The H3K27me3 demethylase UTX in normal development and disease

Abstract

In 2007, the Ubiquitously Transcribed Tetratricopeptide Repeat on chromosome X (UTX) was identified as a histone demethylase that specifically targets di- and tri-methyl groups on lysine 27 of histone H3 (H3K27me2/3). Since then, UTX has been proven essential during normal development, as it is critically required for correct reprogramming, embryonic development and tissue-specific differentiation. UTX is a member of the MLL2 H3K4 methyltransferase complex and its catalytic activity has been linked to regulation of HOX and RB transcriptional networks. In addition, an H3K27me2/3 demethylase independent function for UTX was uncovered in promoting general chromatin remodeling in concert with the BRG1-containing SWI/SNF remodeling complex. Constitutional inactivation of UTX causes a specific hereditary disorder called the Kabuki syndrome, whereas somatic loss of UTX has been reported in a variety of human cancers. Here, we compile the breakthrough discoveries made from the first disclosure of UTX as a histone demethylase till the identification of disease-related UTX mutations and specific UTX inhibitors.

Keywords: :

• UTX
• H3K27
• MLL2
• HOX
• RB
• SWI/SNF
• reprogramming
• embryogenesis
• Kabuki
• cancer

Introduction

A decade ago, the identification of the first histone demethylase Lysine (K)-Specific Demethylase 1 (LSD1) served as a landmark discovery that triggered research on dynamic regulation of histone methylation.¹ In the following years, numerous additional histone demethylases, that execute the removal of methyl groups on specific lysine residues of the histone tails and of non-histone substrates, were characterized in more detail (reviewed in ref. 2). In 2007, several groups identified Ubiquitously Transcribed Tetratricopeptide Repeat on chromosome X (UTX) and Jumonji D3 (JMJD3) as novel histone demethylases that catalyze the removal of di- and trimethyl groups on histone H3 lysine 27, thereby promoting target gene activation.^3-6 Notably, Ubiquitously Transcribed Tetratricopeptide Repeat on chromosome Y (UTY) is a closely related homolog of UTX on the Y-chromosome but, up until now, no enzymatic H3K27me2/3 demethylase activity has been reported for UTY.^3,7

The X-linked H3K27me2/3 eraser UTX is a member of the MLL2 histone H3K4 methyltransferase complex^8,9 and contributes to animal body patterning by regulation of homeobox (HOX) genes.^4,5 In contrast, a histone demethylation-independent role for UTX and JMJD3 has been demonstrated in normal and malignant T-cells through interaction with the BRG1-containing SWI/SNF remodeling complex.¹⁰ UTX knockout (KO) studies have unraveled important roles for UTX in many developmental processes, including cardiac development and hematopoiesis, but also suggested that UTX and UTY might have redundant functions during embryonic development.^11-17

Upon the establishment of UTX as a histone eraser in the context of normal development, a number of studies started to report genetic defects targeting UTX as the underlying cause of specific diseases. In 2009, a role for the histone H3K27me2/3 demethylase UTX as tumor suppressor was initially postulated in several human tumors including multiple myeloma, esophageal and renal cancer.¹⁸ In 2012, specific loss-of-function defects in UTX were identified in patients with a specific hereditary disorder named the Kabuki syndrome.¹⁹ In this review, we summarize the current knowledge on UTX in normal development and highlight recent findings on its implication in cancer and hereditary disease.

UTX Drives Context Dependent Transcriptional Regulation Mainly Through its H3K27 Demethylase Activity

H3K27me2/3 demethylation

Methylation of H3K27 is a critical mediator of transcriptional gene repression and contributes to important biological processes including X-inactivation, genomic imprinting, stem cell maintenance, animal body patterning, circadian rhythms and cancer.^5,20 Regulation of cellular H3K27me3 levels is mainly mediated by the H3K27 methyltransferase Polycomb Repressor Complex 2 (PRC2) and the H3K27me2/3 demethylases UTX and JMJD3 (Fig. 1).^3-6,21-24 Two main classes of histone demethylases have been discovered until now including the flavin-dependent amine oxidases, such as LSD1,¹ and the iron and α-ketoglutarate-dependent dioxygenases with a Jumonji C (JmjC) catalytic domain, such as UTX and JMJD3.^3-7

Figure 1. The UTX family mediates H3K27me2/3 demethylation. Graphical illustration of open and closed chromatin states that are mediated by the histone demethylases UTX and JMJD3 and the histone methyltransferase PRC2 hereby erasing or writing methyl groups on H3K27 enabling activation or blockage of gene transcription, respectively (graphics from www.somersault1824.com).

