Schistosome egg antigens drive the development of Breg cells in vivo

To elucidate the mechanism by which S. mansoni can drive the development of Breg cells, we first investigated whether schistosome eggs or their soluble antigens were sufficient to drive Breg cell development in vivo, without the context of natural infection. Intraperitoneal treatment of C57BL/6 mice with two doses of 5000 S. mansoni eggs or 100 μg of soluble egg antigens (SEA) one week apart efficiently induced IL-10 protein expression in splenic CD19⁺ B cells one week after the last injection. IL-10 protein secreted during 2 days ex vivo restimulation with SEA was 3 to 4-fold increased compared to the amount secreted by restimulated B cells from control-treated animals (Fig 1A), while IL-6 was unchanged. This indicates a typical cytokine expression pattern characteristic for Breg cells. The frequencies of B cells expressing intracellular IL-10 protein were likewise significantly 2-fold increased (Fig 1B and 1C). Also the surface activation marker CD86, often upregulated on activated B cells and Breg cells [33–35], was increased on splenic B cells by egg or SEA treatment, while CD40 expression was not significantly changed (Fig 1D). To verify that the observed effects are specific and exclude a general influence of protein solutions on B cell IL-10 production and activation, we treated mice with human serum albumin (HSA) as infection-unrelated control protein. We did not observe an increased IL-10 secretion (S1A Fig) or CD86 expression (S1B Fig) by B cells compared to the PBS group, and concluded that PBS is a suitable control for subsequent experiments. Furthermore, injection of eggs or egg antigens was as efficient in increasing the frequency of IL-10-expressing B cells as was chronic infection with S. mansoni (S1C Fig), and egg-injected mice continued to have an elevated B cell IL-10 production until at least 4 weeks after the last egg injection (S1D Fig), indicating that this phenotype is persisting over longer periods. SEA was purified from liver eggs and can contain LPS to variable extent. Since the TLR4 ligand LPS is known for its capacity to drive B cell IL-10 expression and Breg cell development [33,36,37], it is crucial to exclude that the Breg driving capacity by SEA in vivo was due to LPS contamination of the schistosome antigen preparations. Therefore, the same experiment was repeated in TLR4-deficient animals and compared to wild-type. Upon SEA treatment, IL-10 secretion of B cells as well as intracellular IL-10 and surface CD86 expression was overall comparable in both groups (S2 Fig), indicating that the Breg-inducing capacity by SEA was largely not attributable to a putative LPS contamination. To confirm the regulatory function, splenic B cells from the various groups were tested for their capacity to drive Treg cell development, an acquired phenotype previously described for splenic B cells during natural schistosome infections [12]. Indeed, splenic B cells from egg- or SEA-injected, but not control-treated, mice induced the development of CD25⁺Foxp3⁺ Treg cells during 4 day co-culture with CD25-depleted CD4 T cells (Fig 1E and 1F), which confirms the regulatory capacity of egg antigen-activated B cells ex vivo. As expected, IL-10 protein concentration in co-culture supernatants was only increased in presence of egg antigen-activated but not control B cells (Fig 1G). Collectively, these data show that schistosome eggs or their soluble antigens on their own are sufficient to induce IL-10-producing Breg cells in vivo, without the context of natural infection, and that these B cells are bona fide Breg cells that can drive Treg cell development.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schistosome eggs and SEA (soluble egg antigen) drive development of Breg cells in vivo.

C57BL/6 mice were i.p. injected with two doses of 5000 S. mansoni eggs or 100 μg SEA in PBS, or PBS as control. At day 14, CD19⁺ MACS-isolated splenic B cells were restimulated with SEA (20 μg/ml) for 2 days. (A) Cytokine concentration in culture supernatants as determined by ELISA. (B, C) Representative FACS plots (B) and summary (C) for intracellular IL-10 expression of B cells after addition of Brefeldin A to the last 4 hours of the culture. (D) Mean fluorescence intensity of CD40 and CD86 expression. (E-G) SEA-restimulated B cells were co-cultured for 4 days with CD25-depleted CD4 T cells. Frequency of CD25⁺Foxp3⁺ Treg cells after co-culture in representative FACS plots (E) and summary (F) is shown. (G) IL-10 concentration in culture supernatants after co-culture. Summary of 4 experiments with N = 12–16 (A-D) or 2 experiments with N = 8–9 (E-G). Significant differences by Mann-Whitney test are indicated with * p < 0.05, ** p < 0.01, *** p < 0.001.

