B-1 cells and B-1 cell precursors prompt different responses to Wnt signaling

Recently several studies demonstrated a role for the Wnt pathway in lymphocyte development and self-renewal of hematopoietic stem cells (HSCs). B-1 cells constitute a separate lineage of B lymphocytes, originating during fetal hematopoiesis, expressing lymphoid and myeloid markers and possessing self-renewal ability, similar to early hematopoietic progenitors and HSCs. A plethora of studies have shown an important role for the evolutionary conserved Wnt pathway in the biology of HSCs and T lymphocyte development. Our previous data demonstrated abundant expression of Wnt pathway components by B-1 cells, including Wnt ligands and receptors. Here we report that the canonical Wnt pathway is activated in B-1 cell precursors, but not in mature B-1 cells. However, both B-1 precursors and B-1 cells are able to respond to Wnt ligands in vitro. Canonical Wnt activity promotes proliferation of B-1 cells, while non-canonical Wnt signals induce the expansion of B-1 precursors. Interestingly, using a co-culture system with OP9 cells, Wnt3a stimulus supported the generation of B-1a cells. Taking together, these results indicate that B-1 cells and their progenitors are differentially responsive to Wnt ligands, and that the balance of activation of canonical and non-canonical Wnt signaling may regulate the maintenance and differentiation of different B-1 cell subsets.


Introduction
B-1 cells constitute a subpopulation of B cells, mostly found in the peritoneal cavity and rarely in the spleen. Besides possessing B lineage markers (CD19 HI IgM HI IgD lo ), their expression is quantitatively different from conventional B cells (CD19 + IgM lo IgD HI ), and B-1 cells also express the myeloid marker CD11b. Further, B-1 cells are subdivided in two subtypes: B-1a and B-1b owing to CD5 expression in the former. This heterogeneity is also seen in the progenitor populations of these cells. Ample evidence supports the existence of distinct ontogenic lineages, in which B-1a cells are generated largely from fetal liver progenitors, and maintained in adult life mostly by self-renewal, while B-1b cells, albeit having self-renewal, can be generated de-novo via HSCs located in the BM [1][2][3][4][5][6]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 maintained under pathogen free conditions. All procedures described here were approved by the Ethical Committee from UNIFESP (2012/712).

Characterization of B-1 and B-1P cells
Cells from the peritoneal cavity or bone marrow from Axin2 +/lacZ Wnt-reporter mice and C57BL/6 wild type (WT) mice were used. The peritoneal cells were harvested by washing out the peritoneal cavity using RPMI1640 medium. The bone marrow cells were isolated from both femurs of each animal, and clear bones were crushed using a mortar and pestle. The crushed bones were rinsed with RPMI1640 medium, and the supernatant were collected and filter using a 40 μm cell strainer. After that, cells were counted and pre-stained with anti-CD16 CD32 mAb to block Fcγ RIII/II receptors and stained on ice for 30 min with the monoclonal antibodies against the following molecules: CD19, CD23, CD11b, CD5 to characterize B-1a (CD19 + CD23 --CD11b +/-CD5 + ) and B-1b (CD19 + CD23 -CD11b +/-CD5 -) cells subsets from peritoneal cavity and with CD3e, CD4, CD11b, Gr-1, IgM, NK-1.1, Ter119, CD45R/B220, CD19 and CD93 (Early B-AA4.1) to characterize B-1 progenitors (B-1P -Lin -CD19 + EarlyB + B220 lo/neg ) from bone marrow. B-1 cells and B-1 cell precursors were also stained with anti-Flt3 and IL7-R antibodies to determine the expression of these receptors. Cells were acquired using the BD FACSCanto™II flow cytometer and data were analyzed with FlowJo software (S1 and S2 Figs).