The H3K27me2/3 demethylases UTX and JMJD3 preferentially demethylate H3K27me3 followed by H3K27me2 in vitro and in vivo. This demethylase activity is dependent on the catalytic JmjC domain, which contains conserved residues for binding with the co-factors iron and α-ketoglutarate.^3-7 Moreover, a newly identified zinc-binding domain within these H3K27me2/3 erasers provides specificity toward the histone lysine H3K27 and excludes interaction with the near-cognate histone lysine H3K9.^25,26 Notably, the family member UTY lacks H3K27me2/3 erasing activity in vitro and in vivo despite a conserved JmjC domain and 88% sequence homology with the UTX protein.^3,7,14 Finally, UTX and UTY proteins contain tetratricopeptide repeats (TPRs) at their N-terminal regions that are important for protein-protein interactions. These TPRs are lacking in the JMJD3 protein,^3-7 which might suggest lack of redundant functions between the H3K27me2/3 demethylases UTX and JMJD3.

UTX, UTY and JMJD3 are evolutionary conserved from Caenorhabditis elegans (C. elegans) to human.^3-7 The mouse genome contains the three H3K27me2/3 demethylase family members Utx, Uty and Jmjd3, whereas the genome of Drosophila melanogaster harbors only one ortholog of the mammalian H3K27me2/3 demethylases called dUTX.²⁷ The C. elegans genome possesses 4 orthologs whereby UTX-1 resembles the mammalian UTX and the three other orthologs are more related to JMJD3.²⁸

UTX is a member of the MLL2 H3K4 methyltransferase complex

In 2007, an MLL2 (also called ALR and formerly called MLL4) complex was identified in different mammalian cell types containing 12 protein members including MLL2, PTIP, UTX, ASC-2, ASH2L, RBBP5, WDR5, DPY30, matrin3, MGC4606, α- and β-tubulin.⁸ Furthermore, the 3 complex members MLL2, PTIP and UTX were shown to co-localize at promoter regions and first exons of MLL2 target genes marked with the transcriptional activation mark H3K4me3.⁸ Another group confirmed the identification of a similar H3K4 methyltransferase complex in which MLL2, MLL3, PTIP, UTX, ASC-2, ASH2L, RBBP5, WDR5, DPY30 and PA1 were also linked to H3K4 methylation activity (Fig. 2).⁹ Interestingly, the composition of the MLL2 complex was confirmed in the Drosophila ortholog Trithorax related (Trr) complex²⁹ and the C. elegans ortholog SET-16 complex.²⁸ Hence, this conserved MLL2 complex acts like a classical H3K4 methyltransferase complex,⁸ which suggests a dynamic interplay between H3K27me2/3 demethylation and H3K4 methylation during transcriptional gene activation.

Figure 2. The chromatin complexes MLL2-UTX, SWI/SNF and PRC2 contribute to open and closed chromatin conformations. Schematic representation of the H3K4 methyltransferase complex MLL2-UTX, the SWI/SNF ATPase remodeling complex and the H3K27 methyltransferase complex PRC2 composed out of different protein-coding and non-protein-coding members. The histone eraser UTX is part of the MLL2 complex leading to a dynamic interplay between H3K4 methylation and H3K27me2/3 demethylation. Furthermore, UTX can cooperate with the BRG1-containing SWI/SNF complex where it plays a role in general chromatin remodeling independent of its H3K27 demethylase function. The MLL2-UTX and SWI/SNF complexes both contribute to an open chromatin formation. The PRC2 complex enables efficient methylation of H3K27 thereby promoting gene silencing and chromatin compaction (graphics from www.somersault1824.com).

In flies, dUTX mutant tissues are marked by an increase in global H3K27me3 levels and surprisingly also a reduction in global H3K4me1 levels.^20,30 The effect on H3K4me1, a mark enriched at enhancer regions and at gene bodies of actively transcribed genes, seems to be independent of the demethylation capacity of dUTX.^20,30 Similarly, loss of the Trr H3K4 methyltransferase complex results in a global profound reduction of H3K4me1 levels in many tissues.^30,31 Notably, Trr and dUTX are enriched genome-wide near transcriptional start sites (TSSs) (enriched for H3K4me3 and RNAP II) but also on active enhancers (enriched for H3K4me1) indicating a promoter-proximal role as well as a promoter-distal role for the Trr-dUtx complex in transcriptional regulation.³⁰ Importantly, loss of Trr provoked the most profound reduction in H3K4me1 levels at the enhancer regions compared with dUTX loss³⁰ indicating that the Trr complex itself is the main driver of H3K4 monomethylation with support of dUTX.