https://doi.org/10.1371/journal.ppat.1006539.g001

Schistosome antigens activate marginal zone B cells in the spleen

Different B cell subsets have been described to give rise to Breg cells, especially in spleen where subsets differ e.g. in tissue localization and pathogen recognition receptor expression [25,38]. We and others had previously identified CD23^lowCD21⁺ marginal zone B cells as the major IL-10-producing splenic B cell subset during chronic Schistosoma infection and mediating protection in a mouse model of airway inflammation [8,12]. To test whether soluble egg antigens act on the same splenic subset, we sorted splenic CD23^lowCD21⁺ marginal zone B cells from egg-, SEA-treated, or control mice for subsequent ex vivo restimulation and cytokine analysis, and compared this with the major splenic B cell subset, CD23^hiCD21^- follicular B cells (Fig 2A). Only marginal zone B cells but not follicular B cells showed significantly increased IL-10 secretion as was measured in culture supernatants after 2 day restimulation with SEA (Fig 2B). As for total B cells (Fig 1A), also for the individual subsets IL-6 expression was not increased (Fig 2C). Intracellular IL-10 expression and CD86 expression was likewise significantly upregulated in marginal zone B cells of egg- or SEA-treated mice compared to control-treated mice. SEA seemed to be more potent than egg injection in activating follicular B cells, as SEA-injection also significantly increased intracellular IL-10 and CD86 expression in this subset, although expression levels remained significantly lower compared to marginal zone B cells (Fig 2D and 2E). For analysis of B cell activation we generally restimulated cells ex vivo with SEA. This increased the baseline expression of CD86 in all groups compared to medium (average MFI of follicular B cells: 1013–1632±62.2; for marginal zone B cells: 2640–4118±176.9, for all groups and without significant differences between groups), but was required for detection of B cell cytokines as a result of the in vivo antigen exposure. Without SEA restimulation, we found a trend of increased IL-10 production by B cells which only reached significance upon additional restimulation (S1E Fig), indicating that renewed exposure to antigen is required to achieve detectable B cell activity and cytokine production. This is also supported by experiments in IL-10 GFP reporter mice, in which IL-10 (GFP) accumulates in B cells during the entire in vivo treatment period. Here, increased IL-10 (GFP) expression, without ex vivo restimulation, was only detectable in B cells of 14 weeks chronically infected mice, but not in B cells of mice that received two injections of eggs within a relatively short period of 2 weeks (S1F and S1G Fig). Altogether, these data indicate that the result of in vivo development of IL-10 producing B cells is in principle detectable without restimulation (S1E and S1G Fig), but the data from egg-injected IL-10 reporter mice also suggest that an ex vivo SEA restimulation is required to visualize its full IL-10 potential, something which will happen in vivo during a natural infection due to the constant production of eggs and the high levels of circulating antigens. Taken together, these data support the notion that B cells, and in particular marginal zone B cells, are responsive to in vivo schistosome antigen stimulation thus supporting the findings in natural schistosome infections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Schistosome antigens mainly activate the marginal zone B cell subset of the spleen.

C57BL/6 were treated as in Fig 1, splenic follicular (FO) B cells (CD19⁺CD23⁺CD21^low) and marginal zone (MZ) B cells (CD19⁺CD23^-CD21^hi) FACS sorted, and restimulated for 2 days with SEA (20 μg/ml). (A) Gating scheme for both B cell subsets within the CD19⁺ gated B cell population. (B, C) Cytokine concentration in culture supernatants as determined by ELISA. (D) Intracellular IL-10 expression of B cell subsets after addition of Brefeldin A to the last 4 hours of the culture. (E) Mean fluorescence intensity of CD86 expression of B cell subsets. Summary of 2 experiments with N = 6 (B, C) or 3–5 experiments with N = 13–21 (D, E). Significant differences by Mann-Whitney test are indicated with * p < 0.05, *** p < 0.001. Significant differences between B cell subsets by Wilcoxon paired test are indicated with ^# p < 0.05, ^## p < 0.01, ^### p < 0.001.

https://doi.org/10.1371/journal.ppat.1006539.g002

Macrophage subsets of the marginal zone bind SEA but are dispensable for Breg cell induction

Molecules secreted by schistosome eggs are highly glycosylated and known to bind to C-type lectin receptors [39,40]. Since B cells show a very restricted expression of those receptors [41], we hypothesized that other C-type lectin receptor-expressing cell types in the splenic marginal zone, such as macrophages or dendritic cells, bind SEA and provide additional signals to the marginal zone B cells to support Breg cell development. Among the accessory cell types, macrophages of the splenic marginal zone were of particular interest because of their known interactions with marginal zone B cells [28,42] and schistosome antigens [43,44]. However, it was unknown whether marginal zone macrophages can capture SEA in vivo and are important for B cell IL-10 expression. To evaluate this, fluorescently labeled SEA was injected i.v. and 30 minutes to 24 hours later various splenic cell types were analyzed for bound SEA using fluorescence microscopy and flow cytometry. Already after 30 minutes of injection, SEA clustered along the marginal zone area of the spleen as detected by fluorescence microscopy of splenic tissue sections (Fig 3A). SEA localized predominantly within two specialized macrophage subsets of the marginal zone: SIGN-R1-expressing MZ macrophages and Siglec-1-expressing marginal metallophilic macrophages (Fig 3B). In contrast, after injection of labeled ovalbumin (OVA) as non-schistosomal control protein no fluorescence signal was detected in the spleen (S3 Fig). Flow cytometry confirmed that 83% of metallophilic macrophages were positive for SEA only 30 minutes after injection, and still 53% of cells after 24 hours (Fig 3C). In addition, F4/80-expressing red pulp macrophages and Ly6C^hi monocytes efficiently bound SEA (81.7 and 82.6%, respectively), while only a small fraction of dendritic cells and neutrophils did bind SEA (Fig 3C, S4A–S4C Fig for gating scheme of cell types). Metallophilic macrophages not only bound SEA abundantly, they also significantly upregulated typical surface activation markers such as CD11c and CD86, but not MHCII, at 30 minutes and 2 hours after SEA injection, respectively (Fig 3D). Because macrophages of the marginal zone were most potent in binding SEA, we next addressed their role for SEA-mediated marginal zone Breg cell induction. To this end, macrophages were depleted in vivo by i.p. injection of clodronate-containing liposomes [45] prior to injection of schistosome eggs. Eggs were injected 3 and 4 weeks after clodronate treatment, at time-points when only macrophages, including metallophilic and MZ subsets, were significantly reduced in spleens, but no other cell types (S4D Fig and [46]). Successful and specific depletion of splenic macrophages was also confirmed by fluorescence microscopy of tissue sections (Fig 3E) and flow cytometry (S4E Fig) at 7 days after the last egg injection, when B cell activity was analyzed. Unexpectedly, IL-10 secretion of splenic B cells from macrophage-depleted mice was equal to that from control liposome-treated mice (Fig 3F). Also the upregulation of intracellular IL-10 (Fig 3G) and CD86 expression in B cells, as well as the induction of Foxp3⁺CD25⁺ and IL-10⁺CD25⁺ CD4 T cells (S4F and S4G Fig) following SEA injection was not affected by the absence of macrophages. In conclusion, splenic macrophages are not essential for schistosome antigen-induced Breg cell development, despite the high binding of SEA by different macrophage subsets.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Macrophage subsets of the MZ bind SEA but are dispensable for schistosome antigen-mediated Breg cell induction.

