The TLR2 ligands Pam2Cys-Sk4, Pam3Cys-SK4, and CL429 were acquired from EMC Microcollections, Germany. The TLR3 ligand polyinosinic-polycytidylic acid (poly I:C) was purchased from Sigma-Aldrich, Switzerland. The TLR4 ligands Kdo2-Lipid A, monophosphoryl lipid A (MPLA), and lipid A detoxified were purchased from Avanti Polar Lipids, USA. The TLR4 ligand LPS (E. coli-derived) was acquired from Sigma-Aldrich. The TLR5 ligand Flagellin (FLA-ST, S. typhimurium-derived) was purchased from Invivogen, France. The TLR7 ligands loxoribine and gardiquimod, as well as the TLR7/8 ligand resiquimod were acquired from Chemdea (Ridgewood, NJ, USA), and the TLR7 ligand imiquimod was purchased from Invivogen. The TLR9 ligands CpG oligodinucleotide (ODN) D32 (5′-ggT GCG TCG ACG CAG ggg gg-3′), CpG 2007 (5′- tcg tcg ttg tcg ttt tgt cgt t -3′), and CpG 2142 (5′-TCGCGTGCGTTTTGTCGTTTTGACGTT-3′) were ordered from Eurofins, Switzerland. ODN D-SL01 (5′-tcgcgacgttcgcccgacgttcggta-3′) and ODN D-SL03 (5′-tcgcgaacgttcgccgcgttcgaacgcgg-3′) were purchased from Invivogen or Eurofins, Switzerland. The positive control for proliferation concanavalin A (ConA) was acquired from Sigma-Aldrich.

CellTrace Violet stain was purchased from Thermofisher Scientific.

Monoclonal antibodies used included CD4 (clone 74-12-4; ATCC), CD14 (clone MIL2; Biorad, USA), CD16 (clone G7, Biorad), CD21 (clone BB6-11C9.6; Southern Biotech, USA), CD40 (clone G28.5, ATCC, USA), CD79a-APC (clone HM57, BioLegend, USA), CD80/86 (detection through human cytotoxic T-lymphocyte associated antigen-mouse immunoglobulin fusion protein; Enzo Biochem, USA), CD115 (clone ROS8G11-1; kindly provided by Prof. David Hume, Roslin Institute, University of Edinburgh, UK), CD172a (clone 74-22-15A, hybridoma kindly provided by Prof. Armin Saalmüller, Veterinary University, Vienna, Austria), CD172a-PE (Southern Biotech, USA), anti-IgM (clone PIG45A, ATCC), anti-IgG (23.7.1b, Thermofisher Scientific, Switzerland), anti-CD25 (clone K231.3B3, Biorad), and CADM1 (clone 3E1, MBL International, USA). Anti-TLR2 antibodies were kindly provided by Dr. Javier Dominguez (Departemento de Biotecnología, Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria, INIA, Madrid, Spain). Recombinant porcine IL-3 his-tagged (IL-3-his) was produced as described (25). Anti-his-tag antibody was purchased from Roche, Switzerland. Phosphorylated phospholipase (PLC)-γ2 was detected using Alexa Fluor^® 647 mouse anti-PLC-γ2 (pY759) (clone K86-689.37, Becton Dickinson). Anti-mouse secondary antibodies were purchased from Invitrogen (AF488, AF647, and PE-Cy5.5), Southern Biotech (FITC and PE), Abcam (PE-Cy7), and BD (V500 and V421 conjugates).

Citrated blood was taken from a total of eight different pigs housed in our specific pathogen-free facility. Blood withdrawal was performed under the guidance of the Swiss animal welfare law, approved by the Cantonal ethical committee for animal experiments (BE88/14 and BE90/14). Peripheral blood mononuclear cells (PBMCs) were prepared using density gradient centrifugation (1,000 × g, 25 min) over Ficoll-Paque (1.077 g/L, GE Healthcare Life Science). CD21⁺ B cells were isolated using magnetic cells sorting (MACS), following the manufacturer’s instructions (Miltenyi Biotec, Germany). PBMCs were cultured in DMEM + GlutaMAX (Thermofisher, Switzerland) supplemented with 10% fetal bovine serum (Thermofisher) and 0.02 mM beta-mercaptoethanol (Sigma-Aldrich) at 39°C and 5% CO₂. CD21-sorted B cells were cultured in medium and conditions as described above, supplemented with HEPES, non-essential amino acids, sodium pyruvate (Thermofisher), and 50 U/ml recombinant porcine interleukin-2 (IL-2, kindly provided by Dr. S. Inumaru, National Institute of Animal Health, Ibaraki, Japan).