UTX interacts with the SWI/SNF remodeling complex and the histone acetyltransferase CBP

Besides regulation of histone modifications like H3K27me3, ATPase-dependent remodeling complexes including the SWI/SNF family contribute to regulation of chromatin accessibility hereby enabling transcriptional activation and repression.³² The SWI/SNF family is mainly linked to transcriptional activation whereby the ATPase complex members BRG1 and BRM interact with acetylated histone tails through their bromodomain (Fig. 2).^33,34 An interaction between the Brg1-containing SWI/SNF remodeling complex and the histone H3K27me2/3 erasers Utx and Jmjd3 was demonstrated promoting general chromatin remodeling that was independent of their H3K27me2/3 demethylase potential. Furthermore, UTX and JMJD3 were shown to be functionally required for transcriptional activation of lineage-defining T-box transcription factors (including T-bet, Eomes and Tbx5),¹⁰ which represent essential transcriptional regulators in early cell-fate decisions, differentiation and organogenesis.³⁵

An interaction between the histone acetyltransferase Creb-binding protein (CBP) and the SWI/SNF remodeling complex is observed in mammals^10,36 and Drosophila.³⁷ In flies, CBP and the PRC2 complex act antagonistically in the regulation of active and repressed chromatin states at Polycomb target genes by acetylation and methylation of H3K27, respectively (Fig. 2).³⁷ H3K27 acetylation is mainly detected at enhancers, promoters and gene bodies of actively transcribed genes.³⁰ To enable efficient H3K27 acetylation, CBP was shown to associate with dUTX and the ATPase Brm of the SWI/SNF complex.³⁷ The ATPase Brm in flies resembles the 2 mammalian orthologs BRM and BRG1 of the SWI/SNF complexes.^32,37 Genome-wide binding studies showed that the chromatin components dUTX, Brm and CBP co-occurred at Polycomb target genes enriched for H3K27ac, suggesting a collaborative effort antagonizing Polycomb mediated gene silencing.³⁷ Of note, loss of dUTX or Trr results in a reduction of global H3K27ac levels in various tissues,^30,37 so dUTX might be capable of promoting H3K27 acetylation and H3K4 monomethylation in concert with Trr, Brm and CBP.

UTX regulates dynamic HOX expression

The HOX genes are highly conserved transcriptional regulators essential during development of the anterior-posterior axis (head-tail).³⁸ The HOX family in humans consists of 39 HOX genes that encode 4 clusters of HOX proteins, i.e., HOXA to HOXD.³⁹ Aberrant expression of HOX genes leads to developmental defects and disease.³⁹ Polycomb and Trithorax proteins are implicated in the complex regulation of spatio-temporal expression of HOX genes thereby regulating animal body patterning and cell differentiation.⁴⁰ Evidence for direct involvement of UTX in regulation of HOX gene activity was demonstrated through UTX knockdown experiments in HEK293T cells in which loss of UTX induced transcriptional repression of HOXA and HOXC clusters.⁵ Furthermore, UTX loss also resulted in an increase of H3K27me2/3 levels at the promoter regions of HOXA13 and HOXC4, whereas no changes in MLL2 and PRC2 occupancy were observed.⁵

Chromatin immunoprecipitation (ChIP) of UTX followed by HOX-specific ultra-dense tiling microarrays (ChIP-chip) showed a strong UTX binding to narrow windows within 500 base pairs downstream of the TSSs of many HOX genes in human fibroblasts, but not in mouse embryonic stem cells (ESCs) where HOX genes are silenced. Importantly, H3K27 trimethylation was concomitantly diminished at UTX binding sites.³ Hence, HOX expression is tightly regulated by opposing activities of the H3K27me3 writer PRC2 and the H3K27me2/3 eraser UTX in order to enable context dependent transcriptional regulation.