Enrichment of B-1 and B-1P cells
B-1 cells population were cell sorted by BD FACSAria III from mice's peritoneal cells. First, cells were collected and processed as described above. Two strategies were used to purify B-1 cells population: negative or positive selection. For the negative selection strategy, cells were stained with CD3 and CD23 antibodies. From the lymphocyte gate, a double negative population (CD3 -CD23 -) was sorted. After that, an aliquot of sorted cells was fully stained to confirm the B-1 cell purity. For the positive selection, CD19 and CD23 antibodies were used, and the CD19 + CD23population from the lymphocyte gate was sorted. In all experiments the B-1 cell purity was around 95% after cell sorting.
B-1 and B-1 cell precursor enriched population were obtained from pooled sorting cells from 5-7 mice, and were considered one biological sample. Each experiment was performed with 2-3 biological samples as indicated in each Fig.

B-1 cell culture
Purified B-1 cells were cultivated in RPMI medium added with 10% fetal calf serum (FCS), 5x10 -5 M 2-β-mercaptoethanol, 1 mM L-glutamine, 100 U mL -1 streptomycin, 100 μg mL -1 penicillin. When indicated in the text, recombinant Wnt3a and Wnt5 (100ng/ml) were added daily for 72 hours. In other experiments, IL7 was also added to the cultures. After this, B-1 cells were collected and submitted to the experiments protocols as described below.
an Applied Biosystems 7500 Fast Real-Time PCR System. The amplification efficiencies were determined by comparing the dilution series of reference and target genes from a reference cDNA template. The amplification efficiency was calculated using the following equation: E = 10 (−1/slope) − 1, in which E is the efficiency and slope is the value obtained by constructing standard curve. A validation was performed to evaluate if the efficiencies of the target and the reference gene were approximately equal (90% E 110%). If the target and the reference genes had comparable amplification efficiencies, relative quantification was determined according to the 2 −ΔΔCt or 2 −ΔCt method, as indicated in each Fig [23,24]. Each reaction was carried out in triplicate using at least three biological samples. The sample used as normalizer was bone-marrow derived total or control B-1 cells, as indicated in each experiment.

Proliferation analysis
After purification, B-1 cells and B-1 precursors were stained using 5 μM of Cell Proliferation Dye eFluor1 670, following the manufacture instructions and submitted to different cell culture conditions, as indicated in each experiment. The maximum of CFSE staining cells in time zero were considered to determine the region gate of non-proliferative cells. To measured the decay of fluorescence, which is not related to proliferation, B-1 cells or B-1 cell precursor cultured in a RPMI medium only was used. The decay of fluorescence in these samples was subtracted to the decay of fluorescence in the experimental groups to determine the gate region of fluorescence cells and also MFI.

Canonical Wnt signaling evaluation
Cells from Axin2 +/lacZ Wnt-reporter mice were obtained as described above (item 2) and the Wnt signaling was evaluated by measurement of the β-galactosidase activity (lacZ), as previously described [25,26]. Briefly, up to 5x10 6 cells suspension was loaded with 2 mM of fluorescein di β-D-galactopyranoside (FDG, Molecular Probes) by hypotonic shock. After one minute, 10x volume of ice-cold medium was added to restore the isotonicity. The reaction was stopped two hours later by adding 1 mM of PETG (phenylethyl β-D-thiogalactopyranoside), following the surface markers staining. The cells were acquired using the BD FACSCanto™II flow cytometer and data were analyzed with FlowJo software. In all experiments cells from WT mice were submitted to the same treatment to determine negative and positive gates. In order to calculate the percentage or absolute number of FDG+ cells of each subset, FDG+ amount was subtracted from amount of FDG+ cells in the WT mice in order to correct differences in background staining.

