Myeloid lineage skewing due to exacerbated NF-κB signaling facilitates osteopenia in Scurfy mice

Abstract

Immune surveillance through Foxp3+ regulatory T cells plays a crucial role in bone homeostasis. Scurfy, the mouse model of autoimmune IPEX syndrome, bears a loss-of-function mutation in Foxp3 that leads to multi-organ inflammation. Herein, we report that scurfy mice exhibit severe bone loss mediated by accelerated osteoclastogenesis. Mechanistically, Foxp3 deficiency results in the upregulation of NF-κB in T helper cells through the loss of repressive Foxp3/NEMO interaction, thereby unleashing NF-κB-mediated over-production of pro-osteoclastogenic cytokines. Flow cytometry analysis shows marked increase in lin-Sca-1+c-kit+ hematopoietic stem cells (LSK HSCs) and granulocyte/macrophage progenitors (GMPs) in bone marrow of scurfy mice with corresponding exacerbated osteoclastogenic potential, implying that osteoclast progenitors are affected at a very primitive stage in this disorder. Scurfy LSK HSCs exhibit greater sensitivity to M-CSF and contain abundant PU.1+ Sf LSK HSCs compared with WT. Accordingly, genetic or pharmacological inhibition of M-CSF or mTOR signaling, but not IL-17 signaling, attenuates osteoclastogenesis and osteopenia in scurfy. Thus, our study suggests that Foxp3 deficiency leads to osteopenia owing to dysregulated NF-κB activity and subsequent cytokine-mediated hyper-proliferation of myeloid precursors, and positions the NF-κB pathway as a potential target for therapeutic intervention for this disorder.

Figure 1

Sf mice exhibit severe bone loss and increased number of osteoclasts in bone. (a) Representative 3-D reconstructed images of the trabecular bone region in the proximity to the growth plate were captured by μCT analysis. (b) Representative images of the proximal and midshaft cortex of the long bones are shown. (c–f) Quantification of bone parameters: (c) μCT analysis depicts ratio of cancellous bone volume/tissue volume (BV/TV) in the same areas shown in panel a, (d) trabecular numbers (Tb N), (e) trabecular separation (Tb Sp), (f) and trabecular thickness (Tb Th). (g–j) Histological examination of osteopenia in Sf mice. (g) Representative H&E- and TRAP-stained histological section near the growth plate. (h) Quantifications of osteoclast number per surface area were plotted for the trabecular bone region. (i) Histological section of ankle bones. (j) Quantifications of osteoclast number per surface area were plotted for the sections near the ankle region. (k) WBM cultures from WT and Sf mice were subjected to vehicle or RANKL stimulation for 4 days before TRAP staining to visualize osteoclast formation. (l) Bar graphs were plotted by counting numbers of multi-nulceated (MNC) TRAP⁺ osteoclasts (OCs) from each of the 4 wells of each RANKL-stimulated groups. All bar graphs were depicted by the average of the numbers, ±S.D. Statistics: n=4 per group; *P<0.05 and **P<0.01 by Student T-test. Scale bars: 200 μm in (a), (i) and (k); 1 mm in (b); 100 μm in (g)

Figure 2

Foxp3 deficiency in Sf mice resulted in the activation of NF-κB pathway and over-production of pro-osteoclast cytokines in CD4⁺ T lymphocytes. (a) Western blot of NEMO and Rel A protein expression in splenic CD4⁺ T cells using WT and Sf-derived CD4+ T cells. Images shown were blots probed with anti-Rel A and anti-NEMO Ab separately. β-actin was used as internal loading control. (b) Intracellular flow analysis was performed on red-cell-depleted WBMs. Cells were fixed, permeabilized and stained with anti-NEMO Ab followed by FITC-conjugated secondary Ab and analyzed on flow cytometer. Histogram of the middle fluorescence intensity (MFI) of FITC was shown. (c) Co-immunoprecipitation of Foxp3 and NEMO protein in primary splenic CD4⁺T cells. CD4⁺T-cell proteins derived from WT and Sf pulled down by anti-NEMO Ab were blotted with either anti-NEMO or anti-Foxp3 Ab. Total lysate from Foxp3⁺Tregs used as positive control. (d) Co-immunoprecipitation of Foxp3 and NEMO protein in transiently transfected 293 T cells. 293 T cells were transfected with both Foxp3 and NEMO (tagged with c-myc) plasmids. Cells transfected with either Foxp3 or NEMO plasmid alone were used as controls. Immunoprecipitation (IP) and immunoblots (IB) were conducted as indicated. In addition, two clones of anti-Foxp3 Ab were used for IP, by which K-13 generated stronger immunoprecipitation than H190. (e) qPCR analysis (fold change) of the indicated factors in MACS purified CD4⁺ T cells from WT and Sf mice. (f) Cytokine production following PMA stimulation was assessed by intracellular flow analysis on MACS-purified, BM-derived CD4⁺ T cells using appropriate antibodies. (g) Intracellular flow analysis on MACS-purified, PMA-stimulated spleen-derived CD4⁺ T cells. (h) qPCR analysis (fold change) of M-CSF expression in MACS-purified, anti-CD3/CD28 stimulated CD4⁺ T cells from WT and Sf mice. Bar graphs were depicted by the average of the numbers, ±S.D. Statistics for all bar graphs: n=3 per group; *P<0.05 and **P<0.01 by Student T-test