UTX regulates a retinoblastoma gene network

Genome-wide ChIP-chip analysis of UTX, H3K4me2 and H3K27me3 in human fibroblasts identified UTX binding at 1945 promoters mostly upstream of TSS.⁴¹ Out of these, the majority of UTX targets were enriched for H3K4me2 (62%) and showed transcriptional activity. This observation is consistent with UTX being a member of the H3K4 methyltransferase MLL2 complex.^8,9,41 Using Ingenuity Pathway analysis (IPA), a RB gene network including RB1, HBP1 and RBBP4, 5, 6 and 9 was discovered as the most significant network of UTX-bound genes enriched for H3K4me2.⁴¹ Loss of UTX in human fibroblasts and mouse embryonic fibroblasts (MEFs) confirmed increased levels of H3K27me3 at promoters of genes from the identified RB-network thereby contributing to deregulated RB-dependent cell cycle arrest leading to ectopic cell proliferation.^41,42

UTX Escapes X-Inactivation

The female genome harbors 2 X-chromosomes of which one X-chromosome is silenced to compensate the difference in gene dosage with males.⁴³ X-inactivation is mediated by the long intergenic non-coding RNA (lincRNA) XIST, which coats the inactive X-chromosome. Next, the PRC2 complex is guided by XIST to the inactive X-chromosome hereby enabling H3K27 methylation leading to the formation of facultative heterochromatin and gene silencing (Fig. 2).⁴⁴

Utx is one of the few genes that can escape X-inactivation in female mice and human.⁴⁵ In agreement, Utx shows higher expression in female tissue of brain, liver, neurons and sexual organs.^46,47 In line with this observation, a female-specific function was reported for Utx in the regulation of the X-linked homeobox genes Rhox6 and Rhox9.⁴⁶ It was shown that Utx is expressed at higher levels in undifferentiated female ESCs as compared with male ESCs, which resulted in a stronger Utx binding in the promoter regions of Rhox6 and Rhox9 in females. Subsequently, Utx mediates a more profound H3K27me3 removal and transcriptional activation of these homeobox genes in female cells.⁴⁶

The Role of UTX in Cellular Reprogramming

Pluripotency is the unique characteristic of ESCs and the early developmental stage embryo to induce the generation of the three germ layers endoderm, ectoderm and mesoderm finally constituting all tissues of the adult organism.⁴⁸ Through in vitro reprogramming of somatic cells via induction of the four transcription factors OCT4, SOX2, KLF4 and MYC; pluripotency can be re-established hereby generating induced pluripotent stem cells (iPSC).^48,49 During reprogramming, genome-wide changes in the transcriptome and the chromatin structure are induced to achieve the switch to the pluripotent state.⁴⁸

The histone H3K27me3 demethylase Utx has a critical role in efficient induction or re-establishment of pluripotency during the generation of iPSCs in vitro, but is dispensable for maintenance of pluripotency.⁵⁰ The requirement for UTX to ensure correct reprogramming depends on its H3K27me3 demethylase activity.⁵⁰ During iPSC induction, Utx target genes that are enriched with H3K4me3 in Utx wild-type MEFs aberrantly accumulate H3K27me3 in Utx mutant MEFs. These loci, including the stem-cell maintenance genes Sall1, Sall4, and Utf1 as well as validated Klf4 and Oct4 target genes, fail to reactivate during reprogramming due to loss of Utx.⁵⁰ Hence, Utx seems to play a crucial role during reprogramming through its H3K27me2/3 demethylation activity in correct re-activation of essential pluripotency genes.

A Critical Role for UTX in Embryonic Stem Cells and Embryonic Development

Embryonic stem cells

H3K27 methylation is important in the maintenance of self-renewing ESCs by repressing tissue-specific developmental genes.^51-53 During ESC differentiation, the re-activation of developmental genes is associated with loss of H3K27me3.⁵¹ Loss of Utx does not influence ESC self-renewal and proliferation, but seems to provoke an effect on the differentiation capacity of ESCs.^11,12,50,54 This differentiation defect is reflected by loss of induction of a set of developmental genes including some ectoderm markers (Otx2, Msi1, Sox1, and Pax6),⁵⁴ endoderm markers (Gata4, Gata6, Foxa2, Sox17, Gsc, and Ncad)^12,55 and mesoderm markers (Vegfr2, Wnt3, and Brachyury),^12,13,54,55 presumably due to loss of Utx binding to the promoter regions of these genes.⁵⁴