Peripheral blood mononuclear cells lacking plasmacytoid dendritic cells (pDCs) were obtained by depleting IL-3 receptor expressing cells as described (26). Briefly, PBMCs were incubated with porcine IL-3-his followed by anti-His-IgG (Roche), anti-IgG-microbeads (Miltenyi Biotec). All incubations lasted for 15 min on ice. After washing, cells were passed through an LD column and the negative fraction was collected on ice.

Purified B-cell subsets were obtained using FACS sorting. After PBMC preparation, T cells were depleted using the MACS system (Miltenyi Biotec). After incubation with anti-CD3 (clone PTT3/FyH2), cells were incubated with anti-mouse IgG magnetic beads (Miltenyi Biotec) and sorted using an LD column. The CD3⁻ fraction was then stained with anti-CD21, IgM, IgG, and CD11R1 as described in the Section “Results”. After doublet exclusion and lymphocyte gating, CD21⁺IgM⁺ (P1), IgG⁺ (P2), CD21⁻IgM⁺CD11R1⁺ (P3), and CD21⁻IgM⁺CD11R1⁻ (P4) cell subsets were defined by electronic gating and the four subsets sorted using a fluorescence activated cell sorting (FACS) Aria (BD Biosciences).

Peripheral blood mononuclear cells were seeded at 1 × 10⁶ cells/500 μl and stimulated with TLR ligands for 17 h. Then, cells were harvested and stained for B cell subsets using anti-IgM, -IgG, CD21, and CD11R1, combined with CD25 or CCR7 as activation markers.

To quantify tonic B cell signaling, freshly isolated PBMC were first stained for surface markers, incubated for 2 min, then fixed and permeabilized (BD Phosphoflow^TM reagents) and stained for PLC-γ2 (pY759).

Peripheral blood mononuclear cells were cultured in medium as described above at a concentration of 2 × 10⁶ PBMCs/ml in round-bottom microtiter plates They were stimulated with TLR ligands at the following concentrations: 10 µg/ml for TLR2 ligands, 1 µg/ml for TLR4 ligands, 1 µg/ml for TLR5 ligands, 5 µg/ml for TLR7/8 ligands, and 5 µg/ml for TLR9 ligands. The cells were incubated for 2 days, and 1 μCi ³H-thymidine per well was added during the final 18–24 h. Cells were harvested on glass fiber filters and radioactivity was quantified in a Microbeta 1640 Trilux counter (Perkin Elmer).

To identify proliferating cell populations, we used CellTrace Violet stain and multicolor flow cytometry. Fifty million PBMCs or enriched B cells were resuspended in 1 ml PBS. CellTrace Violet was added followed by incubation at 37°C for 10 min. Next, 2 ml FBS was added to the suspension and it incubated at room temperature in the dark for 15 min. Cells were washed twice with PBS/EDTA and centrifuged at 350 × g at room temperature for 10 min. Cells were then seeded into round-bottom 96-well plates at 200,000 cells/well in 200 µl final volume, with TLR ligands at the concentrations described above. After incubation at 39°C/5% CO₂ for 5 days, cells were stained with primary and secondary antibodies for B cell subsets corresponding to the desired read-out. IgG block (Jackson Immunoresearch, USA) was performed before adding primary antibodies when using enriched B cells.

