Role of glucocorticoid-induced leucine zipper (GILZ) in bone acquisition

Abstract

Glucocorticoids (GCs) have both anabolic and catabolic effects on bone. However, no GC anabolic effect mediator has been identified to date. Here we show that targeted expression of glucocorticoid-induced leucine zipper (GILZ), a GC anti-inflammatory effect mediator, enhances bone acquisition in mice. Transgenic mice, in which the expression of GILZ is under the control of a 3.6-kb rat type I collagen promoter, exhibited a high bone mass phenotype with significantly increased bone formation rate and osteoblast numbers. The increased osteoblast activity correlates with enhanced osteogenic differentiation and decreased adipogenic differentiation of bone marrow stromal cell cultures in vitro. In line with these changes, the mRNA levels of key osteogenic regulators (Runx2 and Osx) increased, and the level of adipogenic regulator peroxisome proliferator-activated receptor (PPAR) γ2 decreased significantly. We also found that GILZ physically interacts with C/EBPs and disrupts C/EBP-mediated PPARγ gene transcription. In conclusion, our results showed that GILZ is capable of increasing bone acquisition in vivo, and this action is mediated via a mechanism involving the inhibition of PPARγ gene transcription and shifting of bone marrow MSC/progenitor cell lineage commitment in favor of the osteoblast pathway.

Keywords: Arthritis; Glucocorticoid; Inflammation; Osteoblast; Osteoporosis.

FIGURE 1.

Characterization of GILZ transgenic mice. A, schematic diagram showing the DNA construct used for generation of GILZ transgenic mice. Approximate locations of primer pairs used for genotyping (P1 and P2) and qRT-PCR (P3 and P4) are indicated. The expected sizes of PCR products are 500 and 216 bp, respectively. B, genotyping and copy number estimation. Tail genomic DNAs from a WT and three transgenic founders (nos. 196, 47, and 198) were analyzed by PCR. A serial diluted Col3.6-GILZ-HA plasmid DNA containing indicated copies of plasmid was mixed with tail genomic DNA from a WT mouse and used as copy number reference. Glucagon was used as an internal control for equal loading of genomic DNA from the transgenic mice. C and D, real time qRT-PCR and Western blot analysis of GILZ mRNA (C) and protein (D) expression in bone tissues. E–H, immunohistochemical staining of paraffin-embedded bone sections showing the expression of GILZ in bone tissues. The arrows indicate strong GILZ-positive osteoblast-like cells lining on the surface of bone. The scale bars represent 100 μm. Original magnification, ×200 (E–H).

FIGURE 2.

DXA and μ-CT analyses. A and B, left femurs of 3-month-old GILZ Tg and WT littermate control mice were assessed by DXA scan for BMD (A) and BMC (B). C–F, representative three-dimensional reconstructed images of femoral samples from WT and GILZ Tg mice (males) at 3 months (C and D) and 6 months (E and F) of age. G–L, quantitative results showing BV/TV (G), Tb.N (H), Tb.Th (I), Tb.Sp (J), Conn. D (K), and structure model index (SMI, L) at 3 and 6 months of age. The data are shown as means ± S.D. n = 7–12 for DXA and n = 5–8 for μ-CT analysis, respectively. *, p < 0.05; **, p < 0.01; ***, p < 0.001. M, months.

FIGURE 3.

Dynamic analysis of bone formation. Calcein was injected intraperitoneally into 3-month-old male mice 14 and 4 days prior to sacrifice. A–D, representative fluorescent images of plastic-embedded femoral sections of WT and GILZ Tg mice. C and D are enlarged images of boxed areas in A and B, respectively. E–G, quantified results showing mineral apposition rate (MAR, E), mineralizing surface (MS/BS, F), and bone formation rate (BFR, G) of cortical bone. The data are shown as means ± S.D. (n = 7). *, p < 0.05; **, p < 0.01; ***, p < 0.001. The scale bars represent 100 μm. Original magnification, ×40 (A and B). H and I, quantified results showing cortical thickness (H) and represent μ-CT images of midshaft (I). n = 4; p = 0.012.

FIGURE 4.

Histological and histomorphometric analyses of femurs in 3-month-old male GILZ Tg and WT mice. A–D, representative hematoxylin- and eosin-stained femurs from WT (A and C) and GILZ Tg (B and D) mice. C and D are enlarged images of boxed areas in A and B, respectively. Marrow adipocyte is indicated by stars. E–G, bone histomorphometric analysis results showing the numbers of osteoblasts (E), osteoblast surface (F), and the numbers of marrow adipocytes (G). The data are shown as means ± S.D. (n = 9–11). *, p < 0.05; **, p < 0.01. The scale bars represent 100 μm. Original magnification, ×40 (A and B). M, months.

FIGURE 5.

