Hdac3 Deficiency Increases Marrow Adiposity and Induces Lipid Storage and Glucocorticoid Metabolism in Osteochondroprogenitor Cells

Abstract

Bone loss and increased marrow adiposity are hallmarks of aging skeletons. Conditional deletion of histone deacetylase 3 (Hdac3) in murine osteochondroprogenitor cells causes osteopenia and increases marrow adiposity, even in young animals, but the origins of the increased adiposity are unclear. To explore this, bone marrow stromal cells (BMSCs) from Hdac3-depleted and control mice were cultured in osteogenic medium. Hdac3-deficient cultures accumulated lipid droplets in greater abundance than control cultures and expressed high levels of genes related to lipid storage (Fsp27/Cidec, Plin1) and glucocorticoid metabolism (Hsd11b1) despite normal levels of Pparγ2. Approximately 5% of the lipid containing cells in the wild-type cultures expressed the master osteoblast transcription factor Runx2, but this population was threefold greater in the Hdac3-depleted cultures. Adenoviral expression of Hdac3 restored normal gene expression, indicating that Hdac3 controls glucocorticoid activation and lipid storage within osteoblast lineage cells. HDAC3 expression was reduced in bone cells from postmenopausal as compared to young women, and in osteoblasts from aged as compared to younger mice. Moreover, phosphorylation of S424 in Hdac3, a posttranslational mark necessary for deacetylase activity, was suppressed in osseous cells from old mice. Thus, concurrent declines in transcription and phosphorylation combine to suppress Hdac3 activity in aging bone, and reduced Hdac3 activity in osteochondroprogenitor cells contributes to increased marrow adiposity associated with aging. © 2015 American Society for Bone and Mineral Research.

Keywords: AGING; HSD11B1; LIPID DROPLETS; OSTEOPOROSIS; RUNX2.

Fig. 1

Mice lacking Hdac3 in osteochondroprogenitor cells have high bone marrow fat. (A) Sections of tibias from 5-week-old to 6-week-old Hdac3 CKO_Osx, Hdac3 CKO_OCN, and corresponding control mice were stained with Goldner’s trichrome. (B) Tibias from 8-week-old male and female Hdac3 CKO_Col2ERT mice and control littermates were histologically prepared with Safranin O/Fast Green for observation of bone architecture. (C) AV/TV (%) and N.Ad/T.Ar were quantified in histologic sections from Hdac3 CKO_Col2ERT and control mice. Means ± SE are shown (n = 4 female control, 5 female Hdac3_Col2ERT CKO, 3 male control, 6 Hdac3_Col2ERT CKO) *p ≤ 0.05 versus control. Other p values are indicated. AV/TV = adipocyte volume per tissue volume; N.Ad/T.Ar = number of adipocytes per tissue area.

Fig. 2

Hdac3-deficient BMSC form lipid droplets when grown in osteogenic medium. (A) BMSCs from 8-week-old Hdac3 CKO_Osx or control animals were cultured in osteogenic medium for 21 days and stained with Oil Red O. (B) BMSCs from 4-week-old Hdac3CKO_Osx animals were cultured in osteogenic medium for 7 days. The abundance of mRNA transcripts for the indicated genes was determined by qPCR. Data represent means ± SE of 3 replicates. *p ≤ 0.05 versus control. (C, E) BMSCs from 4-week-old Hdac3 CKO_Osx mice and control littermates were transduced once (at the time of seeding) with adenoviruses expressing either V5-tagged Hdac3 (Ad-Hdac3) or GFP (Ad-GFP). Cells were grown in osteogenic medium for 7 days and the indicated transcripts were measured by qPCR. Data represent means ± SE of three replicates. Bars with different letters are significantly (p ≤ 0.05) different from one another.

Fig. 3

Runx2-positive cells can contain cytosolic lipid droplets. (A, B) BMSC from 4-week-old Hdac3 CKO_Osx mice or control littermates were grown in osteogenic medium for 14 days and stained with antibodies against Runx2 (red) and Pparγ2 (green) (A) or antibodies against Runx2 (red) and Perilipin 1 (green), with subsequent staining with monodansylpentane (blue), a neutral lipid stain, to highlight lipid droplets (B). Cells from Hdac3 CKO_Osx cultures demonstrated prevalence of Runx2-positive cells that also showed positive staining for intracellular lipid droplets. (C) The number of Runx2-positive/lipid droplet-positive cells in cultures from Hdac3 CKO_Osx and control mice were counted. Data were collected from over 100 cells in each culture and represent the means ± SE of two independent experiments. *p ≤ 0.05 versus control.