(A, B) Spleens were snap-frozen 30 minutes after i.v. injection of 200 μg fluorescently labeled SEA, and binding of SEA analyzed by fluorescence microscopy. Images are representative for 2 experiments with N = 5 mice and 3 follicles per section imaged. (A) SEA clustered around the marginal zone of the spleen (indicated by white dashed line). (B) SEA localized in the marginal zone to macrophages expressing Siglec-1 (marginal metallophilic macrophages) and SIGN-R1 (MZ). (C, D) Mice were i.v. injected with fluorescently labeled SEA and splenocytes harvested 30 minutes-24 hours later. (C) Frequency of SEA-positive cells in various splenocyte subsets as determined by flow cytometry (gating schemes see S4 Fig). MΦ, macrophages; RP, red pulp; MM, marginal metallophilic; DC, dendritic cells. (D) Mean fluorescence intensity of surface markers on MM macrophages calculated as fold increase versus the expression at time-point 0 hours. Summary of 2 experiments with N = 2. (E-G) Mice were i.p. injected with 200 μl clodronate-containing liposomes or PBS control liposomes. Three weeks later, mice were i.p. injected with two doses of 5000 eggs. Seven days after the last egg injection, splenic B cells were restimulated for 2 days with SEA. (E) Absence of splenic Siglec-1 and SIGN-R1-expressing macrophage subsets at the time-point of spleen collection was confirmed by fluorescence microscopy. B cells were stained with B220. (F) Cytokine concentration in culture supernatants as determined by ELISA. (G) Intracellular IL-10 expression of B cells after addition of Brefeldin A to the last 4 hours of the culture. One representative (N = 5) out of 2 similar experiments is shown. Significant difference by Mann-Whitney test is indicated with * p < 0.05, ** p < 0.01. Significance as determined by one-sample t-test of log-transformed data is indicated with ^# p < 0.05, ^## p < 0.01.

https://doi.org/10.1371/journal.ppat.1006539.g003

MZ B cells bind and take up schistosome antigens

To test whether B cells directly bind and interact with schistosome antigens without the help of surrounding accessory cell types, fluorescently labeled SEA was injected i.v. and its co-localization with B cells analyzed by fluorescence microscopy of splenic tissue sections. Indeed, egg antigens were found to co-localize with some splenic B220⁺ B cells (Fig 4A). Flow cytometry, a more sensitive method compared to fluorescence microscopy, showed that only MZ B cells but not follicular B cells did bind SEA in vivo, with a maximum of 56.4% of cells being positive at 2 hours after SEA injection. SEA was still detectable on 11.5% of marginal zone B cells at 24 hours (Fig 4B). Marginal zone B cells also showed an increased CD86 surface expression following SEA injection, which was significant at 6 hours after SEA injection (Fig 4C). Interestingly, marginal zone B cells that bound SEA showed a higher CD86 expression compared to cells that were found to be negative for SEA, i.e. approximately 3-fold (SEA positive) versus only 1.5-fold (SEA negative) compared to B cells from untreated animals (Fig 4C). This suggests that marginal zone B cells not only efficiently bind egg antigens in vivo, but also become (more) activated as a consequence of this interaction. Binding of SEA to B cells was confirmed in vitro by culturing splenic B cells with fluorescently labeled SEA for 60 minutes, after which up to 16% were positive for SEA as measured by flow cytometry (Fig 4D), which is a similar percentage as found for total B cells after in vivo application of labeled SEA (Fig 4B). By using SEA labeled with the pH-sensitive dye pHrodo, it was shown that egg antigens were not only bound to the surface but were also internalized by B cells into acidic cellular compartments (Fig 4D). As in vivo, also in vitro the marginal zone B cell subset bound SEA more efficiently than the follicular B cell subset, with in average 15.9% versus 6.9% of cells being positive for SEA (Fig 4E). Collectively, these data show that B cells, and in particular MZ B cells, are capable of directly interacting with schistosome antigens by binding and internalization of those antigens, both in vivo and in vitro.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Schistosome antigens bind to B cells and are internalized into acidic compartments.

(A) C57BL/6 mice were i.v. injected with 200 μg of fluorescently labeled SEA and spleens snap-frozen 30 minutes later. Fluorescence microscopic image shows localization of SEA to B220⁺ B cells and is representative for 2 experiments with N = 5 mice and 3 viewing fields per section imaged. (B, C) Splenic B cells were analyzed by flow cytometry at 30 minutes to 24 hours after i.v. injection of labeled SEA for (B) frequency of SEA⁺ cells within total B cells or B cell subsets as gated in Fig 2A, and for (C) fold increase of CD86 expression on total MZ B cells compared to MZ B cells gated for SEA⁺ and SEA^- populations. Summary of 2 experiments with N = 2. (D, E) Splenic B cells from naïve mice were cultured in vitro with SEA (20 μg/ml) labeled with a green dye (PF-488) alone, or co-labeled with a pH-sensitive dye (pHrodo Red). After 60 minutes, the frequency of SEA⁺ B cells was determined by flow cytometry. (D) Gated for total CD19⁺ B cells (1 representative out of 2 experiments shown). (E) Gated for follicular (FO) and marginal zone (MZ) B cell subsets (summary of 3 experiments). Significant differences indicated with ^# p < 0.05 and ^## p < 0.01 are determined by one-sample t-test of log-transformed data (C) or between B cell subsets by Wilcoxon paired test.

https://doi.org/10.1371/journal.ppat.1006539.g004

Breg cells can be generated in vitro by direct interaction with schistosome egg antigens