Wnt components expression by B-1 cells
In order to consider if B-1 cells were able to respond to Wnt signaling, we evaluated the expression of AXIN2, FZD receptors and other Wnt target genes. We first compared bonemarrow-derived total cells and purified peritoneal B-1 cells for expression of Wnt components. As expected, AXIN2 levels were lower in B-1 cells than in bone-marrow derived total cells ( Fig  1A). Despite of B-1 cells express all FZD receptors, LRP5, LRP6, ROR1, ROR2 and RYK; we detected higher levels of expression of FZD6 and LRP6 (Fig 1B). Considering this, we could assume that B-1 cells are capable to respond to Wnt ligands.
In the next step, we investigated if B-1 cells could be responsive to Wnt ligands. Purified B-1 cells were stimulated in vitro with recombinant Wnt3a (100ng/ml) or Wnt5a (100ng/ml) proteins daily. After 72 hours we observed that the Wnt3a stimulus augmented the expression of AXIN1, AXIN2, LEF1, FZD6, FZD9 and decreased the expression of ROR2. Wnt5a augmented the expression of AXIN2, FZD9 and ROR2 ( Fig 1C). These data suggest that the presence of Wnt3a augmented the responsiveness of B-1 cells to canonical Wnt signaling, while Wnt5a augmented at least the expression of ROR2.

Activation of Wnt signaling by WNT3a induces B-1 cell proliferation and increases expression of IL7R in vitro
Considering that the presence of Wnt ligands in the B-1 cell culture stimulated the expression of some genes that prompt B-1 cells to respond to them, our next step was to investigated if Wnt stimuli could modulate cell proliferation. We demonstrated that Wnt3a increased the proliferation of B-1 cells in vitro (Fig 2A, 2B and 2C) and also increased the expression of AXIN2 (Fig 2D), which demonstrates that the Wnt pathway is activated by this stimulus.
Interestingly, stimulation with Wnt5a did not induce any change in B-1 cell viability or proliferation (data not shown). Additionally, Wnt3a augmented FLT3 and IL7R gene expression in B-1 cells (Fig 3A and 3B), but did not modify PAX5 expression (Fig 3C). At protein level we confirmed that Wnt3a induced an increase in the IL7-R expression by B-1 cells, but not Flt3 expression (Fig 3D and 3E).

Wnt3a increases B-1 cell responsiveness to IL7
Considering this, we decided to investigate if the increment in the proliferation in the presence of Wnt3a could be not a direct effect of Wnt ligand stimulus, but due to an induction of IL7-R expression. Corroborating this hypothesis, a higher number of IL7R + cells were detected in Wnt3a treated group (Fig 4A) in comparison to control group (NT-non-treated). As observed in the Fig 4A, 68,5% of B-1 cells from the Wnt3a treated group expressed IL7R, while only 40% of B-1 cells are IL7R + in control group (NT). Corroborating on this, the absolute number of IL7R + B-1 cells are a approximately 6.67x10 4 cells in Wnt3a group and almost 2.5 times less in control group (2.69 x10 4 cells) (Fig 4B). Considering that the absolute number of B-1 cells in the culture after 72 hours of Wnt3a stimulus is higher than control group, we investigated the proliferation and viability of these cells. Wnt3a did not modify the cell viability in culture, however we observed that proliferation index is higher after this treatment. Furthermore, IL7R + B-1 cells are more proliferative than IL7R -B-1 cells, at least in the presence of Wnt3a (Fig 4A-lower panel and Fig 4C). It was observed that both IL7Rand IL7R + B-1 cells are more proliferative in the presence of Wnt3a in comparison to control group. Besides, in the Wnt3a conditions, IL7R + cells are more proliferative than IL7Rcells, as measured by the decay of CFSE fluorescence (Fig 4C).
To elucidate this, proliferation of B-1 cells were analyzed in the presence of Wnt3a only, IL-7 only or Wnt3a + IL-7. Expectedly, B-1 cell proliferation is enhanced by addition of IL- 7 recombinant, as well as in addition of Wnt3a stimulus (Fig 4D and 4E). The addition of Wnt3a+IL7 promotes higher levels of proliferation, which suggests that increased IL7-R expression caused by Wnt3a could promote B-1 cell proliferation in vitro via IL7 signaling. Whether this increase is due to increased expression of IL7-R on all B1 cells or preferentially on cells already expressing IL7-R via an autocrine loop remains to be elucidated. Nevertheless, our current findings are mostly consistent with Wnt3a acting via the IL7/IL7R axis to induce B-1 cell proliferation.