Peripheral blood mononuclear cells or purified B cell subsets were cultured for 5–7 days culture at 39°C/5% CO₂ at the conditions indicated in the figure legends, and supernatants were harvested and frozen until analysis. In some cultures, 50 U/ml recombinant porcine IL-2 (kindly provided by Dr. S. Inumaru, National Institute of Animal Health, Ibaraki, Japan) and 10 ng/ml recombinant porcine B-cell activating factor [BAFF, prepared as previously described (27)] were added. Nunc-Immuno 96-well plates (Sigma-Aldrich) were coated with anti-IgM antibody in PBS (clone 5C9, 1:200). After overnight incubation at room temperature, plates were washed three times with wash buffer (PBS + 0.05% Tween 20) and blocked with PBS/0.5% BSA/0.05% Tween 20 buffer at 37°C for 1 h. After washing, samples were transferred and plates incubated at 37°C for 2 h. Next, plates were washed three times and we added goat anti-pig detection antibody coupled with horseradish peroxidase (Bethyl, A100-117P, 1:20,000) for 20 min at 37°C. After washing, the substrate OPD (Sigma-Aldrich) was added and absorbance was measured at 450 nm using VersaMax reader (Molecular Devices, USA).

Two pigs were vaccinated with a commercial vaccine against FMDV A Iran 96 (kindly provided by Merial, Pirbright, UK) using a prime boost vaccination protocol with 4 weeks between injections. PBMCs from these animals were used 3–7 months after booster vaccination. Cells were cultured in 24-well plates at a concentration of 2 × 10⁶ cells/well and stimulated with purified FMDV antigen (10 µg/ml 146S antigen derived from A Iran 96, kindly provided by Merial) and/or TLR ligands, and incubated for 7 days at 39°C, 5% CO₂.

FMDV-specific antibodies were detected by ELISA. Plates were coated with 100 µl 1 µg/ml FMDV A Iran 146 S antigen in PBS and incubated over night at 4°C. After washing with PBS, the plates were blocked with 1% BSA in PBS for 1 h at room temperature. Then, samples were applied and incubated for 30 min at room temperature. After washing the plates with PBS, peroxidase-conjugated goat anti-swine IgG (Jackson ImmunoResearch, PA, USA) followed by the addition of TMB as substrate.

B cells and monocytes were enriched with MACS using CD21 and CD14 antibodies, respectively. Purified pDCs were obtained using fluorescence activated cell sorting (FACSAria, Becton Dickinson) based on their CD172a⁺CD4^high phenotype (28). The purity of sorted cells was over 98%. The purified B cell subsets P1–P4 were also obtained by FACS sorting as described above. After sorting, cells were washed and resuspended in TriZOL for RNA extraction using an RNeasy mini kit (Qiagen, Switzerland). RNA was treated with RNase-free DNase I (Ambion, UK). RT-PCR was performed on a Thermocycler ABI 7500 (ThermoFischer). RNA expression levels were calculated based on 18S RNA levels. TLR2 forward primer, 5′-TATCCAGCACGAGAATACACAGTTTAA-3′; TLR2 reverse primer, 5′-CGAGTTGAGATTGTTATTGCTAATATCTAAAA-3′; TLR2 probe, 5′-CATTGGCTTCCCCAGACCCTGGA-3′; TLR3 forward primer, 5′-AAAATCTCCAAGAGCTTCTATTAGCAA-3′; TLR3 reverse primer, 5′-TTGTATTTGATTTGATGACAACTCTAATCTTT-3′; TLR3 probe, 5′-CGTGAAGAACTTGATTTCCTTGGCAATTCTTC-3′; TLR4 forward primer, 5′-TGGCAGTTTCTGAGGAGTCATG-3′; TLR4 reverse primer, 5′-CCGCAGCAGGGACTTCTC-3′; TLR4 probe, 5′-CGGCATCATCTTCATCGTCCTGCAG-3′; TLR7 forward primer, 5′-CAGAAGTCCAAGTTTTTCCAGCTT-3′; TLR7 reverse primer, 5′-GGTGAGCCTGTGGATTTGTTG-3′; TLR7 probe, 5′-CAACTGGTACATCAGCACCTCTCAAGCAGA-3′; TLR8 forward primer, 5′-GATACCATTGCGGCGATAATATG-3′; TLR8 reverse primer, 5′-TTTACCTTGGCTAAGCACACATG-3′; TLR8 probe, 5′-ATGTTGGCTGCCCTGGCTCACC-3′; TLR9 forward primer, 5′-GAGACCCTGCTGTTGTCCTACAA-3′; TLR9 reverse primer, 3′-AGTTCCCCCCCACATCAA-5′; TLR9 probe, 5′-CGCCTGAGGACCTGGCCAATGTG-3′.