In vitro bone marrow cell culture and gene expression. A–F, CFU assays. Whole bone marrow cells (each well represents cells from one individual mouse) were cultured in either regular growth media DMEM or in osteogenic or adipogenic induction media. On day 14, cells were fixed and stained with Giemsa for CFU-f (A and B) or with alkaline phosphatase (ALP) substrate (NBT/BCIP) for CFU-Ob (C and D). CFU-Ad assay was performed with a higher cell seeding density (1 × 10⁶ cells/well of 6-well plate). After 2 days of induction, the cells were switched to maintenance media and stained with Oil red O (E and F) on days 10–15 when lipid-filled adipocytes appeared. To further confirm CFU-ob results, osteogenic differentiation experiments were also performed in 96-well plates with a higher initial seeding density and induced for 6 (G) and 21 days (H), respectively, to visualized alkaline phosphatase-positive cells and mineralized bone nodules (ARS stain). I and J, real time qRT-PCR analysis of key genes controlling MSC lineage commitment in cultured bone marrow cells. Whole bone marrow cells from GILZ Tg and WT mice were cultured for 6 days with or without osteogenic induction and analyzed for the expression of Runx2, Osx, and PPARγ2. Expression levels of GILZ mRNA were also examined. The data were calculated using the relative expression software tool and presented as relative expression of target gene in Tg samples over WT control samples. *, p < 0.05; **, p < 0.001.

FIGURE 6.

TRAP stain of osteoclasts. A–D, representative images of TRAP-stained tibiae from 3-month-old GILZ Tg and WT mice. C and D are enlarged images of boxed areas in A and B, respectively. E and F, quantitative results showing the numbers and surface of osteoclasts. G, TRAP stain showing in vitro osteoclast differentiation of BMMs isolated from 3-month-old GILZ Tg and WT mice. H, quantitative results showing the average numbers of osteoclasts in G. I, real time qRT-PCR showing relative levels of GILZ mRNA in BMMs treated with or without RANKL. The data are shown as means ± S.D. The scale bars represent 100 mm. M, months; CTRL, control.

FIGURE 7.

GILZ interacts with C/EBPs. A, co-immunoprecipitation assay. 293T cells were cotransfected with HA-GILZ and Flag-C/EBPα, -β, or -δ. Whole cell lysates were immunoprecipitated with anti-Flag and detected with anti-HA antibody to show co-precipitated GILZ protein. The expression of C/EBPs and GILZ in transfected cells is shown by Western blots using anti-Flag antibody and anti-HA antibody, respectively. B, EMSA. Affinity-purified GST-C/EBPβ and His-GILZ protein was incubated with a 30-bp IRDye-labeled DNA probe containing tandem repeat C/EBP binding site either alone (lanes 1 and 3) or together (lanes 5, 6, and 8). The DNA-protein complex was resolved in native polyacrylamide gel and imaged using an Odyssey Infrared Imaging System. Lane 1 contains 8 μg of His-GILZ, lane 2 contains 8 μg of GST, lane 3 contains 0.5 μg of GST-C/EBPβ, lane 4 contains 8 μg of GST plus 8 μg of His-GILZ, lanes 5–8 contain 0.5 μg of GST-C/EBPβ plus His-GILZ (8 and 12 μg in lanes 5 and 6, respectively). Lane 7 contains 0.5 μg of GST-C/EBPβ and 2 μl of anti-C/EBPβ antibody. Lane 8, contains the same as lane 7 plus 8 μg of His-GILZ. C, DNA pulldown assay showing interactions between GILZ, C/EBPβ, and PPARγ promoter. Cell lysates harvested from GILZ- and GFP-expressing MSCs were incubated with biotin-labeled DNA probe mentioned above, and the protein-DNA complex was purified with magnetic beads conjugated to avidin. The complex was separated on SDS-PAGE, transferred onto membrane, and then detected with antibodies against C/EBPβ or GILZ. The specificity of protein-DNA interaction was confirmed by adding 60× excess of unlabeled DNA probe as indicated. D, GST pulldown assay showing interaction between GILZ and C/EBPs. Purified GST-C/EBPs were immobilized on glutathione-Sepharose beads and then incubated with His-tagged GILZ; after several washes, the bound proteins were eluted by boiling in sample buffer, separated by SDS-PAGE, and detected with anti-GST or anti-His antibodies. E and F, working model illustrating how GILZ may inhibit C/EBP-mediated PPARγ gene transcription. In the absence of GILZ (E), dimerized C/EBPs bind to C/EBP-binding sites and associate with co-activators such as PGC-1 and initiate transcription. However, in the presence of GILZ (F), the association of C/EBPs with co-activators and general transcription machinery (GTM) is disrupted, thus C/EBP-mediated PPARγ gene transcription is inhibited. IB, immunoblot; IP, immunoprecipitation;