Next, we investigated whether the observed direct interaction of B cells with SEA can drive IL-10 expression and induction of regulatory B cell function in vitro. To this end, CD19⁺ splenic B cells from naïve mice isolated using MicroBeads were cultured for 3 days with SEA. SEA-stimulated B cells secreted significantly more IL-10, but not IL-6, compared to non-stimulated B cells (Fig 5A), showing a typical cytokine pattern characteristic for schistosome-induced Breg cells. Separate cultures of sorted MZ and follicular B cells showed once more that the MZ B cell subset reacted more potently to SEA, e.g. with a significantly higher IL-10 secretion compared to the follicular subset (Fig 5B). In addition, frequencies of B cells expressing intracellular IL-10 were likewise increased (Fig 5C). To ensure that the IL-10 phenotype is not reliant on artificially high levels of stimulation with PMA/ionomycin which is added to facilitate detection of intracellular IL-10, we repeated the assay using cells from IL-10-GFP reporter (TIGER) mice with similar results (S5 Fig). CD40 and CD86 expression were upregulated compared to cultures in medium alone (Fig 5D). Importantly, in vitro SEA-activated B cells were also capable of driving Treg cell development during a 4 day co-culture with CD25-depleted CD4 T cells (Fig 5E), thus providing further evidence for a bona fide regulatory function of the in vitro induced Breg cells. Because we had seen internalization of egg antigens into acidic compartments (Fig 4D), we wondered whether lysosomal processing is necessary for induction of IL-10 expression. Addition of chloroquine, an inhibitor of endosomal acidification [47], significantly reduced the IL-10 secretion and frequency of IL-10⁺ B cells induced by SEA and CpG (ligand for endosomal TLR9), but not by Pam3Cys (ligand for surface TLR2) (Fig 5F and 5G). This suggests that internalization and endosomal processing of SEA is required for B cell IL-10 induction. The type of receptor involved in direct activation of Breg cells by SEA remains unknown. Egg antigens are abundantly glycosylated and known to bind to C-type lectin receptors [39,40]. Le^x-motifs, one of the most abundant glycan structures present in SEA, bind to the C-type lectin receptor SIGN-R1. However, when treating SIGN-R1-deficient mice with SEA, IL-10 expression was equally well induced compared to wild-type mice (S6A Fig), suggesting no involvement of the Le^x-motifs. Furthermore, stimulation of various TLRs on B cells, including TLR2, TLR4, TLR7 and TLR9, has been described to induce IL-10 production [30,31]. Because SEA has been reported to contain TLR2 activity [48,49], we compared SEA-induced Breg cell responses in wild-type and TLR2-deficient B cells. We did however not observe any difference in IL-10 secretion between the two strains (S6B Fig), excluding a role of TLR2-triggering SEA components in SEA-induced B cell IL-10 production. This is further supported by the fact that TLR2-mediated B cell activation was independent of endosomal processing while SEA-mediated activation was dependent on it (Fig 5F and 5G). Collectively, these data demonstrate that Breg cells can be generated in vitro by culture with schistosome antigens, that endosomal processing is involved in this process, and that these Breg cells are functional in the sense that they support Treg cell development.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Breg cells can be generated in vitro by stimulation with schistosome egg antigens which involves lysosomal processing.

(A-E) Splenic B cells from naïve mice were cultured for 3 days with 20 μg/ml SEA or medium as control. Cytokine concentration in culture supernatants of (A) total B cells (6 experiments) or (B) sorted B cell subsets (2 experiments with N = 6). (C) Intracellular IL-10 expression of total B cells after addition of PMA, ionomycin and Brefeldin A to the last 4 hours of the culture (4 experiments). (D) Mean fluorescence intensity of the activation markers CD40 and CD86 (4 experiments). (E) Splenic B cells were stimulated in vitro as above, and subsequently co-cultured with CD4⁺CD25^- sorted splenic T cells. After 4 days, the frequency of CD25⁺Foxp3⁺ Treg cells within the CD4 T cell population was determined by flow cytometry. Summary of 4 experiments. (F, G) Splenic B cells from naïve mice were cultured in vitro with SEA (20 μg/ml), CpG (5 μg/ml) or Pam3Cys (10 μg/ml) for 2 days. Every day, chloroquine (5 μM) was added to the culture. Intracellular IL-10 expression after addition of PMA, ionomycin and Brefeldin A during the last 4 hours of the culture (F), and cytokine concentrations in culture supernatants (G). Summary of 2 experiments with N = 3–4. Significant differences are indicated with * p < 0.05, ** p < 0.01 and tested by Mann-Whitney. ^# p < 0.05 indicates significant difference between B cell subsets by Wilcoxon paired test.

https://doi.org/10.1371/journal.ppat.1006539.g005

CD40 ligation enhances the schistosome antigen-induced Breg cell development

Previous studies highlighted a role for CD40 ligation during in vitro Breg cell induction [50,51]. We therefore tested whether CD40 ligation could increase the SEA-mediated effect on Breg cell development. Addition of anti-CD40 stimulatory antibody (Ab) to the 3 day SEA culture increased IL-10 secretion of splenic B cells by 1.7-fold compared to SEA alone. A similar enhancing effect was observed for IL-6 secretion (Fig 6A) and CD86 expression (Fig 6B). In contrast to anti-CD40 Ab, addition of anti-IgM Ab did not significantly enhance the SEA-mediated B cell activation (S7 Fig). To exclude stimulatory effects from LPS in vitro, B cells from TLR4-deficient mice were stimulated with SEA with or without addition of anti-CD40 Ab. The fold increase of IL-10 secretion compared to B cells cultured in medium was similar for wild-type and TLR4-deficient cells, thus excluding a major effect of the TLR4 ligand LPS (S6C Fig). Finally, co-culture of CD25-depleted CD4 T cells with anti-CD40-stimulated B cells increased the frequency of CD25⁺Foxp3⁺ Treg cells, which was further increased if B cells had been stimulated with SEA plus anti-CD40 Ab (Fig 6C). Thus, SEA stimulation plus CD40 ligation of B cells further enhanced the capacity to drive the development of IL-10-producing Breg cells in vitro.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. CD40 ligation enhances SEA-induced Breg cell development.

