Impaired cartilage resorption, vascular invasion and endochondral ossification in Sox9-misexpressing mice. (Aa-h) Alcian Blue staining of longitudinal sections through the center of BAC-Col10-Sox9 transgenic (tg) and wild-type (wt) long bones shows that Sox9 misexpression causes substantial inhibition of cartilage resorption, leaving behind cones of non-resorbed hypertrophic cartilage. At E16.5 (a,b) and E18.5 (c,d) the growth plate organization in the transgenic femur is disturbed. At P8 (e,f), resorption of the transgenic cartilage cone has started by bone marrow sprouts invading from the periphery rather from the diaphyis (see also Fig. S4 in the supplementary material). At P18 (g,h), much of the transgenic cartilage cone has been resorbed and the process of secondary ossification of the epiphysis is significantly delayed. (Ba-e) Delayed vascular invasion into Sox9 transgenic hypertrophic cartilage. Immunofluorescence staining for the endothelial marker CD31 (PECAM) shows the beginning of capillary invasion into the hypertrophic cartilage in a E15.5 wild-type (wt) humerus (b,c), but no capillaries in the diaphysis of a transgenic littermate (a). Only at E16.5 do capillaries first appear in the diaphysis of a transgenic (d) tibia (e, wild-type). (Ca-d) TRAP staining reveals retarded invasion of osteoclasts into the diaphysis of a E15.5 Sox9 transgenic femur (a) as compared with a wild-type littermate femur (b). At E18.5, the density of osteoclasts seems reduced in transgenic bones (c) as compared with wild-type littermates (d). (Da-h) Staining for collagen I indicates somewhat disorganized and thickened cortical bone of Sox9 transgenic mice at E19 (a) and P1 (c) as compared with wild-type (b,d) bones, shown here for the femur. At P8 (e-h), the failure of columnar arrangement of hypertrophic chondrocytes and lack of subchondral trabecular bone in a Sox9 transgenic humerus becomes apparent, while the differences in thickness of cortical bone disappear (g,h: distal humerus). Scale bars: 100 μm in Aa-f,Cc,d; 200 μm in Ag,h,De,g; 50 μm in Bc,e,Ca,b; 0.5 mm in Da-d.

Downregulation of Vegfa, Mmp13, OpnN and Runx2 in transgenic hypertrophic cartilage. (Aa-l) E16.5 and (Ba-l) E18.5 tibia. Expression of transgenic Sox9 (Aa,Ba) in the resting and proliferating zones largely coincides with the Col10a1 expression pattern (Ac,Bc) in the transgenic humerus and exceeds the level of endogenous Sox9 in wild-type animals. (Ae,f) Vegfa mRNA is completely absent from the cartilage cone of the transgenic tibia but strongly expressed in the lower hypertrophic zone of wild-type littermates. (Be,f) Starting at E18, in the centre of the epiphysis above the Sox9 overexpressing zone, Vegfa expression is probably upregulated by hypoxic conditions, both in the wild-type and transgenic tibia. Mmp13 mRNA is seen in bone marrow cells but not in hypertrophic chondrocytes of transgenic bones (Ag,Bg), whereas in the wild-type littermate, the majority of Mmp13 signal is located in hypertrophic chondrocytes (Ah,Bh) (see also Fig. 4A). Osteopontin (Opn) is absent from transgenic hypertrophic cartilage (Ai) but expressed in the lower wild-type hypertrophic chondrocytes (Aj). At E18.5, Opn is strongly expressed in periosteal and subchondral tissue both in transgenic (Bi) and wild-type (Bj) animals. The zone of Runx2 expression that is highest in the wild-type prehypertrophic zone (Al,Bl) appears extended in the upper part of transgenic cartilage towards the diaphysis (Ak,Bk). In the lower part of transgenic cartilage cones, however, Runx2 is strongly reduced (Ak,Bk). tg, transgenic; wt, wild-type. Scale bars: 100 μm.

Altered gene expression in BAC-Col10a1-Sox9 transgenic cartilage. (A,B) Higher magnification of Mmp13 expression starting in the diaphysis of E15.5 and E16.5 tibiae showing that Mmp13 is not expressed in transgenic cartilage, only in bone marrow cells. (C,D) Mmp9 is expressed by osteoclasts but not in hypertrophic cartilage; however, the total expression area seems reduced in transgenic cartilage (C). (E,F) Focal expression of Ihh in the transgenic cartilage cone (arrow in E) indicates that cells might be maintained in a prehypertrophic stage. (G,H) Chondromodulin is absent from both wild-type (H) and transgenic (G) hypertrophic cartilage, whereas its expression in epiphyseal cartilage is similar in wild-type and transgenic bones. (I,J) Antibody staining for SOX9 revealed a strong nuclear reaction and enhanced SOX9 protein levels in transgenic hypertrophic chondrocytes (I), confirming the in situ result (see Fig. 3Aa,Ba). In the wild-type growth plate (J), SOX9 protein was only stained in cells of the resting, proliferating and prehypertrophic zones. h, hypertrophic zone (tibia, E18.5); tg, transgenic; wt, wild-type. Scale bars: 50 μm in A,B; 100 μm in E,I,J; 200 μm in G,H.