Contradicting studies report on whether the effect of UTX loss on ESC differentiation depends on its H3K27me3 demethylase activity.^12,13,54,55 Namely, male UTX KO ESCs that express a catalytic inactive or wild-type Utx protein both could rescue the induction of expression of the mesoderm genes Wnt3 and Brachyury supporting an H3K27me3-demethylase-independent role of Utx during ESC differentiation.⁵⁴ Furthermore, Utx binds directly to the promoters of the ectoderm genes Msi1 and Sox1 and the mesoderm gene Brachyury marked with an increase in H3K4me3 levels but without a change in H3K27me3 levels.⁵⁴ In contrast, a decrease in H3K27me3 levels and an increase in H3K4me3 levels was seen at the promoter of the Utx target Hoxb1 during differentiation of Utx wild-type ESCs confirming that Utx still plays an H3K27me3 demethylase-dependent role in transcriptional regulation of homeotic and developmental genes.⁵⁴ Therefore, UTX regulates transcriptional activation of UTX target genes during ESC differentiation in different ways dependent or independent of its H3K27me2/3 demethylation activity.

Embryonic development in mice

Several groups have shown that complete loss of Utx in Utx^Δ/Δ female mice leads to embryonic lethality between embryonic day (E) 10.5 and E12.5, whereas Utx^Δ/+ heterozygous female mice are viable and fertile.^11-15 Surprisingly, Utx^Δ/Y male mice are marked by a diverse outcome ranging from embryonic lethality to tumor formation in adults.^11-15 The viable adult Utx^Δ/Y male mice are smaller in size compared with their littermates^12,14 and remain fertile indicating that Utx is not required for male fertility.¹² Interestingly, Utx^Δ/UtyΔ male mice, in which both Utx and Uty are eliminated, phenocopy the Utx^Δ/Δ female mice,¹⁴ suggesting that the Y-linked Uty gene can partially rescue the effect of complete loss of Utx.¹¹ In addition, Utx and Uty can both interact with Rbbp5 (member of the Mll2 complex) regulating H3K4 methylation rather than H3K27me2/3 demethylation at Utx and Uty target genes.¹⁴

Embryonic development in Drosophila melanogaster

Homozygous dUTX Drosophila mutants are lethal, with only 5% of the animals that survive the pupal stage although not reaching adulthood.²⁰ More specifically, dUTX^Δ homozygotes that lack both maternal and zygotic dUTX protein (dUTX^Δmat-zyg-) die as larvae, whereas dUTX^Δ homozygotes that lack only zygotic dUTX protein (dUTX^Δmat+zyg-) can develop further but die quickly after the pupal stage.⁵⁶ dUTX was shown to be necessary for the controlled expression of 2 specific homeotic genes Ubx and Abd-B in very early fly development.⁵⁶ Furthermore, sex combs malformation, rough eyes, wrinkled wings and wing vein defects are observed in the dUTX mutants which resembles some characteristics of Trithorax mutants and is in concordance with the effect of these proteins on Hox gene regulation.^20,56

Embryonic development in Caenorhabditis elegans

In agreement with Drosophila, UTX-1 mutant worms lacking both maternal and zygotic UTX-1 protein (UTX-1^m-z-) die as late stage embryos or malformed L1 larvae. Mutants with loss of only the zygotic UTX-1 (UTX-1^m+z-) protein are viable and reach adulthood, but a reduction in fertility is observed due to defects in gonad migration and oocyte organization. The UTX-1^m-z- mutants show an increase in global H3K27me2/3 levels at the embryonic stage, but introduction of a catalytic inactive mutant of UTX-1 in UTX-1^m+z- mutant animals rescued the fertility indicating that the UTX-1 demethylase activity is not necessary for this developmental process. Interestingly, loss of UTX-1 complex members SET-16 (MLL2–3), PIS-1 (PTIP), WDR-5.1 (WDR5) and F21H12.1 (RBBP5) also resulted in posterior and gonadal defects further confirming that these genes are acting in the same genetic complex.²⁸

A Critical Role for UTX in Tissue-Specific and Developmental Processes

UTX contributes to a variety of tissue-specific and developmental processes including cardiac development,^11-15 hematopoiesis,^13,15,57,58 myogenesis,^16,17,27,59 osteogenic differentiation,⁶⁰ wound healing,⁶¹ and aging.^62,63 Cardiac development and hematopoiesis will be discussed in further detail.