(A, B) Splenic B cells from naïve mice were cultured with 20 μg/ml SEA or medium as control, with or without addition of anti-CD40 co-stimulatory antibody. After 3 days of culture, supernatants were analyzed for IL-10 and IL-6 by ELISA (A) and B cell CD86 expression by flow cytometry (B). Summary of 6 experiments. (C) Splenic B cells were stimulated for 3 days with anti-CD40 mAb (2 μg/ml) and SEA (20 μg/ml) or anti-CD40 alone, and subsequently co-cultured with CD4⁺CD25^- sorted splenic T cells. After 4 days, the frequency of CD25⁺Foxp3⁺ Treg cells within the CD4 T cell population was determined by flow cytometry. Summary of 4 experiments. (D-F) Mice were i.p. injected with two doses of 100 μg SEA or PBS as control. CD40 ligand was blocked in vivo by i.p. injection of 200 μg hamster anti-mouse CD40 ligand or 200 μg hamster IgG as control (Ctrl) for 4 times, every 4 days starting at day -1 before SEA/PBS treatment. At day 14 after the first SEA/PBS injection, splenic B cells were in vitro restimulated with SEA for 2 days. Summary of 2 experiments (N = 8). (D) Fold increase of IL-10 in supernatants from B cells of SEA-injected versus PBS-injected mice. (E) Intracellular IL-10 expression of B cells after addition of Brefeldin A to the last 4 hours of the culture. (F) Mean fluorescence intensity of CD86 on B cells. Significant differences are indicated with * p < 0.05, ** p < 0.01, *** p < 0.001 as tested by Mann-Whitney test. ^# p < 0.05, ^## p < 0.01 and ^### p < 0.001 indicates significant difference by one-sample t-test of log-transformed data (D) or to respective medium control by Mann-Whitney test.

https://doi.org/10.1371/journal.ppat.1006539.g006

Previous reports only addressed the role of CD40 ligation in vitro, but the relevance for in vivo Breg cell induction was not investigated. We therefore blocked CD40 ligand in vivo by i.p. injection of a hamster anti-mouse CD40 ligand blocking mAb (200 μg; every 4 days starting at day -1 prior to 1^st SEA injection) during SEA treatment of mice and analyzed the effect on Breg cell activation. In hamster IgG-injected control mice, SEA treatment increased the amount of B cell-derived IL-10 secretion by 3.3-fold compared to PBS treatment. Upon anti-CD40 ligand administration this increase was only 1.6-fold and thereby significantly lower (Fig 6D). Importantly, the SEA-mediated upregulation of intracellular IL-10 and CD86 expression by B cells was even fully abolished when CD40 ligand was blocked (Fig 6E and 6F). Taken together, CD40 ligation enhances the Breg cell-inducing effect of SEA both in vitro and in vivo.

The secretory schistosome egg antigen IPSE/alpha-1 induces Breg cell development in mice and humans

SEA is a complex mixture of several different antigens. In the next step, we therefore aimed to identify specific antigens in SEA that are relevant for Breg cell induction. We focused on three major antigens that provoke an antibody response in nearly all infected patients: omega-1, kappa-5 and IPSE/alpha-1 [52,53]. B cells were able to bind fluorescently labelled natural IPSE/alpha-1 (nIPSE), a secreted egg antigen we purified from egg extracts, in a dose-dependent manner during 60 minutes in vitro culture (Fig 7A). During 3 days culture however, nIPSE induced significantly elevated IL-10 but not IL-6 secretion by B cells in a concentration dependent manner (Fig 7B and S8A Fig). Importantly, recombinant IPSE/alpha-1 expressed in tobacco plants (pIPSE), which behaves as nIPSE in terms of protein dimerization and human basophil activation (S9 Fig), had similar effects to the natural molecule on B cell IL-10 and IL-6 secretion (Fig 7C). Both nIPSE and pIPSE-stimulated B cells were capable of driving Treg cell development during B cell-T cell co-culture (Fig 7D). Interestingly, SEA depleted of IPSE/alpha-1 (SEAΔIPSE) was as efficient as total SEA in inducing IL-10 secretion and CD86 expression by B cells, which suggests that also SEA antigens other than IPSE/alpha-1 can activate B cells (Fig 7B). However, as opposed to IPSE/alpha-1, other major components in SEA, such as omega-1 and kappa-5, did not increase IL-10 in any of the tested concentrations (1–20 μg/ml) (S8B Fig), thus excluding a role for these antigens in SEA-mediated Breg cell induction. Notably, omega-1 is toxic to B cells at concentrations of 5 μg/ml and above, and was therefore only tested at 1 μg/ml. For better comparability, we determined the following average relative amounts of IPSE/alpha-1, omega-1 and kappa-5 within SEA, based on the yields of several purifications of these molecules from SEA: 1.2% IPSE/alpha-1, 0.6% omega-1 and 1.8% kappa-5. Furthermore, to proof that the Breg cell-inducing effect is specific for molecules in SEA, we stimulated B cells in vitro with adult worm antigen (AWA) as control of an S. mansoni-derived antigen mixture not containing IPSE. AWA was unable to induce IL-10 secretion when tested in the same concentration as used for SEA (S8C Fig). For a possible future therapeutic application of antigen-activated Breg cells against e.g. allergic diseases, it is crucial to confirm the IL-10-inducing effect in human B cells. After 3 days in vitro stimulation with SEA and anti-CD40, we found a significant increase in the percentage of total human IL-10⁺ CD19⁺ B cells compared to cells cultured with anti-CD40 alone. Comparing different B cell subsets, which have previously been attributed with regulatory properties [54,55], we found the increase in IL-10⁺ B cells after SEA stimulation to be most pronounced among CD1d⁺ B cells rather than CD24⁺CD27⁺ and CD24⁺CD38⁺ B cells. Both nIPSE and pIPSE significantly increased the fraction of IL-10-expressing cells among CD1d⁺ B cells, whereas neither had an effect on the other two subsets investigated. As CD1d⁺ B cells only comprise a very small fraction of all B cells (CD24⁺CD27⁺ B cells >> CD24⁺CD38⁺ B cells > CD1d⁺ B cells), the effect of nIPSE and pIPSE does not translate into an increase in the percentage of IL-10⁺ cell in the total B cell pool in contrast to SEA (Fig 8). This also suggests that additional molecules in SEA may have an IL-10-inducing effect. Collectively, we demonstrated that SEA was bound to and internalized by B cells, and that this direct interaction drives the development of Breg cells. Furthermore, we identified IPSE/alpha-1 as a single molecule of SEA that induces Breg cells in mice and humans, both as a natural and a recombinant molecule.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. The secretory SEA component IPSE/alpha-1 induces Breg cell development.

