryopreserved peripheral blood mononuclear cells were stimulated ex vivo with a combination of CpG, PMA, and Ionomycin (CpG+P/I) and analyzed for numbers and IL-10 expression levels of Breg by flow cytometry-based analyses.

INTRODUCTION
According to the WHO 260 million people are chronically infected with HBV and 887,000 people are dying each year due to hepatitis B virus (HBV) infection (1). Second-generation hepatitis B vaccines composed of the small HBV envelope protein (hepatitis B surface antigen; HBsAg) are currently used for universal vaccination and reduce the overall incidence of both hepatitis B and the associated long-term consequences such as chronic hepatitis B and liver cirrhosis (2,3).
Vaccination with recombinant HBsAg triggers the production of anti-HBs with an anti-HBs titer of > 10 IU/L being a very reliable surrogate marker for vaccine-induced protective immunity (4). Approximately 5-10% of vaccinees are defined as "non-responders", i.e. they do not develop a protective anti-HBs titer after completing a full primary series of the hepatitis B vaccine (5)(6)(7)(8). Third-generation hepatitis B vaccines containing additional HBV envelope proteins (pre-S1 and pre-S2) are known to be more immunogenic and superior in inducing protective antibody titers also in non-responders to the conventional vaccines (9). Despite the success of universal immunization programs leading to high regional vaccination coverage rates in most Western countries, non-responsiveness to hepatitis B vaccination is a major problem, especially for health care workers for whom successful hepatitis B vaccination is mandatory (10,11). Genetically determined resistance, advanced age, gender, obesity, smoking, chronic/systemic disease, and immunosuppressive therapies are known factors contributing to non-response to immunization (6,(12)(13)(14)(15)(16)(17)(18).
Several aspects like failure of antigen presentation or costimulatory signals, defects in the generation of HBsAgspecific CD4 T helper (Th) cells and insufficient production of Th1 and Th2 cytokines upon hepatitis B vaccination have been discussed, but the exact underlying immunological and molecular mechanisms contributing to hepatitis B vaccine non-responsiveness remain largely unclear (10,(19)(20)(21)(22)(23)(24). Further studies investigating immune cells and immunological mechanisms involved in non-responsiveness to vaccines are urgently needed and will help to improve immunogenicity of existing or the development of new vaccines to overcome nonresponsiveness to hepatitis B vaccines.
In the last years, studies revealed that regulatory B cells (Breg) have an immunosuppressive capacity and help to maintain immunological homeostasis (25,26). Breg suppress immunopathology by skewing T-cell differentiation, induction and maintenance of regulatory T cells, as well as suppression of pro-inflammatory cells, mainly mediated through regulatory cytokines IL-10, TGF-ß, and IL-35 (25)(26)(27)(28)(29). There is an ongoing debate on the phenotypic characterization of different Breg subsets, specific Breg markers, and the question whether all B cells can acquire suppressive function in response to environmental triggers, or whether Breg represent a distinct lineage (28,(30)(31)(32). CD24 +/high CD27 + and CD24 high CD38 high Breg have been described as distinct Breg subpopulations and investigated in different disease entities like autoimmune diseases (e.g. multiple sclerosis, rheumatoid arthritis, systemic lupus erythematosus, etc.) and cancer (25,30,(33)(34)(35). For both, CD24 +/high CD27 + and CD24 high CD38 high Breg it is known that they can suppress effector CD4 T cells and dendritic cells, and CD24 high CD38 high Breg can additionally induce regulatory T cells (Treg) and suppress virus-specific CD8 T cells (30,33,36). Rosser et al. (37) postulated three potential mechanism for Breg-mediated suppression of antibody responses, which could also play a role in the case of hepatitis B vaccine nonresponsiveness: (i) Breg may alter the cytokine microenvironment in which plasma cell maturation takes place; (ii) Breg could suppress CD4 Th cells, which leads to a diminished maturation of B cells into antibody producing plasma cells; (iii) Breg could induce Treg which may contribute to an indirect suppression of antibody production (25,37).
In the last years, some studies investigated the role of Breg in hepatitis B vaccine non-responsiveness, but findings were contradictory and in-depth analyses remain elusive (12,38,39 (38). In contrast, Garner-Spitzer et al. reported clearly elevated frequencies of CD24 high CD38 high Breg in hepatitis B vaccine non-responders which might contribute to the increased baseline levels of IL-10 in these individuals and also lead to an induction of regulatory T cells (39).
The current study aimed to investigate different Breg subpopulations in hepatitis B vaccine responders and nonresponders before and after booster vaccination with a secondversus a third-generation hepatitis B vaccine.