Quantitative mRNA analysis of epiphyseal and hypertrophic chondrocytes from transgenic and wild-type cartilage by real-time PCR. Runx2, Vegfa, Mmp9, Mmp13 and RANKL were strongly downregulated in hypertrophic chondrocytes prepared from Sox9-transgenic epiphyses in comparison to mRNA levels of wild-type littermates. Sox9 mRNA levels were several fold higher in transgenic hypertrophic chondrocytes than in wild-type chondrocytes, as expected. The data represent typical results from 1 out of 4 independent real-time PCR reactions, each in triplicate, obtained with two different chondrocyte preparations. Runx2 was standardized to cyclophilin mRNA, Vegfa and RANKL to actin, and the others to GAPDH. Standard deviations were calculated from triplicate values. P values were calculated by the Student's t-test; P<0.01 is considered significant. The low levels of Col1a1 in transgenic hypertrophic chondrocytes might indicate that the transgenic cartilage cones prepared from P5 transgenic animals contain less endochondral bone trabeculae than that of wild-type hypertrophic cartilage. Col2a1 mRNA levels in hypertrophic cartilage were not altered in tg chondrocytes; Ihh levels were altogether reduced in tg hypertrophic cartilage despite focal upregulation (see Fig. 4E). Epi, epiphyses; Hyp, hypertrophic chondrocytes; Tg, transgenic; Wt, wild-type.

Direct binding of SOX9 to SRY sites in the VEGFA gene and repression of VEGFA transcription activity. (A) (Top) Genomic map of the human VEGFA gene with cis acting elements (boxes). Numbers indicate the distance (bp) of fragments relative to transcription start site (arrow). (Bottom) Reporter gene constructs containing a +1/379 or a +1/230 bp wild-type (wt) fragment of the VEGFA gene, or a SRY mutant of the +1/230 fragment, followed by an SV 40 large T promoter and firefly luciferase (Luc) gene. The SRY sites are conserved in three species. (Ba-c) (a) SOX9 suppresses the transcriptional activity of the +1/230 luciferase reporter construct in COS7 cells in a dose-dependent manner. (b,c) Reporter gene assays in COS7 cells with various VEGFA reporter gene constructs indicate that the SRY elements in the +1/230 region are responsible for the high transcriptional activity of the +1/230 reporter gene; their activity is reduced after co-transfection with Sox9 (b) and the suppressive effect of SOX9 is reduced after mutation of the SRY element (c). (C) Gel shift assay using recombinant Sox9 and a ³²P-labeled VEGFA promoter-2SRY oligonucleotide (5 fmol). SOX9 binds to the SRY probe (lane 3) and addition of SOX9 antibody (2.5 μg) shows a supershift band (lane 5*). The shift is effectively competed by the addition of the non-labelled wild-type probe (lane 9) but not by the mutant SRY oligonucleotide (lane 10). (Da-c) Transfection of primary mouse rib chondrocytes with Sox9 siRNA enhanced Vegfa expression (a) but caused reduced Sox9 mRNA (b) and protein (c) levels (western blotting). Sox9 and Vegfa mRNA were measured by quantitative RT-PCR; the mean value of three independent experiments is shown. (Ea,b) (a) Chromatin immunoprecipitation (ChIP) analysis shows that SOX9 binds to the SRY sites in the Vegfa gene in mouse primary chondrocytes in vivo. PCR analysis of the precipitate (IP) shows specific binding of the SRY-containing Vegfa promoter region indicated in A in Sox9-HaloTag-transfected cells (+), but not in untransfected cells (−). Total, total genomic DNA of Sox9-transfected and - untransfected cells (control). (b) ChIP of endogenous Sox9 with chromatin from primary mouse chondrocytes isolated from the resting and hypertrophic zones of newborn rib cartilage, using anti-SOX9 antibodies for immunoprecipitation and the same PCR primers as in Ea. Lane1, PCR of total cell lysate; lane 2, PCR of anti-SOX9 immunoprecipitate; lane 3, PCR of non-immune IgG precipitate.