Wnt signaling is not activated in B-1 cells in vivo
The canonical Wnt signaling in B-1 cells was evaluated using the Axin2 +/lacZ Wnt-reporter mouse [28]. No difference in the amount of peritoneal B and T lymphocytes was observed in these mice compared to wild type strain. In both, B-1 cells correspond approximately to 60% of total lymphocytes (22% of total cells from peritoneal cavity (S3 Fig) in the peritoneal cavity. To assess activation of canonical Wnt signaling, Axin2 expression was measured by the βgalactosidase activity, using the fluorescein di β-D-galactopyranoside (FDG) as a fluorogenic β-galactosidase substrate. Axin2 activation was detected in less than 3% of B-1 cells (Fig 5A-5B), indicating that the canonical Wnt signaling was not activated in these populations under steady state conditions (ex vivo). Despite of this, we observed activation of Wnt signaling in the B-1 cell precursor population (Fig 5C-5D). In agreement with the literature, more mature stages of B cells have reduced Wnt signaling levels, while this pathway is active in the pre-proB stages (12).

Wnt3a induce B-1 precursors to originated B-1a cells in vitro
To further investigate if the Wnt pathway is more activated in the early stages of B-1 cells, the activity was assessed in the B-1 cell precursors in vitro [29]. A co-culture experiment using Wnt-transduced OP9 cells, which constitutively express Wnt3a (OP9-Wnt3a) and Wnt5a (OP9-Wnt5a) [27] was performed. Considering the initial input (3x10 3 cells), B-1P cells cultivated with OP9-WT increased almost 15x (43x10 3 ± 8 cells) and on OP9-Wnt5a more than 40x (93x10 3 ± 10 cells). The number of B-1P cells in the OP9-Wnt3a co-culture was roughly maintained at 2.5x10 3 cells (Fig 6A and 6B). Furthermore, B-1P cells cultivated in the presence of OP9-Wnt5a mostly lost the expression of AA4.1 (Fig 6C). However, the expression of AA4.1 is sustained in B-1P cells cultivated in the presence of OP9-WT and OP9-Wnt3a. This data is suggestive that Wnt5a induced proliferation and differentiation in the B-1 precursors.
Additional flow cytometer analysis on B-1 precursors differentiated in vitro showed that only in the Wnt3a treated group CD5+ cells (B-1a) were generated, despite the lack of proliferation showed earlier (Fig 7). In order to separate B-1 cell precursor from a population of B-1 cells that could be generated in vitro, AA4.1+ cell population was excluded from the next analysis. From the AA4.1population, expression of CD19 and CD5 was analyzed to determine the frequency of B-1a (CD19 + CD5cells) and B-1b (CD19 + CD5cells). It is important to mention that all CD19 + cells were also IgM + (data not shown). Based on this analysis, B-1 cell precursors Wnt3a promotes B-1 cell self-renewal and Wnt5a induces B-1 precursors proliferation co-cultivated with OP9-WT and OP9-Wnt5a showed only 1,5% of CD19 + CD5 + cells (Fig 7A  and 7C), while in the presence of Wnt3a 36% of CD19 + cells are also CD5+. However it is important to remember that the total number of cells in these latter cultures was reduced. Considering this, Wnt5a induced a pronounced expansion of B-1 cell precursors in vitro, which did not occur when these precursors were in the presence of Wnt3a (Fig 6). However, an interesting generation of CD5+ B-1 (B-1a) cells was observed in the presence of Wnt3a (Fig 7).