Study Cohorts
Two groups of hepatitis B vaccinated individuals were enrolled in this study. The first group comprised eleven non-responders to second-generation hepatitis B vaccine (2 nd HBvac NR) (9 women, 2 men; average age of 25.7 years) ( Table 1). All 2 nd HBvac NR subjects received a full primary series of hepatitis B vaccine and were repeatedly vaccinated with Engerix ® or Twinrix ® in the past, without developing protective anti-HBs levels (> 10 IU/L). The second group consists of eight responders to second-generation hepatitis B vaccine (2 nd HBvac R) (6 women, 2 men; average age: 30.0 years), who received a full series of Twinrix ® vaccination more than ten years in the past ( Table 1). On day 0, group 1 (2 nd HBvac NR) received a booster vaccination with the third-generation recombinant hepatitis B vaccine Sci-B-Vac ™ (VBI Vaccines Inc., Rehovot, Israel), whereas group 2 (2 nd HBvac R) was revaccinated with the second-generation recombinant hepatitis B vaccine Twinrix ® (Glaxo Smith Kline, Brentford, UK). Peripheral blood was taken by venipuncture at baseline (day 0), and on day 7, and 28 after vaccination with the approval of the local ethic committees (School of Medicine, Technical University of Munich and Innsbruck Medical University). Data on age, weight, and height (BMI (body mass index) measurement), smoking status, alcohol consumption, medical co-morbidities, and medication were collected at baseline ( Table 1). Informed consent was obtained from all participating individuals prior to their inclusion.

Determination of Serum Anti-HBs Levels
Serum levels of HBsAg-specific antibodies (anti-HBs) were quantified using the Architect ® chemiluminescence microparticle immunoassay (Abbott Diagnostics, Wiesbaden, Germany). The detection limit was 10 IU/mL.

Isolation and Cryoconservation of PBMC
Within 4 h after collection of heparinized whole blood human peripheral blood mononuclear cells (PBMC) were separated by Ficoll density gradient (human Pancoll, PAN-BIOTECH, Aidenbach, Germany) as described previously (40). PBMC were frozen in aliquots of 5 x 10 6 PBMC per vial in 1.8 mL cryotubes (Thermo Scientific, Roskilde, Denmark) in a concentration of 1 x 10 7 PBMC per 1 mL freezing medium (fetal calf serum (FCS) (Life Technologies, Darmstadt, Germany) supplemented with 10% DMSO (Sigma-Aldrich, Steinheim, Germany) in a freezing container (Mr. Frosty, Thermo Scientific, Roskilde, Denmark) and put on -80°C. After 48 h PBMC were stored in the vapor phase of a liquid nitrogen tank until further use.

Thawing and Resting of PBMC
PBMC were thawed at 37°C using CTL Anti-Aggregate Wash ™ 20x Solution (Cellular Technology Limited (CTL) Europe, Bonn, Germany) diluted in RPMI-1640 medium (Life Technologies, Darmstadt, Germany) (1:20). Cells were counted with an automated cell counter (Vi-cell XR, Beckman Coulter, Krefeld, Germany) in CTL-Test ™ Medium (Cellular Technology Limited (CTL) Europe, Bonn, Germany). The median cell recovery after thawing was 4.4 x 10 6 PBMC per vial with a median viability of 92%. For a standard resting procedure PBMC were incubated for 18 h at 37°C in a humidified atmosphere at 5% CO 2 in a concentration of 2 x 10 6 PBMC/mL CTL-Test ™ medium supplemented with 10% FCS and 1% penicillin-streptomycin (PenStrep, Life Technologies, Invitrogen, Darmstadt, Germany) (abbr.: RPMI-10) RPMI-10. The median cell recovery after resting was 4.2 x 10 6 PBMC per vial with a median viability of 94%.

Data Acquisition
Acquisition of samples was performed within 4 h after staining using a BD LSR Fortessa flow cytometer (Becton Dickinson, Franklin Lakes, USA) equipped with a 96-well plate reader and FACSDiva Software V.6.0 (Becton Dickinson, Heidelberg, Germany). On a weekly basis, the flow cytometer's performance was checked and settings were configured with Cytometer, Setup & Tracking beads (Becton Dickinson, Franklin Lakes, USA). Photomultiplier voltages were adjusted with the help of unstained cells for all parameters.