(A) Splenic B cells were incubated with PF-647-labeled SEA (20 μg/ml), natural IPSE/alpha-1 (nIPSE, 1–5–10 μg/ml) or left untreated and the frequency of SEA- and IPSE-positive cells measured after 60 minutes by flow cytometry. Representative FACS plots and summary of 3 experiments is shown (N = 3). (B) Splenic B cells from naïve mice were cultured for 3 days with 10 μg/ml nIPSE, 20 μg/ml SEA, 20 μg/ml SEAΔIPSE or medium as control, with or without addition of anti-CD40 mAb (2 μg/ml). Cytokine concentration in supernatants after 3 days as determined by ELISA. Summary of 4 experiments. (C) Splenic B cells from naïve mice were cultured for 3 days with 10 μg/ml recombinant, plant-derived IPSE (pIPSE), 20 μg/ml SEA or medium as control. Cytokine concentration in supernatants after 3 days as determined by ELISA. Summary of 4 experiments. (D) Splenic B cells were cultured for 3 days with nIPSE (10 μg/ml), SEA (20 μg/ml) or medium alone as control, and subsequently co-cultured with CD4⁺CD25^- sorted splenic T cells from C57BL/6 or DEREG mice. After 4 days, the frequency of CD25⁺Foxp3⁺ Treg cells within the CD4 T cell population was determined by flow cytometry. Summary of 3 experiments. Significant differences to medium control are indicated with * p < 0.05, as tested by Mann-Whitney test. ^# p < 0.05 indicates significant difference by one-sample t-test of log-transformed data.

https://doi.org/10.1371/journal.ppat.1006539.g007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Human CD1d⁺ Breg cells increase IL-10 expression after SEA and IPSE/alpha-1 stimulation.

B cells were isolated from PMBCs of healthy volunteers and stimulated for 3 days with SEA (20 μg/ml), natural IPSE (nIPSE, 10 μg/ml), plant-derived IPSE (pIPSE, 10 μg/ml), or left untreated. Different Breg cell subsets and intracellular IL-10 expression were assessed by FACS. Summary of 3 donors. Significant differences to the respective medium control are indicated with ^# p < 0.05, or between conditions with and without anti-CD40, with * p < 0.05, ** p < 0.01 as tested by paired t-test.

https://doi.org/10.1371/journal.ppat.1006539.g008

Animals

Female C57BL/6OlaHsd mice from Harlan, TLR4-deficient mice (on C57BL/6 genetic background, kindly provided by Dr. S. Akira, Osaka, Japan), TLR2-deficient mice (on C57BL/6 genetic background, kindly provided by the group of Dr. K. Willems van Dijk), SIGN-R1-deficient mice (on C57BL/6 genetic background, kindly provided by the group of Dr. W. Unger), DEREG (DEpletion of REGulatory T cells) mice (on C57BL/6 genetic background, kindly provided by Dr. T. Sparwasser) and IL-10-GFP reporter (TIGER) mice (on C57BL/6 genetic background, kindly provided by Dr. R.A. Flavell) were housed under SPF conditions in the animal facilities of the Leiden University Medical Center in Leiden, The Netherlands, and used for experiments at 8–14 weeks of age. Percutaneous infection of mice with S. mansoni was performed as described elsewhere [12], and mice sacrificed in the chronic phase (14–15 weeks) post infection.

Preparation and purification of schistosome eggs, egg and worm antigens

Freshly isolated S. mansoni eggs from trypsinized livers of hamsters infected for 50 days were washed in RPMI medium with 300 U/ml penicillin, 300 μg/ml streptomycin (both Sigma-Aldrich, Zwijndrecht, The Netherlands) and 500 μg/ml amphotericin B (Thermo Fisher Scientific, Breda, The Netherlands) and then kept at -80°C. SEA, AWA, omega-1, kappa-5 and IPSE/alpha-1 were prepared and isolated as described previously [53,71,72,73]. The purity of the antigen preparations was checked by SDS-PAGE and silver staining, and protein concentrations determined using the BCA procedure. The antigen preparations had an endotoxin content of less than 150 ng/mg protein (SEA) or 3 ng/mg protein (purified molecules) as tested by Limulus Amoebocyte Lysate (LAL) test and TLR4-transfected HEK-reporter cell lines (kindly provided by Prof. Golenbock, University of Massachusetts Medical School, Boston, USA).

Production and purification of recombinant IPSE from N. benthamiana plants

Recombinant IPSE was produced by transient expression in Nicotiana benthamiana and purified according to the methods described in [70]. In short, the complete sequence encoding the 134 AA mature Schistosoma mansoni IPSE (Smp_112110) was codon optimized and preceded by a signal peptide from the Arabidopsis thaliana chitinase gene (cSP) and a N-terminal 6x histidine-FLAG tag (H6F) was included. The full sequence was synthetically constructed at GeneArt and cloned into a pHYG expression vector. In all experiments the silencing suppressor p19 from tomato bushy stunt virus in pBIN61 was co-infiltrated to enhance expression. For gene expression the two youngest fully expanded leaves of 5–6 weeks old N. benthamiana plants were infiltrated by injecting Agrobacterium tumefaciens containing the IPSE expression plasmid. N. benthamiana plants were maintained in a controlled greenhouse compartment (UNIFARM, Wageningen) and infiltrated leaves were harvested at 5–6 days post infiltration. Plant-produced recombinant IPSE was obtained by applying leaf apoplast fluid containing IPSE to Ni-NTA Sepharose (IBA Life Sciences) in 50 mM phosphate buffered saline (pH 8) containing 100 mM NaCl. Bound IPSE was eluted with phosphate buffered saline (pH 8) containing 0.5M imidazole. Total soluble apoplast proteins and purified IPSE were separated under reducing/non-reducing conditions by SDS-PAGE on a 12% Bis-Tris gel (Invitrogen) and subsequently stained with Coomassie brilliant blue staining.