Peripheral blood mononuclear cells were cultured in 12-well plates at 2 × 10⁶ cells/ml in 2 ml. Cells were stimulated with TLR ligands and incubated for 5 h at 39°C/5% CO₂. Next, cells were harvested and stained with markers for CD4, CD14, CD80/86, CD172a, CCR7, and CADM-1. Samples were read using a FACSCanto II and data were analyzed with FlowJo software. Monocytes were identified and gated as CD14 cells, conventional DC1 (cDC1) as CD14⁻CD4⁻CADM1⁺CD172^low/− cells, cDC2 as CD14⁻CD4⁻CADM1⁺CD172⁺ cells, and pDC as CD14⁻CD4⁺CADM1⁻CD172^low/+ cells (26, 29).

Analysis was performed in Prism 7 (GraphPad, USA). All statistical analyses used parametric tests for paired samples. ANOVA with Greenhouse-Geisser correction followed by Tukey’s test was performed for multiple comparisons of the mean of each group with the control, and Dunnett’s test was used generate P values for each comparison. Where applicable two-way ANOVAs were performed followed by Sidak’s multiple comparison tests. P-values were defined as follows: ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05; ns: not significant. Correlation was calculated using the linear regression method.

Stimulating TLRs in immune cells engages various pathways, resulting in surface marker upregulation, cytokine production, and cell proliferation. As a first step, we tested the stimulatory potential of TLR ligands by assessing their ability to induce proliferation of whole PBMCs. We screened a set of 18 TLR ligands using a ³H-thymidine incorporation assay. We tested ligands described as stimulating TLR2/1 and TLR2/6 heterodimers, TLR3, TLR4, TLR5, TLR7, TLR8, and TLR9 (Figure 1A).

Pam3Cys-SK4, a TLR2/1 ligand, and Pam2Cys-SK4, a TLR2/6 ligand, induced proliferation 8–18-fold over control values. CL429 (a conjugate of a TLR2/1 ligand with a NOD ligand) induced lower levels of activation (2–7-fold over controls). No significant difference between Pam3Cys-SK4 and Pam2Cys-SK4 was detected (Figure 1B). The TLR3 ligand Poly I:C, all TLR4 ligands and the TLR5 ligands flagellin induced no significant proliferation compared to controls. Among the TLR7 ligands, imiquimod and gardiquimod induced potent proliferative responses in the range of those induced by Pam3Cys-SK4, while loxoribine was not active. Resiquimod, a TLR7/8 ligand, induced the highest proliferation of all ligands tested, with values 11–27-fold those of controls. For stimulation of TLR9, various CpGs were tested. The type A CpG D32 and the type B CpG2007 did not induce significant responses. CpG2142 induced a proliferation up to 5-fold over controls. Type B ODNs D-SL01 and type C D-SL03 induced the highest proliferation of this group, with proliferation values 7 and 12 times those of controls, respectively.

We identified B cells as the main proliferating population after TLR ligand stimulation. To this end, we employed multicolor staining for B cells (IgM⁺) and T cells (CD3⁺), with CellTrace Violet stain as a proliferation marker. Interestingly, up to 80% of IgM⁺ cells proliferated when stimulated with the TLR7/8 ligand resiquimod (Figure 1C), whereas no response or even suppression was found in T cells with the selected TLR ligands (Figure 1D).

