Smoothened signaling in the mouse osteoblastoid lineage is required for efficient B lymphopoiesis

The morphogen Hedgehog (Hh) plays critical roles in development and stem cell maintenance.¹  Hh signaling proceeds through the essential activator Smoothened (Smo), which in the absence of ligand is inhibited by the receptor Patched (Ptch); the binding of ligand relieves inhibition of Smo, which in turn stimulates downstream transcription.²  Hh signaling has been implicated in the expansion³  and differentiation^4,5  of hematopoietic stem cells, but cell-autonomous Hh signaling is not required for B lymphopoiesis in the mouse.^6-8  We have shown that extinction of Hh signaling in stromal cells impairs their ability to support B lymphopoiesis from hematopoietic stem progenitor cells in vitro. Moreover, depletion of Smo from stromal cells is associated with downregulation of markers associated with osteoblastoid identity.⁶  Taken together, these observations prompted the hypothesis that B lymphopoiesis in the bone marrow (BM) is promoted by Hh signaling in stromal osteoblasts (OBs). We have now tested this by employing a mouse model in which Smo is selectively removed from the osteoblastoid lineage.

The transgenic mouse strain B6.Cg-Tg(Sp7-tTa, tetO-EGFP/cre)1Amc/J, hereafter termed Osx1-GFP::Cre, was obtained from The Jackson Laboratory (Bar Harbor, ME). Osx1-GFP::Cre^+/0, Smo^fl/fl and Osx1-GFP::Cre^+/0, Smo^+/+ mice were obtained by interbreeding Osx1-GFP::Cre^+/0 and Smo^fl/+ mice (gift of N. Watkins, Garvan Institute). C57BL/6 mice bearing a GFP cassette within the Rag1 locus were gifts of K. Medina (Mayo Medical School, Rochester, MN). All mouse lines were constructed on a C57BL/6 background. For partial BM ablation, mice were treated with 5-fluorouracil (5-FU; APP Pharmaceuticals) at 150 mg/kg by intraperitoneal injection. Mice were housed in accordance with policies of the Johns Hopkins Animal Care and Use Committee.

All antibodies were from BD Biosciences. Data were collected using an LSRII flow cytometer (BD Biosciences) and analyzed with FlowJo v10.1 software (Tree Star).

Lin⁻Sca-1⁺c-Kit⁺ (LSK) hematopoietic progenitors were purified from BM of 6- to 10-week-old mice.⁶  Assays for B-lymphoid differentiation on primary OBs were performed as described elsewhere.^6,9 

Bone cells were isolated from 6- to 10-week-old mice as described previously,¹⁰  except that the final incubation was in medium containing 20% fetal bovine serum, 50 µg/mL ascorbic acid, and 10 mM β-glycerophosphate. OB differentiation was induced as described elsewhere.¹¹  For immunofluorescence analysis, cells were fixed in methanol, incubated with 1.5% blocking serum in phosphate-buffered saline for 1 hour, and then with anti-mouse osteocalcin antibody (Takara) overnight at 4°C. Primary antibody was detected with a phycoerythrin-conjugated secondary antibody.

To test the prediction that Hh signaling in OBs promotes B lymphopoiesis, we deleted Smo by Cre-mediated excision in the mouse osteoblastoid lineage (Figure 1A; supplemental Figure 1A-C, available on the Blood Web site). In these animals, Cre was expressed from the osteoblastoid-specific Osx1 promoter.¹²  OB cells prepared from Osx1-Cre, Smo^fl/fl (Smo^Ob-ko) mice showed homogeneous staining for osteocalcin, a specific OB marker, and the identity of these cells was confirmed by flow cytometry (Figure 1B; supplemental Figure 2). Excision of floxed Smo alleles was observed in OB cells from Smo^Ob-ko mice, but not in OB cells from Osx1-Cre, Smo^+/+ (Cre control) mice or in spleen, thymus, and BM B-lymphoid cells from either strain (Figure 1C). Primary OBs from Cre control mice expressed Ptch1, Smo, Hh ligands, and the Hh target Gli1; OBs from Smo^Ob-ko mice exhibited a 10-fold reduction in Smo transcripts and a similar reduction in the expression of Gli1 (supplemental Figure 3).

Smo^Ob-ko mice exhibited reduced body weight (supplemental Figure 4A). Tooth elongation was impaired (supplemental Figure 4B) and femurs were osteopenic, with thin growth plates and decreased trabecular bone (supplemental Figure 4C). In Smo^Ob-ko animals, splenic B-cell areas (the marginal zone and the periphery of the periarteriolar lymphoid sheath) were reduced, whereas the thymus showed a moderate decrease in cortical thymocytes (supplemental Figure 4D). Despite these abnormalities, the proportions of major lymphoid subsets in spleen and thymus of Smo^Ob-ko mice were normal (supplemental Figure 5A-B); in blood, the proportions of B and T cells or CD4⁺ and CD8⁺ T-cell subsets were likewise unaffected (supplemental Figure 5C-D). In contrast, the percentage and number of steady-state B progenitors at pre-B and later developmental stages (supplemental Figure 1D)¹³  were reduced in the BM of Smo^Ob-ko mice (supplemental Figure 5E-F), whereas the proportion of myeloid progenitors was unaltered (supplemental Figure 5G).