Discussion
Here we report that the responsiveness of B-1 and B-1 precursors to Wnt ligands is different, which points to a role for the Wnt pathway in the regulation B-1 cell fate. As Malhotra et al [14] described on conventional B cells, at least two Wnt ligands can differentially regulate the B lymphopoiesis: Wnt3a and Wnt5. In summary, results described here demonstrated that: 1) The canonical Wnt pathway is not activated in B-1 cells in vivo, but it is in the B-1 progenitors (B-1P FDG+ cells were detected); 2) Wnt3a induces an increase in IL-7R expression followed by increased in B-1 cell proliferation in vitro; 3) Wnt5a induces an expansion of B-1 cell precursors, while Wnt3a is able to induce de novo B-1a generation in vitro.
The lower frequency of FDG + B-1 cells is in accordance with data from literature, which show that the responsiveness to Wnt ligands diminishes along hematopoietic development, so

Fig 4. Wnt3a stimulation increment proliferation of B-1 cells in response of IL7.
Purified B-1 cells from C57BL/6 mice were treated with Wnt3a (100ng/ml), IL-7 (50ng/ml) and Wnt3a+IL7 during 3 days. Non-treated cells (NT) were used as control group. A) Contour plots analysis of expression of CD19 and IL7R by non-treated B-1 cells or Wnt3a-treated B-1 cells. Histograms show the proliferation (CFSE decay) of IL7R + (blue) or IL7R -(red) B-1 cells in control group or Wnt3a group. The maximum of CFSE staining cells in time zero were considered to determine the region gate of nonproliferative cells. To measured the decay of fluorescence, which is not related to proliferation, B-1 cells cultured in a RPMI medium only was used. The decay of fluorescence in these samples was subtracted to the decay of fluorescence in the experimental groups to determine the gate region of fluorescence cells and also MFI. B) Absolute number of B-1 cells IL7Ror IL7R + cells from non-treated group (NT) or Wnt3a treated group (Wnt3a). C) Proliferation of IL7Ror IL7R + B-1 cells from non-treated group (NT) or Wnt3a treated group (Wnt3a) measured by decay in the MFI value. D) Histograms of CFSE decay (proliferation) of B-1 cells in the different groups: NT (non-treated), Wnt3a, IL-7, Wnt3a+IL7. E) Proliferation of B-1 cells was represented by MFI value in different conditions: NT (non-treated), Wnt3a, IL-7, Wnt3a+IL7. ÃÃÃ p<0,001, ÃÃ p<0,01 and Ã p<0,05 (One way ANOVA). Data from a representative of 3 independent experiments performed in A,B and C and 2 independent experiments performed in D and E. Each experiment was performed using 3 biological samples.
https://doi.org/10.1371/journal.pone.0199332.g004 Fig 5. Canonical Wnt signaling is activated in B-1 precursors, but not in B-1 cells. B-1 progenitors from bone marrow and peritoneal B-1 cells from Axin2+/lacZ and WT mice were isolated and the canonical Wnt signaling analyzed based on β-galactosidase (FDG+) activity. The percentage (A-C) and absolute number (B-D) FDG+ of each cell population were then calculated. n = 3 mice per group. Ã p 0,01 (Mann-Whitney test). WT mice not carrying the reporter transgene (Axin2+/lacZ) were used to define the FDG − population. In order to calculate the percentage or absolute number of FDG+ cells of each subset, FDG+ amount was subtracted from amount of FDG+ cells in the WT mice in order to correct differences in background staining. https://doi.org/10.1371/journal.pone.0199332.g005 Wnt3a promotes B-1 cell self-renewal and Wnt5a induces B-1 precursors proliferation it is predictable that canonical Wnt pathway could be detected in B-1 cell progenitors, but not later in B-1 cells.
Nevertheless, we demonstrated that both B-1 cells and B-1 cell precursors are responsive to Wnt ligands in vitro. B-1 cells proliferate in response to Wnt3a, and also augment the expression of IL7-R. Considering that IL7 is an important factor for proliferation of B cells, we could postulate that up regulation of expression of IL7R could be a mechanism related to an increment in the proliferation index of B-1 cells in the presence of Wnt3a. We also speculate that Wnt3a could be an important factor for the maintenance of self-renewal of B-1 cell population in the peritoneal cavity. Unfortunately, no data was found in the literature about the expression of Wnt ligands in the peritoneum in normal