Gating Strategy
Flow cytometry-based analysis was performed on at least 1.0 x 10 5 living lymphocytes using the software FlowJo version 10 (FlowJo LLC, BD, Ashland, USA). Gating strategy for analysis of ex vivo re-stimulated PBMC is shown in the supplementary information (Additional file 2: Figure S1). Each gate was set in the negative control sample and then adjusted to the antigen stimulated samples. In detail, we firstly set a broad forward-side scatter gate to prevent excluding cells of interest, followed by an exclusion of dead cells. Using a FSC-W against FSC-A plot, we excluded doublets. Next, we gated on CD3 -CD14cells and CD19 + B cells. Within the CD19 + cells, we gated on CD24 + CD27 + and CD24 high CD38 high Breg, respectively. FMO controls were used to gate on CD24 + , CD27 + , and CD38 + cells, respectively. For functional analysis, we gated on IL-10, IL-35, and TGF-ß expression in the CD19 + B cell, CD24 + CD27 + , and CD24 high CD38 high Breg population. Two independent audits were performed to control the gating.

Data Interpretation
We detected significantly higher levels of IL-10, IL-35, and TGF-ß expression of CD24 + CD27 + and/or CD24 high CD38 high Breg upon CpG+P/I re-stimulation compared to unstimulated PBMC in both cohorts. In line with this, we used CpG+P/I induced frequencies for further analysis without further background subtraction. Detected frequencies of cytokine expressing Breg which derived from rare events (< 20 events) were excluded from further analysis to avoid misleading interpretations. The median fluorescence intensity (MFI) of CD24 + CD27 + IL-10 + , CD24 + CD27 + IL-35 + , CD24 + CD27 + TGF-ß + and CD24 high CD38 high IL-10 + Breg of second-generation hepatitis B vaccine non-responders and responders before (day 0) and 7, and 28 days after vaccination with a third-generation hepatitis B vaccine are shown in Additional file 3: Figure S2.

Statistical Analysis
All results were included in the analysis, as no attempt was made to exclude outliers. All tests were two-sided and conducted on exploratory 5% significance levels. Effect measures are presented with 95% confidence intervals. Nonparametric statistical tests were applied in all cases. Unpaired Mann-Whitney U test was used to define significance of values between the 2 nd HBvac NR and 2 nd HBvac R group. One-Way ANOVA (Friedman test) was applied for analyses within the two groups for the different time points. In case of resulting p-values < 0.05, paired Wilcoxon signed rank tests were performed to assess significance of change in values within the 2 nd HBvac NR and 2 nd HBvac R group, respectively. The software Graph Pad Prism 9.1.0 (GraphPad Software, La Jolla, California, USA) was used for statistical analyses. Interpretation of Spearman`s correlation coefficient was performed according to Cohen (41).
The only one remaining non-responder from study group 1 (2 nd and 3 rd HBvac NR) had risk factors (smoking, obesity, hypertension) known to impair vaccine efficacy.

Vaccine Non-Responders Showed Higher Numbers of CD24 high CD38 high Breg And Lower Breg-Derived IL-10 Expression Compared to Vaccine Responders
High levels of immunomodulatory lymphocytes and related cytokines such as regulatory B cells (Breg) and IL-10 have been associated with serological non-response to hepatitis B vaccination. To investigate this, we analyzed frequencies of two populations of Breg (CD24 + CD27 + and CD24 high CD38 high Breg) and respective IL-10 levels in a cohort of 2 nd HBvac NR (n = 11) at baseline visit and on day 7 and 28 post Sci-B-Vac booster vaccination. For comparative analysis, we additionally determined frequencies of Breg and IL-10 expression levels on day 0, 7, and 28 also in a cohort of second-generation hepatitis B vaccine responders (referred to as "2 nd HBvac R"; n = 8) receiving a second-generation hepatitis B booster vaccination.
Frequencies of CD24 + CD27 + Breg tend to be higher in 2 nd HBvac NR compared to 2 nd HBvac R at all three time points (d0: p = 0.062; d7: p = 0.059; d28: p = 0.109) (Figure 2A). In parallel, w e o b s e r v e d s i g n i fi c a n t l y h i g h e r f r e q u e n c i e s o f CD24 high CD38 high Breg in the 2 nd HBvac NR compared to the 2 nd HBvac R group on day 0 (p = 0.004) and day 28 (p = 0.012), but not on day 7 (p = 0.051) ( Figure 2B and Table 2).
Besides IL-10, other cytokines like IL-35 and TGF-ß could mediate immunosuppressive function of Breg. CD24 + CD27 + IL-35 + Breg were detectable in nine 2 nd HBvac NR and three 2 nd HBvac R ( Figure 3A). Median frequencies of CD24 + CD27 + IL-35 + Breg were comparable between both groups and time points, but reactivity rates in the 2 nd HBvac R group were too low for statistical analysis ( Figure 3A and Table 1). CD24 high CD38 high IL-35 + Breg were largely undetectable in both groups (1/11 2 nd HBvac NR and 0/8 2 nd HBvac R; data not shown). Regarding TGF-ß expression, we detected CD24 + CD27 + TGF-ß + Breg in ten 2 nd HBvac NR but only one 2 nd HBvac R ( Figure 3B and Table 2). CD24 high CD38 high TGFß + Breg were detectable in six 2 nd HBvac NR subjects with comparable frequencies at all time points, but in none of the 2 nd HBvac R subjects ( Table 2).
Correlation analysis revealed no correlation of prevaccination levels of CD24 + CD27 + IL-10 + Breg and anti-HBs levels on day 28 (r s = 0.083), but a moderate correlation of pre-vaccination levels of CD24 high CD38 high IL-10 + Breg and anti-HBs levels on day 28 (r s = 0.300) (Additional file 4: Figures  S3A, B).