Figure 3

Bone marrow CD11b^-, not CD11b⁺, cells from Sf are highly proliferative and osteoclastogenic. (a) Flow analysis was performed on bone marrow myeloid populations from 12 Sf and 11 WT controls stained with anti-CD11b and anti-Gr1 antibodies. Statistics was generated by Student T-test. (b) BrdU incorporation in Sf WBM cultures was carried out as described under methods. (c) Bar graph derived from BrdU flow assays shown in panel b. n=4 per group; *P<0.05 by Student T-test. (d) MACS was performed first with biotin-conjugated lineage antibodies and anti-biotin microbeads followed by anti-CD11b microbeads to obtain lin^-CD11b^-/lo and lin^-CD11b^Hi fractions. Osteoclatogenic potential of each MACS fraction was subsequently tested by ex vivo osteoclastogenesis assay. WBM cultures were used as control. Scale bars: 200 μm in (d)

Figure 4

Alteration of hematopoietic progenitor pool in Sf bone marrow. (a, b) Flow cytometry of HSC progenitors was performed as described in Material and Methods. Percentage of LSK hematopoietic progenitor cells were first gated and analyzed, followed by further sub-gating of the lin^-c-kit⁺ cell population (LK) into megakaryocyte/erythroid progenitors/CMP/GMPs. (c) LT-HSC, ST-HSCs and mutipotent progenitors were sub-gated from LSKs for analysis. Bar graphs were depicted by the average of the numbers, ±S.D. Statistics for all bar graphs: n=6 for WT group and n=7 for Sf group; *P<0.05 and **P<0.01 by Student T-test

Figure 5

Sf LSK HSCs are potent osteoclast founders. One thousand FACS-sorted LSK and MPs (lin^-c-kit⁺CD11b^lo described in Supplementary Figure S4) from WT and Sf bone marrows and mixed with 9000 total bone marrow cells from WT and treated with RANKL. WBM cultures were used as control. Bar graph was depicted by the average of OC numbers, ±S.D. Statistics; n=3 per group; *P<0.05 and **P<0.01 by Student T-test. Scale bar: 200 μm

Figure 6

Neutralization of M-CSF significantly rescued osteopenia in Sf mice. (a) Sf and WT mice were injected with M-CSF-neutralizing Ab or PBS as described in Material and Methods. Cross-sections of femur bone extending from the growth plate were generated by 3-D reconstructed images through μCT analysis. BV/TV ratios were depicted as bar graph. (b) Ex vivo osteoclastogenesis assay post in vivo administration of M-CSF-neutralizing Ab. (c) Flow cytometric analysis of WBM hematopoietic progenitors were performed on WBMs derived from M-CSF Ab-administrated Sf mice with the rest of experimental control mice. Statistics for all bar graphs: n=3 for WT group and n=4 for Sf group; *P<0.05 and **P<0.01 by Student T-test. Scale bars: 1 mm in (a); 200 μm in (b)

Figure 7

Rescue of the osteopenic phenotype in Sf via genetic ablation of M-CSF. (a) Representative μCT scanned 3-D images of the femur trabecular area of WT, Sf and Op/+:: Sf mice. (b) Ex vivo osteoclastogenesis assay were performed using WBMs from WT, Sf, Op/+ and Op/+::Sf mice. (c) Frequency of LSK HSCs in Op/+ Sf mice using flow cytometry. (d) Profile of megakaryocyte/erythroid progenitors/CMP/GMPs of WBMs derived from WT, Sf, Op/+ and Op/+ Sf mice, *P<0.05 and **P<0.01 by Student T-test. Scale bars: 200 μm in (a and b)

Figure 8

In vivo administration of rapamycin ameliorated osteopenia in Sf mice through reduction of osteoclastogenic activity in the bone marrow. (a) Representative images from μCT analysis of rapamycin versus vehicle-treated Sf mice. (b) Trabecular BV/TV values of vehicle- and rapamycin-treated mice. (c) Ex vivo osteoclastogenesis assay. (d) Hematopoietic progenitor cell profiling by flow cytometry depicting megakaryocyte/erythroid progenitors, CMP and GMP cells from the various experimental groups described in panel a. Statistics for all bar graphs: n=4 for per group; *P<0.05 and **P<0.01 by Student T-test. Scale bars: Scale bars: 200 μm in (a and c)