Isolation of splenocytes, B cells and T cells

Single cell suspensions of murine spleens were prepared by dispersion through a 70 μm cell strainer (BD Biosciences, Breda, The Netherlands), and erythrocytes depleted by lysis. For analysis of splenic myeloid cell populations, spleens were digested for 1 hour at 37°C by incubation with collagenase D (2 mg/ml; Roche, Woerden, The Netherlands) and DNase I (2000 U/ml; Sigma-Aldrich) before dispersion. B cells were purified from spleens by using anti-CD19 MicroBeads (Miltenyi Biotec, Leiden, The Netherlands) following the manufacturer’s protocol. Purity was routinely ~95–98%. After a typical CD19 MACS sort, circa 83% of all contaminating cells were CD3⁺ T cells, 7% CD11b⁺CD11c⁺ cells, 2% CD11b⁺ CD11c^- cells, and 8% other cells. To determine cytokine secretion of splenic B cell subsets, CD19⁺ B cells were subsequently sorted by flow cytometry for follicular B cells (CD23⁺CD21^low) and marginal zone B cells (CD23^-CD21^hi) which resulted in purities of > 98%. CD4⁺ T cells were purified from spleens by negative selection and depleted of CD25-expressing cells using anti-CD25 MicroBeads (Miltenyi Biotec).

Isolation of human B cells from PBMCs

Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized blood of healthy volunteers by Ficoll gradient centrifugation, and B cells were purified from PBMCs by using anti-CD19 MicroBeads (Miltenyi Biotec, Leiden, The Netherlands) following the manufacturer’s protocol.

In vitro murine and human B cell stimulation

Mouse splenic CD19⁺ B cells (1.5x10⁶/ml) were cultured in medium (RPMI 1640 glutamax; Thermo Fisher Scientific), containing 5% heat-inactivated Fetal Bovine Serum (FBS; Greiner Bio-One, Alphen aan den Rijn, The Netherlands), 5 × 10⁻⁵ M 2-Mercaptoethanol (Sigma-Aldrich) and antibiotics (100 U/mL penicillin and 100 μg/mL streptomycin; Sigma-Aldrich). Human B cells (1.5 x 10⁶/ml) were cultured in medium (RPMI 1640; Thermo Fisher Scientific), containing 10% heat-inactivated Fetal Bovine Serum, pyruvate (1 mM), glutamate (2 mM) and antibiotics (100 U/mL penicillin and 100 μg/mL streptomycin; all Sigma-Aldrich). The following stimuli were added as indicated in the figure legends: SEA (20 μg/ml), SEA depleted of IPSE (SEAΔIPSE, 20 μg/ml), natural (1, 5, 10, 20 μg/ml) or plant-derived IPSE (10 μg/ml), omega-1 (1 μg/ml), kappa-5 (1, 5, 10, 20 μg/ml), AWA (5, 10, 20 μg/ml). For some conditions, co-stimulatory rat anti-mouse CD40 antibody (2 μg/ml; clone 1C10; BioLegend, Uithoorn, The Netherlands) or goat anti-mouse IgM (0.5, 1, 2 μg/ml; Jackson ImmunoResearch, Suffolk, UK) was added to the culture. After 3 days culture at 37°C, supernatants were collected for cytokine analysis by ELISA. Cells were restimulated with PMA (100 ng/ml) and ionomycin (1 μg/ml) for 4 hours in the presence of Brefeldin A (10 μg/ml; all Sigma-Aldrich) for flow cytometric analysis of intracellular IL-10. In experiments addressing involvement of lysosomal acidification, the inhibitor chloroquine (5 μM; Sigma) was added at the start of a two days culture and refreshed after 24 hours, and the TLR ligands CpG ODN 1826 (5 μg/ml; Invivogen) or Pam3Cys (10 μg/ml; Invivogen) used as control stimuli next to SEA.

In vivo experiments

For in vivo stimulation of B cells, mice were i.p. injected with two doses of 5000 eggs or 100 μg SEA in PBS, determined as optimal doses where B cell IL-10 production plateaued in prior dose-titration experiments, and PBS or 100 μg human serum albumin (HSA) in PBS as control 7 days apart. At day 14 after the first injection, splenic B cells were harvested and cultured in medium at 1.5 x 10⁶ cells/ml or restimulated for 2 days with SEA (20 μg/ml) to allow detection of cytokines, as established for in vivo schistosome-exposed B cells before [12]. Supernatants were collected to determine cytokine concentration by ELISA. Cells were cultured for additional 4 hours with Brefeldin A (10 μg/ml; Sigma-Aldrich) to detect intracellular IL-10 by flow cytometry. In some experiments, mice were treated i.p. with 200 μg of hamster anti-mouse CD40 ligand blocking antibody (clone: MR1) or 200 μg hamster IgG as control (Jackson ImmunoResearch) 4 times, every 4 days starting one day before SEA or PBS treatment. For in vivo depletion of macrophages, mice were i.p. injected with 200 μl clodronate-containing liposomes and control mice with 200 μl of PBS liposomes (ClodronateLiposomes.com, Amsterdam, The Netherlands) [45] three weeks prior to egg antigen treatment. Successful and specific depletion of splenic macrophage subsets was confirmed by fluorescence microscopy and flow cytometry.

Co-culture with CD4⁺CD25^- T cells

In vitro or in vivo SEA-stimulated CD19⁺ B cells were co-cultured with MACS-sorted CD4⁺CD25^- T cells at 1:1 ratio (each 1 x 10⁶/ml) to test for in vitro Treg cell induction. After 4 days, Treg cell frequencies were determined by flow cytometry by gating for Foxp3⁺CD25⁺ cells in the CD3⁺CD4⁺ T cell population, and culture supernatants collected for subsequent ELISA.