Cardiac development

The histone eraser Utx seems to play an important role in cardiac development. Utx^Δ/Y ESCs that in vitro differentiate into mature cardiac cells show absence of effective heart-like rhythmic contractions due to a failure of inducing cardiac-specific gene expression.¹¹ Furthermore, Utx^Δ/Δ female mice and Utx^Δ/UtyΔ male mice exhibited strong cardiac developmental and neural tube closure defects (cranioschisis) at day E9-E10.5.^11-15 This cardiac malformation is presumably caused by 2 mechanisms. First, loss of Utx disables the induction of transcriptional activation of cardiac-specific genes mediated by H3K27me2/3 demethylation, whereby the core cardiac transcription factors Nkx2-5, Tbx5, Gata4 and Srf specifically guide Utx to the promoter regions of cardiac-specific genes (Fig. 3).¹¹ Second, loss of Utx disturbs the interaction between the Brg1-containing SWI/SNF chromatin remodeling complex and the core cardiac transcription factor Tbx5, an effect that is thought to be independent from the H3K27me2/3 demethylase activity of UTX (Fig. 3).^10,11 This dual role of UTX is further supported by an independent study showing that UTX and UTY can physically interact with BRG1, NKX2-5, SRF and TBX5 hereby inducing expression of cardiac-specific genes.¹⁴ Hence UTX enables transcriptional activation of cardiac-specific genes through H3K27me3 demethylase dependent and independent mechanisms during cardiac development.

Figure 3. The involvement of UTX in genetic regulation of cardiac development. During cardiac development UTX, guided by the core cardiac transcription factors NKX2–5, TBX5, GATA4 and SRF, promotes specific gene activation of cardiac-specific genes through demethylation of H3K27me2/3 at their promoter regions. In addition, UTX and UTY can interact with the BRG1-containing SWI/SNF chromatin remodeling complex enabling general chromatin remodeling at these cardiac-specific gene loci (graphics from www.somersault1824.com).

Hematopoiesis

Utx expression levels peak in hematopoietic stem and progenitor cells and drop during hematopoietic differentiation.⁵⁸ In Utx^Δ/Δ female mice embryos (E10.5), a lack of red bloods cells (anemia) was observed indicating a critical role for Utx in hematopoiesis.¹³ Furthermore, female adult mice in which Utx was homozygously eliminated at 11–14 wk of age showed an enlarged spleen, reduced hemoglobin levels, anemia, thrombocytopenia and mild leukocytopenia compared with normal controls. In addition, hematopoietic transcription factors Tal1, Gata1, and Lyl1 were downregulated in the bone marrow of Utx KO female mice.¹⁵ Furthermore, Utx was identified as an essential regulator of stem cell migration since loss of Utx blocked the migration potential of hematopoietic progenitor cells in vitro and of primordial germ cells in vivo.¹⁵

Constitutional UTX Defects Cause the Kabuki Syndrome

Constitutional loss-of-function defects in UTX cause the so-called Kabuki syndrome. This rare congenital anomaly syndrome, also called the Kabuki make-up syndrome, was first described by Niikawa and Kuroki in 1981.^64,65 The Kabuki syndrome is characterized by moderate-to-severe mental retardation, visceral and skeletal abnormalities, postnatal growth impairment (short stature) and facial abnormalities including large protruding ears.^64-66 The Kabuki syndrome has a prevalence of about 1 in 32 000 live births.⁶⁵ Using whole-exome sequencing, nonsense and frameshift mutations in the MLL2 gene were first identified as a cause of the Kabuki syndrome⁶⁷ in 74% of these patients.⁶⁸ Two years later, 3 focal deletions targeting the histone demethylase UTX were identified in 1 male and 2 female Kabuki patients.⁶⁹ Subsequently, 2 nonsense mutations and a small indel mutation were discovered in 1 female and 2 male Kabuki patients.¹⁹

As discussed in the next part, both the MLL2 and the UTX gene have been reported as important histone modifier genes involved in cancer pathogenesis. Until now, 6 Kabuki patients have been described that developed different types of cancer including pre-B-ALL,⁷⁰ hepatoblastoma,⁷¹ neuroblastoma,^71,72 Burkitt lymphoma,⁷³ and fibromyxoid sarcoma⁷⁴ marking the Kabuki syndrome as a cancer predisposition syndrome.

