ESET histone methyltransferase is essential to hypertrophic differentiation of growth plate chondrocytes and formation of epiphyseal plates

Abstract

The ESET (also called SETDB1) protein contains an N-terminal tudor domain that mediates protein-protein interactions and a C-terminal SET domain that catalyzes methylation of histone H3 at lysine 9. We report here that ESET protein is transiently upregulated in prehypertrophic chondrocytes in newborn mice. To investigate the in vivo effects of ESET on chondrocyte differentiation, we generated conditional knockout mice to specifically eliminate the catalytic SET domain of ESET protein only in mesenchymal cells. Such deletion of the ESET gene caused acceleration of chondrocyte hypertrophy in both embryos and young animals, depleting chondrocytes that are otherwise available to form epiphyseal plates for endochondral bone growth. ESET-deficient mice are thus characterized by defective long bone growth and trabecular bone formation. To understand the underlying mechanism for ESET regulation of chondrocytes, we carried out co-expression experiments and found that ESET associates with histone deacetylase 4 to bind and inhibit the activity of Runx2, a hypertrophy-promoting transcription factor. Repression of Runx2-mediated gene transactivation by ESET is dependent on its H3-K9 methyltransferase activity as well as its associated histone deacetylase activity. In addition, knockout of ESET is associated with repression of Indian hedgehog gene in pre- and early hypertrophic chondrocytes. Together, these results provide clear evidence that ESET controls hypertrophic differentiation of growth plate chondrocytes and endochondral ossification during embryogenesis and postnatal development.

Fig. 1. ESET gene structure, genotyping and transient upregulation in prehypertrophic chondrocytes

a, diagrams of ESET protein domains and the floxed allele. b, PCR products for WT, floxed-ESET alleles and Cre transgene were shown along with the DNA size markers. c, proximal tibia in newborn mice was stained with an anti-ESET antibody to show transient upregulation of ESET protein in prehypertrophic chondrocytes in wild-type animal and absence of ESET staining in knockout littermate.

Fig. 2. Skeletal defects in ESET-deficient newborn mice

a, photograph of wild-type and (exons 15&16)^CKO/CKO mutant at birth. b, newborn skeletal preparations were stained with alizarin red for comparison of bone elements. Double staining with alizarin red (bone) and alcian blue (cartilage) was carried out for the skull (c), rib cages (d), forelimbs (e) and hind limbs (f). Genotypes of the embryos are indicated on the top.

Fig. 3. Disorganization of growth plates in ESET-deficient mice

a, the humerus of newborn wild-type and (exons 15&16)^CKO/CKO littermate were stained by H&E, and representative regions are enlarged in the corresponding panels. b, H&E staining of femurs from newborn wild-type and (exons 15&16 )^CKO/CKO littermate, with typical areas enlarged for better comparison. c, newborn distal humerus was stained by H&E, by an anti-active Caspase 3 antibody then mounted in DAPI medium to show apoptotic cells as well as general morphology.

Fig. 4. Ectopic expression of type × collagen in ESET-deficient mice

a, proximal humerus, distal femur and proximal tibia from E15.5 embryos were stained with an antibody against type × collagen. Genotypes of the embryos are indicated on the top. b, growth plates in proximal humerus, distal femur and proximal tibia from newborn mice were stained with the same antibody against type × collagen. Note that type × collagen-positive chondrocytes are limited to the hypertrophic zone in wild-type animal, and diffusion of type × collagen-positive cells is observed throughout the growth plate in the knockout littermate.

Fig. 5. Long bone defects and absence of epiphyseal plates in ESET-deficient mice

a, Safranin O staining of cartilage in the distal femur of mice at different stages of postnatal development. b, 5-week old wild-type and mutant mice were exposed to the same x-ray film, with enlargement of the forelimb exposure. Arrows indicate positions of epiphyseal plates at the ends of long bones. c-d, forelimbs and hind limbs from these 5-week old mice were stained with alcian blue and alizarin red to show cartilage and long bone defects in the mutant.

Fig. 6. Trabecular bone defects in ESET-deficient adult mice

a, whole skeletal μCT scan of one month-old wild-type and (exons 15&16)^CKO/CKO littermate. Arrow indicates the incomplete fontanelle closure in the knockout skull. b, 3-D reconstruction of one month-old tibia from wild-type and knockout mice. Position of the epiphyseal plate is indicated by an arrow. c, longitudinal view through the mid-point of the tibia. d-e, decalcified sections of proximal tibia from one month-old mice were stained by H&E to show general cell morphology, and by anti-osteocalcin to show differences in trabecular bone. Note absence of the epiphyseal plate in the wild-type animal. Genotypes are indicated on the top.

Fig. 7. Physical and functional interactions of ESET with Runx2 and HDAC4

a, nuclear extract from 293T cells that co-express Flag-ESET and HA-Runx2 were incubated with a mouse monoclonal anti-HA, a mouse monoclonal anti-Flag, normal mouse IgG, or a mouse monoclonal anti-HDAC4. The immunoprecipitates were then blotted with HRP-conjugated anti-Flag, anti-HA or a goat anti-HDAC4 for detection of the co-immunoprecipitated proteins. b, COS-7 cells were co-transfected with the mOG2-Luc firefly luciferase reporter plus Runx2 and ESET plasmids, and trichostatin A was added 24 hrs post transfection. Total amount of DNA in each transfection was kept constant by the addition of empty vector DNA, and firefly luciferase activities were normalized according to the Renilla luciferase controls. Shown are results from three independent experiments. Western blotting of lysates from a typical experiment was carried out to show comparable protein expression in each sample.

Fig. 8. Disruption of Ihh expression in ESET deficient mice

a, sections of proximal tibia from wild-type newborn mice were stained by H&E to show cell morphology, by an anti-Ihh antibody to show Ihh protein expression in pre- and early hypertrophic chondrocytes, by DAPI to show distribution of nuclei within the growth plate. The anti-Ihh staining was merged with the DAPI image for better localization of Ihh-positive cells. b, sections of proximal tibia from newborn (exon 15&16)^CKO/CKO mutant were stained by H&E, anti-Ihh and DAPI to show decreased Ihh expression in the growth plate.