Antigen binding assays

SEA, IPSE/alpha-1 and ovalbumin (OVA) were fluorescently labeled with PF-488 or PF-647 using the PromoFluor labeling kits (PromoCell, Heidelberg, Germany) according to the manufacturer’s protocol. For some experiments, SEA was co-labeled with the pH-sensitive pHrodo Red dye (Thermo Fisher Scientific). After protein labeling, non-reacted dye was removed using Zeba desalt spin columns (Thermo Fisher Scientific). For analysis of binding in vitro, CD19⁺ splenic B cells were cultured for 60 minutes at 37°C with 20 μg/ml fluorescently labeled SEA or 1–10 μg/ml of IPSE antigen, then washed in ice-cold PBS before analysis by flow cytometry. For analysis of in vivo binding, mice were i.v. injected with 200 μg of fluorescently labeled SEA or OVA as non-schistosomal control protein and spleens snap-frozen 30 minutes to 24 hours later. Binding of SEA to B cells was analyzed by confocal fluorescence microscopy of tissue sections and by flow cytometry.

Induction of IL-4 release from human basophils

Basophils were purified from 250 ml of peripheral blood of healthy human donors to a mean purity of 99% by a three-step protocol consisting of a density gradient centrifugation via Ficoll/Percoll (100/6, density 1.080 g/l), followed by enrichment of the basophils via counter flow elutriation and final purification by magnetic cell sorting using the basophil isolation kit II for negative selection of basophils (Miltenyi-Biotech). Purified basophils were cultured in Iscove's Modified Dulbecco's Media (IMDM; PAA) containing 2 mM glutamine (PAA), 5 μg/ml insulin (Gibco), 50 μg/ml apo-transferrin (Sigma-Aldrich), 100 μg/ml Pen/Strep (PAA), 10% heat-inactivated Fetal Calf Serum (FCS-Gold; PAA) and 2.5 ng/ml IL-3 (kind gift of Kirin Brewery, Japan). Basophils were pre-incubated for regeneration for 30 min at 37°C, 6% CO₂, and then stimulated at a concentration of 0.025 x 10⁶ basophils /ml in 96well flat-bottom culture plates in 100 μl at 37°C, 6% CO₂. Concentration of stimuli was as indicated. Culture supernatants were collected after 18h and stored at -20°C.

Flow cytometry

Flow cytometric analysis of murine B cells was performed by staining with fluorochrome-labeled antibodies against CD19, CD21 (both BD Biosciences), CD23, CD40, CD86 or IL-10 (all eBioscience) after fixation with 1.9% paraformaldehyde and permeabilization with 0.5% saponin (Sigma-Aldrich). Human B cells were stained for CD19, CD38 (both BD Biosciences), CD24, CD27 (both eBioscience), CD1d, IL-10, TNF (all Biolegend), and CD39 (Sony Biotechnology, San Jose, USA). Splenic myeloid cell subsets were discriminated using fluorochrome-labeled antibodies against CD11b, CD11c (both eBioscience), CD8, Ly6C (both Biolegend), F4/80 (AbD Serotec, Puchheim, Germany), Gr-1 (BD Biosciences), and Siglec-1 (Dr. J. den Haan, VUMC, Amsterdam, The Netherlands). Treg cells were fixed and permeabilized with the eBioscience Foxp3 fixation/permeabilization kit and stained using fluorochrome-labeled antibodies against CD3, CD4, Foxp3 (all eBioscience) and CD25 (BD Biosciences). All cells were stained with Aqua dye (Thermo Fisher Scientific) prior to fixation to discriminate dead cells. For all flow cytometric stainings, FcγR-binding inhibitor (2.4G2) was added and FMOs were used for gate setting. Flow cytometry was performed using a FACSCanto or Fortessa (BD Biosciences).

ELISA

The concentration of murine IL-6 and IL-10 as well as human IL-4 present in culture supernatants was quantified by commercial ELISA kits according to the manufacturer’s instructions (BD Biosciences or Eli-Pair, Diaclone).

Confocal microscopy

Spleens were snap-frozen in O.C.T. medium (Tissue-Tek; Sakura, Alphen aan den Rijn, The Netherlands). Cryosections (10 μm) were fixed in ice cold acetone for 10 minutes, air-dried, and blocked in 1% BSA plus 20% FBS in PBS before staining with Abs at room temperature. Cryosections were incubated with rat anti-mouse Siglec-1 (clone SER-4; provided by Dr. J. den Haan, VUMC, Amsterdam, The Netherlands) followed by Alexa555-conjugated goat anti-rat IgG (Invitrogen), anti-SIGN-R1 Alexa647 (clone 22D1; Dr. J. den Haan) and anti-B220 eFluor450 (eBioscience). Images were acquired using a Zeiss LSM 710 confocal laser scanning microscope with Zen software (Carl Zeiss Microimaging, Jena, Germany).

Statistical analysis

All data are presented as mean ± standard error of the mean (SEM). Statistical analysis was performed with GraphPad Prism version 6.00 for Windows (GraphPad Software, La Jolla, CA, USA) using nonparametric Mann-Whitney U test to compare different groups and Wilcoxon paired test to compare B cell subsets. One-sample t-test of log-transformed data was applied to calculate significant changes for data which are expressed as fold increase. All p-values < 0.05 were considered significant.

Ethics statement

All animal studies were performed in accordance with the Animal Experiments Ethical Committee of the Leiden University Medical Center (DEC-12204). The Dutch Experiments on Animals Act is established under European Guidelines (EU directive no. 86/609/EEC regarding the Protection of Animals used for Experimental and Other Scientific Purposes). For the isolation of B cells from PBMCs, human subjects were recruited within the framework of the study P09.170, which was approved by the Medical Ethical Committee of the Leiden University Medical Center. For the isolation of basophils, donors were recruited under approval by the Ethics Committee of the University of Luebeck (AZ-12-202A). Studies were performed according to the declaration of Helsinki and all participants were adults and have given written informed consent.