To identify the developmental stage at which B lymphopoiesis is impaired in Smo^Ob-ko animals, we examined B progenitors at steady state and during recovery from ablation. Early lymphoid progenitors (ELPs) and common lymphoid progenitors (CLPs) are marked by accumulation of Rag-1 transcripts before stable acquisition of lineage identity.¹⁴  We introduced a Rag-1 allele bearing a GFP cassette (Rag-1^GFP) onto Smo^Ob-ko and Cre control backgrounds (supplemental Figure 1C). In Smo^Ob-ko, Rag-1^GFP mice, BM pro-B and pre-B cells were significantly reduced (Figure 1D; supplemental Figure 6A), indicating an impairment of B lymphopoiesis at least as early as the pre-/pro-B to pro-B transition. The percentages and numbers of total GFP⁺ (Rag-1⁺) cells and GFP⁺, B220⁺ (Rag-1⁺ B220⁺) cells were also greatly diminished in Smo^Ob-ko, Rag-1^GFP mice, whereas under these steady-state conditions, GFP⁺, B220⁻ cells, which represent ELPs/CLPs, showed no obvious difference among genotypes (Figure 1E; supplemental Figures 6B and 7A).

Primary OB cells from Cre control and Smo^Ob-ko mice were tested for their ability to support B-lymphoid differentiation from LSK progenitors in vitro. Control OB cells robustly supported B lymphopoiesis from LSK cells of wild-type, Cre control, or Smo^Ob-ko origin (Figure 1F, left), whereas Smo-depleted OB cells were deficient in their support of B lymphopoiesis (Figure 1F, right). This indicated that Hh signaling in primary OB cells promotes B lymphopoiesis and suggested a causal relationship between the extinction of Hh signaling and the paucity of B-lymphoid progenitors in Smo^Ob-ko mice.

We examined recovery of BM B progenitors after chemical ablation by 5-FU (Figure 2A). At 7 days, pro-B cells were 20- to 30-fold less abundant in Smo^Ob-ko mice than in wild-type and Cre control animals (Figure 2B; supplemental Figure 6C). By day 14, pre-B cells had emerged but were markedly less numerous in Smo^Ob-ko mice than in controls (Figure 2C; supplemental Figure 6D). The relative sizes of B progenitor subsets at day 21 (Figure 2D; supplemental Figure 6E) were similar to those observed at day 14, suggesting that repopulation had reached steady state. In contrast, myeloid repopulation was unimpaired in Smo^Ob-ko mice (Figure 2E; supplemental Figure 6F).

At 7 days after BM ablation, the total GFP⁺ population was reduced in Smo^Ob-ko, Rag-1^GFP mice relative to controls (Figure 2F, left; supplemental Figure 6G, left), consistent with a paucity of B progenitors (Figure 2B). Strikingly, ELPs/CLPs were overrepresented in Smo^Ob-ko animals (Figure 2F, middle; supplemental Figure 6G, middle; supplemental Figure 7B). The distributions of B progenitors at 7 days after ablation were similar to those in mice lacking the Rag-1^GFP allele (Figure 2B,F; supplemental Figure 6C,G). These results confirm that extinction of Hh signaling in the osteoblastoid compartment impairs the emergence of pro-B cells. Moreover, the increase in accumulation of ELPs/CLPs in Smo^Ob-ko, Rag-1^GFP mice at 7 days after ablation suggests that Hh signaling in OB cells promotes the exit of B-lymphoid progenitors from the CLP compartment.

Taken together, our results indicate that Hh signaling in osteoblastoid cells promotes B lymphopoiesis; this requirement may be enforced as early as the transition from the CLP to the pre-/pro-B cell stage. Deletion of Ptch1, an inhibitor of Smo, either nonspecifically or in stratified epithelia has been associated with depletion of B- and T-cell progenitors.¹⁵  Our study differs with respect to the lineage specificity of the deletion and the lymphopoietic defect, which was restricted to the B lineage. Although Osx-1-Cre–mediated deletion may occur in other BM cell types,¹⁶  previous studies uncovered a critical role for OBs in B-cell development,^11,17  and we demonstrate that OBs from Smo^Ob-ko mice are defective in Hh signaling and in promoting B lymphopoiesis. Our findings indicate that the ability of osteoblastoid cells to promote B lymphopoiesis is conferred, at least in part, by Smo, whose sustained activity is likely to support expression of factors essential for B-lineage development.

The authors thank Kay Medina (Mayo) for the gift of Rag-1-GFP mice and members of the Department of Molecular Biology and Genetics for stimulating discussions.

This work was supported by National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases grant R01DK099188 (S.D.).

Contribution: W.L. designed and executed experiments, interpreted results, and wrote the manuscript; D.D. designed and executed experiments and interpreted results; D.H. performed histopathologic analysis; and S.D. designed experiments, interpreted the results, and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

David Huso died on 27 January 2016.

Correspondence: Stephen Desiderio, 733 North Broadway, Baltimore, MD 21205; e-mail: sdesider@jhmi.edu.