Genome-wide occupancy of UTX and histone modifications define distinct classes of UTX target genes. (A) Distribution of UTX occupancy in association with H3K27 and/or H3K4 methylation. (B) UTX target genes with H3K4 methylation tend to be expressed, whereas other classes are silent. Box plot showing 25th, 50th, and 75th percentile expression levels for different target gene categories. Whiskers show the 2.5 and 97.5 percentiles. (C) Gene Ontology terms enriched in genes with occupancy of UTX and H3K4me2. (D) Distinct gene cluster families are enriched for UTX occupancy and bivalent histone H3 modifications (protocadherins) or neither histone modification (olfactory receptors).

UTX regulates many RB-binding proteins, which are coordinately decreased in expression in human cancers. (A) Top network of genes occupied by UTX organized by Ingenuity Pathway Analysis. (Center) Forty-nine UTX-occupied genes are known to be associated with RB. The solid line indicates known physical interaction, while the dashed line indicates known interaction by regulation of expression. Genes of special interest are in large font. (B) Validation of UTX occupancy of RB pathway genes by ChIP-qPCR. CDKN2D is not a UTX target gene and served as negative control. (C) The UTX–RB gene network is coordinately repressed in human cancers. We interrogated a compendium of 1973 microarrays representing 22 human tumor types and diverse normal controls for coordinate regulation of the genes occupied by UTX. In the bottom matrix, each column is a sample showing significant coordinate induction (red) or repression (green) of UTX network genes (P < 0.05, FDR < 0.05); each row is a UTX network gene. The middle bar shows the average level of activity of the UTX–RB network in each sample. The top matrix displays whether the coordinate induction or repression of UTX–RB network is enriched for specific clinical annotations. Each brown hatch mark indicates a sample with the indicated annotation; each row is an enriched annotation (P < 0.05, FDR < 0.05, hypergeometric distribution) that shows selective deactivation of the network in cancers relative to their normal counterparts. (D) RB pathway gene expression is correlated with UTX. Genes are sorted and ranked based on Pearson correlation to UTX. Gene set enrichment analysis of the Pearson correlation of UTX-bound genes in 295 human breast cancer patients shows that UTX-bound genes are significantly positively correlated with UTX expression (P < 0.001). (E) UTX expression level is a predictor of patient survival. Kaplan-Meier analysis of patients with high versus low UTX expression is shown for the 295 breast cancer patients (160 patients have “high” UTX expression; 135 patients have “low” UTX expression).

UTX is necessary and sufficient to mediate cell cycle arrest of primary human fibroblasts. (A) Depletion of UTX protein by siRNAs. (B) UTX depletion causes increased H3K27me3 levels at UTX-occupied sites of target genes. ChIP-qPCR of H3K27me3 at the indicated promoters is shown; HOXA9 and GAPDH served as positive and negative controls, respectively. (C) UTX depletion causes decreased mRNA levels of HBP1 and RBBP6 as assayed by qRT–PCR. (D) UTX depletion increases cell proliferation. Serial cell counts of fibroblasts transfected with one of two siRNAs targeting UTX versus control siRNA targeting GFP. (E) Increased S-phase entry of UTX-depleted cells. BrdU immunofluorescence (left) and quantification (right) showing increase percentage of S-phase cells in population versus control. n = 1000+ for each condition. (F) Overexpression of UTX results in decreased cell proliferation. Serial cell counts of fibroblasts transduced with UTX or puromycin-resistant vector alone. (G) Catalytic activity of UTX is required to induce cell cycle arrest, as measured by BrdU incorporation. (UTXmut) UTX^H1126A. n = 1000+ in each condition. (H) UTX overexpression does not induce apoptosis. Staurosporine treatment served as positive control. Mean ± SD is shown in all bar graphs.

Genetic interaction between UTX, RB, and RB-binding proteins. (A) Depletion of RB1 or HBP1 by siRNAs. (siluc) Control siRNA targeting luciferase. (B) RB or HBP1 depletion reverses UTX-mediated cell cycle arrest. Quantifications of percentage of cells in S phase as measured by BrdU incorporation is shown. n = 1000+. (C) UTX modulation of cell cycle progression in mouse embryonic fibroblasts (MEFs) requires the Rb family members. (Left) Wild-type MEFs. (Right) Rb, p107, and p130 triple knockout (TKO) MEFs. (D) C. elegans UTX-1 is synMuv. RNAi of utx-1 was performed on eight characteristic synMuv strains and N2 wild-type worms. Wild-type worms were exposed to utx-1 RNAi and showed a very weak Muv phenotype (<1% penetrance). Strains representing class A (lin-15A, lin-8, lin-38, and lin-56; genes labeled in red), class B (lin-35 Rb, lin-53 RbBP4, and lin-15B; genes labeled in blue), and class C (mys-1 TIP60; gene label in green) were used to test for UTX-1 synMuv activity. All of these strains present a normal vulva (Ferguson and Horvitz 1989), except for mys-1, which produces <10% Muv (Ceol and Horvitz 2004). Penetrance of the synMuv interactions vary between 8% and 24% for class A and 18% and 48% for class B, and is 57% for class C. Overall, utx-1 genetically interacts with classes A, B, and C, and behaves similarly to set-16 mll (K Fisher and GB Poulin, pers. comm.). The number of worms analyzed is indicated above each bar. (E) Photomicrograph of representative synMuv phenotype in an early L4 stage worm. Arrows indicate extra cells adopting the vulval fate. The normal vulva is underlined.