Frequencies of CD24 + CD27 + and CD24 high CD38 high Breg Decreased Upon Booster Vaccination With Third-Generation Vaccine but Remained Stable Upon Second-Generation Vaccine Booster Immunization
Next, we assessed the impact of booster vaccination with second-(2 nd HBvac R group) and third-(2 nd HBvac NR group) generation hepatitis B vaccines on Breg and B10 + cell frequencies as well as Breg-derived IL-10 expression Upon booster vaccination with the third-generation vaccine (2 nd HBvac NR group), we observed significantly higher frequencies of CD24 + CD27 + Breg prior (d0) compared to day 7 (p = 0.014) and day 28 (p = 0.032) post vaccination, but no difference between day 7 and 28 (p = 0.123) (Figure 2A). Frequencies of CD24 high CD38 high Breg were also significantly higher prior vaccination (d0) compared to day 7 post vaccination (p = 0.045), but comparable between the other time points (d0 vs. d28: p = 0.175; d7 vs. d28: p = 0.520) ( Figure 2B).
We also observed no significant differences in frequencies of B10 + cells between the different booster vaccines and time points (2 n d HBvac NR: p = 0.256 and 2 n d HBvac R: p = 0.967) (Figure 4).

Third-Generation Hepatitis B Vaccine
Low-and High-Responders Showed Different Frequencies of CD24 + CD27 + And CD24 high CD38 high Breg Before and One Week After Booster Vaccination Finally, we analyzed differences in frequencies of Breg and B10 + cells between third-generation hepatitis B vaccine low- (3 rd HBvac LR) and high-responders (3 rd HBvac HR) (n = 5, respectively). Frequencies of CD24 + CD27 + Breg were significantly higher in the 3 rd HBvac LR compared to the 3 rd HBvac HR group on day 0 and day 7, but not on day 28 (d0: p = 0.032; d7: p = 0.024, and d28: p = 0.548, respectively) ( Figure 5A   and Table 3). In contrast, we observed a tendency of lower frequencies of CD24 high CD38 high Breg in the 3 rd HBvac LR compared to the HR group at all three time points (p = 0.095, respectively) ( Figure 5B and Table 3).