and health conditions that could support this hypothesis. Conversely, Wnt5a stimulus did not modify the B-1 cell proliferation activity in vitro. It is widely believed that Wnt5a could antagonize the canonical Wnt pathway. Liang et al [30] demonstrated that Wnt5a inhibits the pro-B cell response to IL7, via noncanonical Wnt/Ca +2 pathway. We here showed that in a Wnt3a enriched milieu, proliferation of B-1 cells could be favored by augmented IL7R expression, in accordance with the notion that Wnt5a signals diminish the responsiveness to IL7R. , the OP9-Wnt3a cell line has over 1,000 fold higher expression of Wnt3a than OP9-WT (nontransduced). It is important to mention that the timing and concentration of ligand exposure is a determinant to reflect the effect of Wnt pathway on the cell behaviour. The generation of B-1a cells could be a result of differentiation of B-1 precursors into B1a cells, but also it is possible that Wnt3a exposure in our co-cultures expanded rare and preexisting cells B-1a cells that could not be detected at the start of the culture.
We observed a marked expansion of B1-P progenitors in the presence of Wnt5a, accompanied by loss of AA4.1 expression. Previous reports demonstrated that murine CD19 + and human CD34 + CD38fail to differentiated into CD19 + B cell lineage on OP9-Wnt3a co-cultures, but were favored on OP9-Wnt5a and OP9-Dkk1 ones [14,31]. It could be considered Wnt3a promotes B-1 cell self-renewal and Wnt5a induces B-1 precursors proliferation that the bone marrow microenvironment is rich in Wnt5a [16] and it could support the B cell development. Reya et al [16] also demonstrated that bone marrow stromal cells expressed Wnt5a, but not Wnt10b and Wnt3a, which would support B cell development. In this context, bone-marrow could sustain the maintenance of B-1 cell precursor by expression of Wnt5a.
Despite of the reduced overall number of cells, we demonstrated that only in the presence of Wnt3a, B-1 cell precursors give rise to B-1a cells in vitro. Yoshimoto et al [3] described the emergence of B-1 progenitors before the HSC stage, at embryonic day 9.0-9.5 from yolk sac and intraembryonic para-aortic splanchnopleura (PSp) tissues. Recently the Herzenberg group showed that fetal liver CD150 -, but not CD150 + LT-HSCs were able to reconstitute the B-1a cell population [6], reinforcing the separate origins of B-1 cell progenitors. Interestingly, it has been demonstrated that the Wnt signature of stromal cells in the adult and fetal bone marrow-derived mesenchymal stromal cells are different. Therefore, Wnt ligands in the fetal liver and bone marrow could result in differential responsiveness of the precursors, and the balance of Wnt ligands could influence B-1 population expansion and govern the B-1 cell precursor development.
Based on this, we postulate that the canonical Wnt pathway could be important in the development of B-1 cell precursors, and somehow could interfere in the proliferation of B-1 cells in response to IL7. It could be suggested that Wnt5a regulates the expansion of B-1 progenitors in the adult bone marrow, while Wnt3a could interfere in the generation of B-1a cells in the bone marrow and also in the proliferation of the B-1 cells, perhaps by controlling the self-renewal activity. It is intriguing to conceive that HSCs must have an intrinsic control to maintain the steady-state hematopoiesis during lifetime, and be able to promote reconstitution of cell populations after an injury. Similar mechanisms may operate for B-1 cells. Moreover, a loss of control in the self-renewal or differentiation process could lead to generation of malignant cells which for B-1 cells could lead to B-CLL, the malignant counterpart of normal B-1 cells. It remains unclear whether Wnt signaling is indispensible for B-1 cell development. However our results shed some light on this issue and show how members of this pathway could determine the development and maintenance of B-1 cells. Whether such regulation of B-1 cell lineage fate decisions by Wnt signaling also occurs in vivo, waits for complicated loss-of-function models that are currently unavailable, especially for the non-canonical pathway. Given the cross talk and redundancy in the Wnt pathway, this question is a challenging one to address in vivo.