To identify different cell subsets in peripheral blood, we established multicolor staining for surface IgM, IgG, CD21, and CD11R1 expression. The CD11R1 molecule in pigs is recognized by cross-reactive CD11b antibodies but differ in its cellular distribution (30). After excluding doublets, we identified four B cell populations (Figure 2A). P1 was defined as IgM⁺CD21⁺ cells and is composed of mainly naïve B cells (31). P2 represented IgG-switched memory B cells. The two novel B-cell subsets P3 and P4 were defined as CD21⁻IgM⁺CD11R1⁺ and CD21⁻IgM⁺CD11R1⁻, respectively.

To further characterize P3 and P4, we analyzed various parameters by flow cytometry (Figure 2B). Forward-scatter analysis showed that the two IgM⁺CD21⁻ subsets were bigger compared P1 and P2. Both P3 and P4 also expressed higher levels of IgM when compared to P1. CD11c was shown to be expressed at high level on P3 and on a major part of P4 but not on P2. P3 and P4 were also found to express higher levels of CD80/86, when compared to P1 and P2. Only P3 expressed CD14.

The higher size and surface IgM levels is a characteristic of B-1 cells (1) indicating that both P3 and P4 could represent the porcine equivalent of B-1 cells. Furthermore, as shown in Figure 2B, the phenotype of P3 was reminiscent to that of a subset of human B-1 cells expressing high levels of CD11b, CD11c, CD14, and CD86 (32, 33). We therefore also tested if P3 and P4 had a higher level of tonic intracellular signaling, a characteristic of B-1 cells (32, 34) by analyzing their expression of phosphorylated PLC-γ2, and found this to be elevated when compared to P1 and P2 subsets (Figure 2B).

Contrasting with monocytes and blood DC (29), none of the B-cell subsets expressed CD115 and CD172a. We therefore propose that P3 and P4 represent innate B cell subsets with B-1 like functions.

To evaluate the responses of the B-cell subsets P1–P4 to TLR stimulation, PBMCs were stimulated with the most potent TLR ligands identified in the screening experiments shown in Figure 1. Pam3Cys-SK4 and resiquimod induced strong upregulation of CD25 in all subsets (Figures 3A,B), while responses to ODN D-SL03 induced upregulation in all subsets with the exception of P3. This population was also less responsive to Pam3Cys-SK4 when compared to resiquimod. We also compared the relative responsiveness of the B-cell subsets by calculating the fold-increase in CD25 expression induced by TLR stimulation. While the response of P1 and P2 were comparable (6–7-fold increase of CD25 following resiquimod stimulation), P3 was found to be particularly responsive to resiquimod (17-fold increase). P4 was the subset with the overall lowest TLR responsiveness (3.3-fold increase).

Also, when analyzing CCR7 we found the strongest induction by Pam3Cys-SK4 and resiquimod (Figure 3C). In this assay, P1 and P2 were unresponsive to ODN D-SL03. Furthermore, no upregulation of CCR7 was detected in P4 after stimulation.

We next investigated the proliferative response of B cell subsets using multicolor flow cytometry with CellTrace Violet staining. The identification of P1-P4 B-cell subsets was not possible due to an apparent loss of CD11R1 expression in culture after 5 days (data not shown). We therefore gated for CD21⁺IgM⁺, CD21^−/lowIgM⁺, and IgG⁺ cells subset (Figure 4A). A representative CellTrace Violet staining of the three subsets is shown in Figure 4B and a summary of all biological replicates in Figure 4C. Similar to the induction of CD25 and CCR7 expression, Pam3Cys-SK4 and resiquimod were found to be the most potent inducers of proliferation. Only the IgG⁺ B-cell subset was unresponsive to the CpG stimulation. When comparing the three gated subsets, we found the CD21^−/lowIgM⁺ subset to have the strongest response to all three TLR ligands.

Considering the uncertainty with respect to the identity of the above B-cell subsets, we FACS-sorted the four B-cell subsets P1–P4 before stimulation with resiquimod and determined the proliferation by ³H-thymidine incorporations as in Figure 1A. The results demonstrated that P3 was the B-cell subset with the highest proliferative potential (Figure 4D).