DISCUSSION
Although several potential immunological mechanisms associated with hepatitis B vaccine non-responsiveness, like failure of antigen presentation or costimulatory signals, impact of certain HLA class II alleles, or lack of specific CD4 Th cells have been investigated, the decisive underlying immunological mechanisms remain unclear (10,19,20,39,43). In the current study, we asked whether non-responsiveness to hepatitis B vaccine is associated with a dysregulation of certain Breg subpopulations and analyzed their frequency and function in relation to secondversus third-generation hepatitis B vaccine induced immunity Breg known to mediate immunosuppressive functions and the effect of different hepatitis B vaccine formulations on their frequency and function have not yet been investigated in parallel as possible factors involved in vaccine non-responsiveness. We performed comparative phenotypic and frequency analysis of CD24 + CD27 + and CD24 high CD38 high Breg in 2 nd in HBvac NR and R before and after booster vaccination to investigate whether non-responsiveness to hepatitis B vaccination is associated with alterations of Breg frequencies and their cytokine expression levels.
One of our main findings was the detection of higher frequencies of CD24 + CD27 + and CD24 high CD38 high Breg accompanied by lower levels of IL-10 expression in 2 nd HBvac NR compared to R. So far, very few studies investigated the association of Breg frequencies and IL-10 expression levels with non-responsiveness to hepatitis B vaccination (12,38,39). In contrast to our results, Bolther et al. observed no significant differences in frequencies of CD24 high CD38 high Breg between 2 nd HBvac non-/low-responder and high-responder (38). But since the authors combined non-and low-responder in one group, the finding of comparable frequencies of CD24 high CD38 high Breg in non-/low versus high-responder group could be biased by the very low number of non-responder subject included (38). Correlation analysis revealed no significant correlation of anti-HBs levels and frequencies of IL-10 + CD24 high CD38 high Breg, which is in line with our results (38).
In a study cohort of hepatitis B and tick-borne encephalitis (TBE) vaccine non-responders Garner-Spitzer et al. investigated, whether non-responsiveness is an antigen and/or vaccinespecific phenomenon. In line with our observations, Garner-Spitzer et al. reported clearly elevated frequencies of CD24 high CD38 high Breg in HBvac NR pre-and post-booster TBE vaccination compared to TBE non-or high-responders (39). Since HBvac NR developed sufficient anti-TBE titers (neutralization test (NT) titers ≥ 1:10) and Breg frequencies did not decline post TBE booster vaccination, they concluded that the underlying immunological mechanisms of nonresponsiveness rather depend on the applied vaccine antigen and host-related genetic predispositions (39).
Interestingly, TBE non-and high responders had comparable median frequencies of CD24 high CD38 high Breg pre-booster which only increased in the non-responder group after TBE booster vaccination. In our study, we observed a decline of both, CD24 + CD27 + and CD24 high CD38 high Breg frequencies, post third-generation booster vaccination in 2 nd HBvac NR, which could indicate a positive effect of third-generation hepatitis B vaccines on Breg-mediated immunomodulation in HBvac NR. Interestingly we did not observe decreased numbers of CD24 + CD27 + and CD24 high CD38 high Breg post-boost with the second-generation hepatitis B vaccine.
Breg modulate immune responses predominantly, although not exclusively, via IL-10, therefore IL-10 expression is often used as a marker for Breg function (25). IL-10, a pleiotropic cytokine, is known to act as an immunoregulatory molecule which inhibits the production of inflammatory cytokines by T cells and monocytes and suppresses antigen-presentation e.g. by downregulation of MHC II expression (12,44), but it is also known to promote B-cell differentiation into antibody secreting plasmablasts (12,(44)(45)(46). Several subsets of B cells produce IL-10, including B10 + cells, CD24 + CD27 + and CD24 high CD38 high Breg (30,34). Within our study cohorts, we observed significantly lower IL-10 expression levels of CD24 + CD27 + and CD24 high CD38 high Breg and CD19 + B10 + cells in HBvac NR compared to HBvac R pre-booster. Heine et al. reported that genes associated with the differentiation into antibody-secreting cells are upregulated in IL-10-secreting B cells and concluded that autocrine and paracrine IL-10 signaling contributes to differentiation of B cells into antibody secreting cells (45). Therefore, low IL-10 expression levels may additionally favor non-responsiveness to hepatitis B vaccination.
The fact, that we observed significantly lower IL-10 expression levels of CD24 + CD27 + and CD24 high CD38 high Breg and B10 + cells in 2 nd HBvac NR support the hypothesis that IL-10 may contribute to B-cell differentiation into anti-HBs secreting plasma cells (12). Of note, also the HBvac NR who was the only individual without protective anti-HBs titer even after booster with third-generation hepatitis B vaccine had very low frequencies of IL-10 expressing CD24 + CD27 + and CD24 high CD38 high Breg and B10 + cells. However, we could not observe a significant increase in IL-10 expression levels of CD24 + CD27 + and CD24 high CD38 high Breg in 2 nd HBvac NR who seroconverted after third-generation booster vaccination.