Fig. 4

Lipid droplet formation in Hdac3-insufficient cells is dependent upon glucocorticoid signaling. (A, B) BMSCs from 4-week-old Hdac3 CKO_Osx mice and control littermates were transduced once (at the time of seeding) with either Ad-GFP or Ad-Hdac3 and cultured in osteogenic medium for 7 days. Hsd11b1 expression was measured by qPCR (A) and Western blotting (B). (C) C2C12 cells were transfected with Hsd11 b1 reporter constructs and 300 ng pcDNA3 (control) or Hdac3 cDNA. Firefly luciferase values were normalized to Renilla luciferase. Data represent the mean change ±SE of three replicates from each respective control. Results are representative of at least three experiments; *p ≤ 0.05 versus control. (D) BMSC from 4-week-old Hdac3 CKO_Osx mice and control littermates were cultured in osteogenic medium with and without 5 µM carbenoxolone (CBX) for 7 days. Expression levels of Plin1, Cidec, and Pparg were measured by qPCR. Data represent the means ± SE of three replicates. *p ≤ 0.05 versus Control. (E) BMSCs from 4-week-old Hdac3 CKO_Osx mice and control littermates were cultured in osteogenic medium with and without dexamethasone (Dex) for 7 days. Expression levels of Plin1 and Cidec were measured by qPCR. Data represent the means ± SE of three replicates. Bars with different letters are significantly (p ≤ 0.05) different from one another. (F) C2C12 cells were transfected with glucocorticoid-responsive MMTV-luciferase reporter construct and 300 ng pcDNA3 (Control) or Hdac3 cDNA. Cells were treated with 100 nM Dex or vehicle (ethanol) for 24 hours. Firefly luciferase values were normalized to Renilla luciferase. Data represent the mean change ± SE of three replicates from the vehicle-treated control. Results are representative of at least three experiments. *p ≤ 0.05 versus control.

Fig. 5

Osteoblastic Hdac3 expression decreases with age. (A) HDAC3 mRNA levels were measured in bone cores from the ilium of old (64 to 88 years old) and young (22 to 40 years old) women by qPCR. Plots show the median ± interquartile range of 10 independent samples per group. *p= 0.045. (B) Hdac3 mRNA levels were measured in BMSCs from 22-month-old and 1-month-old mice that were cultured in osteogenic medium for 7 days. Data represent the means ± SE of three replicates, and are representative of aging-related changes in at least three independent experiments. *p ≤ 0.05 versus younger mice. (C, D) Cortical bone-derived osteoblasts (C) or BMSC-derived osteoblasts cultured for 14 days in osteogenic medium (D) were lysed for Western blotting with antibodies recognizing pS424-Hdac3, Hdac3, and beta-actin. Relative band intensity (normalized to beta-actin) is expressed beneath each band. (E, F) After growth in osteogenic medium for 14 days, BMSC cultures from 26-month-old and 2-month-old mice were stained with Oil Red O (red), DAPI (blue), and/or antibodies to Runx2 (green) and imaged by epifluorescence and confocal microscopy. (G) BMSCs from 22-month-old and 1-month-old mice were cultured in osteogenic cell culture medium in the presence or absence of dexamethasone. Data represent the means of triplicate samples ± SE. Bars with different letters are significantly (p ≤ 0.05) different from one another.

Fig. 6

(A) Proposed model for interaction between Hdac3, Hsd11b1, glucocorticoid signaling, and lipid storage mechanisms in osteoblasts. Hdac3 suppresses Hsd11b1 expression to control intracellular glucocorticoid activation and lipid metabolism. (B) Model for Hdac3’s role in the differentiation patterns of skeletal cell populations. Hdac3 is required for proper osteoblast and chondrocyte maturation and inhibits the formation of lipid droplets in Runx2+ osteoblast progenitor cells. 11-DHC = 11-dehydrocorticosterone.