Fig 2: DOT1L negatively regulates Wnt signalling in articular cartilage.(a) KEGG pathway enrichment analysis of microarray data obtained from human articular chondrocytes treated with EPZ-5676 or vehicle. Nominal P values by EASE modified Fisher Exact test using the DAVID analysis tool (see Methods section) are shown. n=5 independent patient-derived cell cultures. (b) Co-immunoprecipitation (Co-IP) using an anti-DOT1L antibody showing interaction between DOT1L and ß-catenin in human articular chondrocytes, that is increased upon Wnt activation by LiCl and disrupted upon DOT1L inhibition. The image is representative of three experiments. (c) TOP/FOP reporter assay in human articular chondrocytes after Wnt stimulation by LiCl and DOT1L inhibition by EPZ. Activity is compared to untreated cells (dotted line). n=3 biologically independent experiments. ***P<0.001 by one-way ANOVA. (d,e) LEF1, TCF1 and c-MYC expression measured by quantitative PCR in chondrocytes treated with EPZ-5676 and LiCl (d) or in LiCl-treated chondrocytes transfected with siRNA directed against DOT1L or scrambled siRNA (siDOT1L or siSCR, respectively) (e). Data are from one experiment with three technical replicates. (f) Immunohistochemistry demonstrating increased TCF1 levels in the articular cartilage of C57/Bl6 wild-type mice after injection of EPZ-5676. The images are representative of three different animals. Scale bar, 200 µm.

Fig 3: Clinical implications of modulating the DOT1L network in cartilage.(a–c) Inactivation of SIRT1 protects against DOT1L inhibitor-induced osteoarthritis: (a) C57/Bl6 wild-type mouse knees were injected with DOT1L inhibitor EPZ-5676 (5 mg kg-1) and SIRT1 inhibitor EX527 (1.25 mg kg–1), or vehicle (V) and killed after 4 weeks. Knees were sectioned and stained with Hematoxylin-Safranin O (a). Scale bar, 200 µm. Cartilage damage was scored (see Methods section) and is shown in (b). One experiment was performed with n=3 (vehicle), 8 (EPZ) and 10 (EPZ+EX527). *P<0.05 by one-way ANOVA. Error bars indicate mean±s.e.m. (c) Immunohistochemistry of TCF1 in the indicated groups. TCF1 levels are increased after EPZ treatment and normalized by additional EX527 treatment. The images are representative of three different animals. Scale bar, 200 µm. (d,e) Loss of DOT1L function causes severe growth retardation as demonstrated by skeletal staining (d) and histology of the growth plate (e) of 4-week-old Dot1lfl/fl;Col2-Cre-/- (Cre-neg) and Dot1lfl/fl;Col2-Cre+/- (Dot1lCart-KO) mice. (f,g) Increased TCF1 levels in Dot1lCart-KO mice as shown by immunohistochemistry in the indicated mice strains in the articular cartilage (f) and growth plate (g). The images are representative of three different animals. Scale bar, 200 and 100 µm.

Fig 4: Loss of DOT1L disrupts chondrocyte homeostasis and triggers osteoarthritis.(a) Immunohistochemistry showing reduced methylated H3K79 levels (H3K79me2) that reflect loss of DOT1L activity in damaged areas from osteoarthritic patients (OA) as compared to their corresponding preserved areas and to cartilage from non-OA patients. Images are representative of images from four different patients. Scale bar, 400 µm. (b) Heat maps of differential mRNA expression determined by quantitative PCR in chondrocytes treated with DOT1L inhibitor EPZ-5676 (EPZ) or vehicle (V) from passage 0 (P0) until P2, and from preserved versus damaged areas in OA cartilage. The colour code represents the mean expression level of six and four independent patient samples respectively. (c) Immunoblot analysis showing decreased methylated H3K79 levels in mouse articular chondrocytes after intra-articular injection of EPZ into C57Bl/6 wild-type mouse knees. The image is representative of one experiment with protein extracts pooled from two or three mice per condition. Unprocessed original scans of blots are shown in Supplementary Fig. 10. (d,e) C57/Bl6 wild-type mouse knees were injected with EPZ (5 mg kg–1) or vehicle and killed after 2 or 4 weeks. Knees were sectioned and stained with Hematoxylin-Safranin O (d). Scale bar, 200 µm. Cartilage damage was scored (see Methods section) and is shown in (e). One experiment was performed with n=10 and 5. Representative images from the 4 week evaluation are shown. *P<0.05 (two-tailed t-test). Error bars indicate mean±s.e.m.

Fig 5: DOT1L directly controls Wnt targets by negative regulation of SIRT1.All experiments were performed in healthy human articular chondrocytes: treated as indicated with DOT1L inhibitor EPZ-5676, Wnt activator LiCl, SIRT1 antagonist EX527 or SIRT1 agonist SRT1720; or transfected with DOT1L or scrambled siRNA. All data are presented as mean±s.e.m. (a) Chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) analysis of DOT1L and methylated H3K79 and (b) acetylated H3K9 (H3K9Ac) and methylated H3K4 (H3K4me3) as markers of active transcription on the transcriptional start site (TSS) of Wnt target genes. Data are from two to five experiments. (c) Expression levels of TCF1 Wnt target gene measured by quantitative PCR in chondrocytes transfected with indicated specific or scrambled siRNA (siSCR). Data are from one experiment with technical triplicates. (d) TCF1 expression measured by quantitative PCR in the presence of SIRT1 agonist and antagonist. Data from two experiments each with technical triplicates. (e) Co-IP analysis using the indicated antibodies demonstrating the interaction of DOT1L and SIRT1. The image is a representative image of three biologically independent experiments. (f) SIRT1 activity relative to vehicle-treated cells (dotted line). Data are from three biologically independent experiments. *P<0.05, ***P<0.001 by one-way ANOVA. (g) ChIP-qPCR analysis of SIRT1, PPARGC1A, GCN5 and EP300 binding on the TCF1 promoter and (h) TCF1 expression after siRNA transfection with indicated specific or scrambled siRNA. Data from two biologically independent experiments.