Autoamplification of NFATc1 expression determines its essential role in bone homeostasis

Abstract

NFATc1 and NFATc2 are functionally redundant in the immune system, but it was suggested that NFATc1 is required exclusively for differentiation of osteoclasts in the skeletal system. Here we provide genetic evidence that NFATc1 is essential for osteoclast differentiation in vivo by adoptive transfer of NFATc1(-/-) hematopoietic stem cells to osteoclast-deficient Fos(-/-) mice, and by Fos(-/-) blastocyst complementation, thus avoiding the embryonic lethality of NFATc1(-/-) mice. However, in vitro osteoclastogenesis in NFATc1-deficient cells was rescued by ectopic expression of NFATc2. The discrepancy between the in vivo essential role of NFATc1 and the in vitro effect of NFATc2 was attributed to selective autoregulation of the NFATc1 gene by NFAT through its promoter region. This suggested that an epigenetic mechanism contributes to the essential function of NFATc1 in cell lineage commitment. Thus, this study establishes that NFATc1 represents a potential therapeutic target for bone disease and reveals a mechanism that underlies the essential role of NFATc1 in bone homeostasis.

Figure 1.

In vivo evidence for the essential role of NFATc1 in osteoclast differentiation. (A) Radiographic analysis of Fos ^−/− mice transferred with NFATc1 ^+/− or NFATc1 ^−/− FLCs. Severe osteopetrosis persists after NFATc1 ^−/− FLC transfer. (B) Microradiographic analysis of femur (left: microcomputed tomography; right: microradiograph). Bone marrow cavity is filled with unresorbed bone in NFATc1 ^−/− FLC chimera. (C) Histologic examination of tibia (top: toluidine blue staining; middle: tartrate-resistant acid phosphatase [TRAP] staining for detection of osteoclasts and lumbar vertebra; bottom: toluidine blue staining). No osteoclasts are observed in NFATc1 ^−/− FLC chimera. (D) Histomorphometric evaluation of osteoclasts in NFATc1 ^+/− and NFATc1 ^−/− FLC chimera. (E) Reconstitution of the hematopoietic system of chimeric mice by the donor cells. PCR primers specific for joint region between the NFATc1 locus and the neomycin-resistant gene (NFATc1 neo) and for wild-type (WT) NFATc1 gene are used. The similar chimeric ratios were confirmed by quantitative PCR. (F) Complete blockade of in vitro osteoclastogenesis in osteoclast precursor cells from NFATc1 ^−/− FLC chimera. Splenocyte-derived osteoclast precursor cells were cultured in RANKL (50 ng ml⁻¹) and M-CSF (10 ng ml⁻¹). We counted multinucleated (>3 nuclei) cells (MNCs) positive for TRAP staining. (G) Radiographic analysis of neonates generated by Fos ^{− / −} blastocyst complementation. Note the difference in radio-opacity at the femur (arrowheads). Right panels show the magnified view of the femur of neonates and tooth eruption of 6-wk-old mice. All of the Fos ^−/−/NFATc1 ^−/− chimeric mice exhibited osteopetrosis, but osteoclastogenesis is restored in Fos ^{− / −} /NFATc1 ^{+ / +} chimeric mice when ES cells are transmitted to the hematopoietic system of recipient mice. **P < 0.001 versus control.

Figure 2.

In vitro compensation of NFATc1 deficiency by forced expression of NFATc2. (A) Phylogenetic tree analysis for the DNA binding (Rel homology) region of NFAT family proteins (top). A wide-range genomic view of NFATc1 and NFATc2 genes suggests that these genes were generated by chromosomal gene duplication (bottom). (B) Loss-of-function analyses of NFATc2 in osteoclast differentiation. Microradiographic analysis of femur derived from NFATc2 ^−/− mice at 5 wk of age (see Fig. 1 B). (C) Histologic examination of tibia from NFATc2 ^−/− mice (left: toluidine blue staining; right: tartrate-resistant acid phosphatase staining [TRAP]). (D) Recovery of osteoclastogenesis in NFATc1 ^−/− FLCs by retrovirus-mediated expression of NFATc1 or NFATc2. Forced expression of NFATc2 as well as that of NFATc1 induces formation of osteoclasts with pit-forming activity on dentin slices (RANKL, 50 ng ml⁻¹). Infection efficiency was monitored by GFP, which is expressed bicistronically. **P < 0.001 versus mock.

Figure 3.

Autoamplification of NFATc1 during osteoclastogenesis. (A) GeneChip analysis of mRNA expression of NFATc1 and NFATc2 in RANKL-stimulated BMMs. Strong induction of NFATc1 by RANKL is inhibited by FK506 (2.5 μg ml⁻¹), which inhibits NFAT activity. NFATc2 expression is constitutive and is not affected by FK506. (B) Semiquantitative RT-PCR analysis of the mRNA of NFATc1 isoforms in BMMs. NFATc1/A is induced ∼10-fold by RANKL. (C) Immunoblot analysis of NFATc1 isoforms during RANKL-induced osteoclastogenesis in BMMs. A Daudi lymphoma cell line is shown as a positive control for all three isoforms (A, B, and C) of NFATc1 (22). (D) A schematic illustration of putative transcription factor binding sites in the 5′ flanking region of the NFATc1 (20) and NFATc2 genes. Binding sites in the 5′ flanking region of the NFATc2 gene were determined by the TRANSFAC retrieval program. Arrows indicate primers for ChIP experiments. (E) ChIP assay of NFATc1 and NFATc2 promoters in RANKL-stimulated BMMs. (F) Luciferase assay of NFATc1 and NFATc2 promoters in HEK293T cells.

Figure 4.

Epigenetic regulation of NFATc1 promoter underlies the selective autoamplification of NFATc1. (A) Analysis of chromatin modification–related factors in NFATc1 and NFATc2 promoters by ChIP in RANKL-stimulated BMMs during osteoclastogenesis. (B) ChIP assay for acetylated histone and methylated histone H3 lysine 4. (C) ChIP assay for MeCP2 and methylated histone H3 lysine 9. (D) A schematic diagram of three stages of osteoclast differentiation that are governed by NFATc1. (i) NFATc2 is recruited to the NFATc1 promoter at the very early phase, but this is not enough to activate the NFATc1 promoter. The binding of RANKL to receptor activator of NF-κB (RANK) results in the recruitment of TNF receptor–associated factor (TRAF) 6, and leads to the activation of downstream molecules, such as NF-κB (25, 34). Cooperation of NFATc2 and NF-κB activates the initial induction of NFATc1, but because NFATc2-deficient mice have no obvious defect in osteoclast differentiation, unknown factor(s) (shown as X) may compensate for the loss of NFATc2 in these mice. (ii) RANKL–RANK interaction cooperates with immunoreceptors to activate the calcium signals (28), which stimulate the NFATc1 activation by way of calcineurin (5). NFATc1 bind to its own promoter, which leads to the robust induction of NFATc1; AP-1 (containing c-Fos) is critical for this autoamplification. Selective recruitment of NFATc1 to the promoter of NFATc1, but not NFATc2, is explained in part, by epigenetic regulation. (iii) Several osteoclast-specific genes, such as cathepsin K, TRAP, and calcitonin receptor, are activated by a transcriptional complex that contains NFATc1 and other cooperators, such as AP-1, PU.1, and microphthalmia-associated transcription factor (MITF).