Although the immunoregulatory role of Breg is predominantly driven by IL-10, other cytokines, like IL-35 and TGF-ß are known to impact responsiveness to hepatitis B vaccination (10,(26)(27)(28)(29). Breg-derived TGF-b can lead to an expansion of Treg linked with impaired antibody production (26,35). In line with reports of Jarrosson et al. (10), we observed robust frequencies of IL-35 and TGF-ß expressing CD24 + CD27 + Breg in the 2 nd HBvac NR whereas in 2 nd HBvac R group, these Breg were mostly undetectable.
Another important result of our study was the high seroconversion rate of HBvac NR after booster vaccination with a third-generation hepatitis B vaccine which was also observed in previous studies, using Sci-B-Vac ™ (9, 47-49). Commonly used hepatitis B vaccines belong to the secondgeneration vaccines and are composed of the yeast-derived recombinant non-glycosylated small (S) HBV envelope protein (HBsAg) (3), whereas third-generation hepatitis B vaccines like Sci-B-Vac ™ , consist of glycosylated and non-glycosylated pre-S1, pre-S2, and S proteins produced in mammalian cells (48). Differences in glycosylation patterns and physical properties of these vaccine antigens might result in higher immunogenicity and successful induction of protective anti-HBs titers (9,47,(49)(50)(51). The exact underlying immunological mechanisms of higher anti-HBs seroconversion rates in formerly non-responders are still unclear, but it is assumed that the additional pre-S1 and pre-S2 domains increase immunogenicity and may evade genetic non-responsiveness to the S antigen (52,53).
As postulated by Garner-Spitzer et al., non-responder status does probably not reflect a basic defect in antibody production, as HBV-NR respond well to TBE and influenza vaccine. Our data support this hypothesis since ten of eleven 2 nd HBvac NR showed a seroconversion after vaccination with Sci-B-Vac ™ . There are reports suggesting that a real hepatitis B non-responder status is quite rare and that most non-responders are indeed lowresponders where higher vaccine doses, intradermal vaccine application or repeated vaccinations could overcome the NR status (54,55). At least for our study cohort this does not apply, as most study participants already received several booster vaccinations with second-generation vaccines without showing seroconversion. The fact, that vaccination with a thirdgeneration hepatitis B vaccine lead to anti-HBs seroconversion in more than 90% of our study participants also argues against an overall defect in the priming and generation of vaccine-induced memory B cells. Valats et al. reported that 2 nd HBvac NR showed substantial numbers of HBsAg-specific memory B cells able to differentiate into anti-HBs secreting plasma cells upon in vitro re-stimulation (54). Since production of anti-HBs is known to be Th cell-dependent (19,(56)(57)(58)(59) also a dysfunction in T-cell help mediating hepatitis B vaccine non-responsiveness was discussed (60,61). But this is contradicted by the fact that several studies reported vaccine-induced, HBs-Ag-specific T cells being detectable in second-generation hepatitis B vaccine nonresponders (10, 62). We have not investigated Th cell function or immunoregulatory T-cell subsets in our study cohorts as our main focus was the possible involvement of Breg on vaccine nonresponsiveness. It remains to be investigated whether and how Breg-mediated effects described here interact with other immunoregulatory mechanisms (e.g. a vaccine-induced, dysfunctional T-cell response) which might synergistically favor hepatitis B vaccine non-responsiveness.
In summary, we report significantly higher frequencies of CD24 high CD38 high Breg in parallel with significantly lower IL-10 expression levels of CD24 + CD27 + and CD24 high CD38 high Breg in 2 nd HBvac NR compared to 2 nd HBvac R. Anti-HBs seroconversion accompanied by a decrease of Breg numbers after booster immunization with a third-generation hepatitis B vaccine could indicate a positive effect of third-generation hepatitis B vaccines on Breg-mediated immunomodulation in hepatitis B vaccine non-responders. Further studies investigating Breg-mediated mechanisms involved in non-responsiveness to vaccines also during primary series of hepatitis B vaccination may help to further improve immunogenicity of existing or the development of new vaccines to overcome non-responsiveness to hepatitis B vaccines.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by ethic committees: School of Medicine, Technical University of Munich and Innsbruck Medical University. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
NK, TB, and UP conceived and designed the experiments. BW and BG-L provided blood samples. NK and LP performed the experiments and acquired the data. NK, LP, and TB analyzed the data. NK, LP, BW, BG-L, PK, HR, UP, and TB interpreted the data. NK, LP, and TB drafted the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This study was supported by Helmholtz Initiative for Personalized Medicine (iMed, project title: "Molecular basis and early predictors of non-responsiveness to hepatitis B vaccination").