Considering that spontaneous IgM secretion has been found as a characteristic of murine and human B-1 cells (33, 35), we tested this function in sorted porcine B cell subsets P1–P4. As shown in Figure 5A, high levels of IgM were only found in the supernatants of TLR-stimulated P3 and P4 B-cell subsets. Even the addition of recombinant porcine IL-2 and BAFF, both previously shown to be required for in vitro activation of porcine memory B cells to differentiate into antibody-producing cells (36), did not permit spontaneous IgM secretion by any B-cell subset. Nevertheless, in the presence of the cytokines the CD21⁺IgM⁺ P1 subset produced 38 ng/ml, but this value was still 39 and 85 times lower that the production found with P3 and P4, respectively. Altogether, this support that P3 and P4, although apparently unable to constitutively produce IgM, possess B1-like characteristics not found with conventional B cells.

Next, we compared polyclonal IgM secretion and restimulation of FMDV-specific memory B cells to produce antibodies following TLR stimulation and found potent responses with Pam3Cys-SK4, resiquimod and ODN D-SL03 (Figures 5B,E). Considering that pDCs play an important role in human B cell activation in vitro (37), we tested the effect of pDC depletion on antibody responses. Also in absence of pDC, B cells were able to produce IgM and virus-specific antibodies following TLR stimulation. Interestingly, antibody levels were significantly increased following resiquimod stimulation but not with the other TLR ligands (Figures 5C,F). When the sorted ILR3⁺ fraction containing the pDC was pooled with the negative fraction (“reconstituted”), the values obtained were again comparable to the PBMC controls indicating that the higher levels of IgM following resiquimod stimulation was not an effect of the sorting process (Figures 5D,G). In contrast to the promoting effect of pDC depletion on resiquimod stimulation, it resulted in reduced responses to FMDV antigen (PBMC compared to pDC-depleted P = 0.019 and pDC-depleted compared to reconstituted P = 0.01).

To better understand B cell response patterns to TLR ligand stimulation, we investigated TLR expression at the transcriptional level using RT-qPCR, and compared results to other antigen-presenting cells. We first compared MACS-sorted CD21⁺ B cells, CD14⁺ monocytes, and FACS-sorted CD172a⁺CD4⁺ pDCs (Figure 6A). B cells expressed high levels of TLR7, intermediate levels of TLR8 and TLR9, and low but detectable levels of TLR2, TLR3, and TLR4. Dominant TLR genes on monocytes were TLR4 and TLR8 (although statistically not significant different from other TLRs) and on pDC TLR3, TLR7, and TLR9.

Considering the strong responses of B cells to TLR2 ligands, we tested if TLR2 was upregulated following stimulation, as has been reported for human B cells. PBMCs were stimulated with Pam3Cys-SK4 and surface expression of TLR2 on CD21⁺ cells was analyzed by flow cytometry. After 24 h, a potent upregulation of TLR2 was detected (Figure 6B).

We next analyzed TLR expression by RT-qPCR on sorted B-cell subsets as defined in Figure 2. Only the expression of TLR3 on the P4 subset was found to be significantly higher when compared to P1 and P2. Nevertheless, TLR3 transcripts were also found in P3 but were absent in P1 and P2. This contrasted with the TLR4 expression, which was undetectable in the B-1 like subsets but present in P1 and P2, although at low levels (Figure 6C).

In the perspective of vaccine development, and to compare B cell TLR responsiveness to other antigen-presenting cells, we decided to investigate activation markers of blood dendritic cell subsets and monocytes after stimulation with the TLR ligands used for the B cell stimulations. The subsets definitions are shown in Figure 7A and are based on previous work (26, 29). After 5 h of stimulation with all ligands tested, CD40 expression was increased on cDC1 cells and monocytes. cDC2 cells responded to all ligands, with the exception of CpG ODN D-SL03. CD40 was also upregulated on pDCs in response to most ligands, with the exception of MPLA. The highest responses were seen with resiquimod (Figure 7B). CCR7 was not induced in cDC1 cells by MPLA and ODN D-SL03. Also, in cDC2 and pDC, MPLA was unable to induce CCR7 (Figure 7C). Interestingly, high induction of CCR7 occurred in pDCs, particularly after stimulation with gardiquimod and resiquimod.