UTX as a Bona Fide Tumor Suppressor Gene in Cancer Biology

Over the last years, deregulated histone methylation became a major theme in cancer biology research including the balance of H3K27 methylation. In 2009, somatic loss-of-function mutations and deletions targeting the UTX gene were identified in multiple cancer types including multiple myeloma, esophageal and renal cancer.¹⁸ This notion was further supported by subsequent identification of recurrent inactivating UTX mutations and deletions in several leukemia as well as solid tumor types (Fig. 4; Fig. S1; Tables S1 and 2), mainly through exome- or genome-wide sequencing strategies.^75-84 Furthermore, a recent study of The Cancer Genome Atlas (TCGA) in which whole-exome sequencing was performed on 3281 tumors derived from 12 tumor types identified 127 significantly mutated genes including the tumor suppressor gene UTX.⁸⁵

Figure 4.UTX defects in primary cancer samples. Schematic representation of the presence of UTX defects (%) in primary cancer samples of a broad range of leukemic and solid cancer types investigated in multiple genomic studies.

Bladder cancer has one of the highest frequencies of UTX mutations (Fig. 4) typically consisting of truncating mutations in the region coding for the functional JmjC domain of UTX. Interestingly, more than 50% of bladder cancer patients harbor genetic aberrations in a bigger set of chromatin remodeling genes including UTX, CREBBP, EP300, ARID1A, CHD6, MLL1, MLL3, and NCOR1.⁷⁷ Furthermore, multiple myeloma cases harboring mutations in UTX or in the Trithorax complex members MLL1, MLL2, and MLL3 were linked with high expression of HOXA9⁷⁶ confirming the role of UTX in HOX gene regulation.³

Interestingly, a study in chronic myelomonocytic leukemia (CMML) identified loss-of-function mutations in both UTX and EZH2 despite their opposing roles on H3K27me3 regulation. Notably, these mutations occurred in a mutually exclusive manner in CMML patient samples.⁸⁶ UTX and EZH2 defects in CMML were associated with ASXL1⁸⁷ and TET2⁸⁸ loss-of-function mutations, two genes also implicated in epigenetic regulation of gene expression.⁸⁶ These studies indicate that the chromatin regulatory machinery is a recurrent mutational target in a broad range of cancer types highlighting the importance of correct regulation of chromatin remodeling in the homeostasis of normal tissues.³²

Whole-exome and whole-genome next-generation sequencing studies in medulloblastoma identified mutations in different chromatin regulators including UTX, MLL2, MLL3, SMARCA4 (BRG1), ZMYM3, and CREBBP.^78-80 In general, four prognostically relevant subgroups (SHH-subgroup, WNT-subgroup, subgroup 3 and subgroup 4) are defined in this malignant childhood brain tumor.⁸⁰ Of particular interest, medulloblastoma is marked by a disturbed gender distribution toward males in subgroups 3 and 4⁸⁹ and mutations targeting the X-linked UTX gene were exclusively identified in these 2 subgroups.^79,80,90 In line with the fact that UTX escapes X-inactivation,⁴⁵ female medulloblastoma patients harboring UTX defects showed bi-allelic UTX inactivation (splice-site UTX mutation and deletion⁸⁰ or missense mutation and loss of chromosome X⁷⁹), suggesting that complete loss of UTX is required for malignant transformation. In addition, more than 50% of UTX mutant male medulloblastoma patients had a deletion of the Y-chromosome (including the UTX family member UTY) compared with less than 10% of UTX wild-type males.⁸⁰ Furthermore, gain or overexpression of the H3K27 methyltransferase EZH2 as well as inactivating mutations in H3K4me3 regulators CHD7 and ZMYM3 were also present in subgroups 3 and 4 of medulloblastoma. All together, these data indicate that deregulation of H3K27 and H3K4 methylation is a core oncogenic component of medulloblastoma. Importantly, mutations targeting the X-linked gene UTX might partially explain the higher male prevalence in medulloblastoma subgroups 3 and 4.⁸⁰

So far, the functional consequences of cancer related UTX mutations have been poorly characterized. Western blot analyses show lack of UTX protein expression in UTX mutant cancer cell lines.¹⁸ Furthermore, re-expression of UTX in UTX mutant cancer cell lines was associated with lower H3K27me3 levels at the promoters of Polycomb target genes and concomitant inhibition of cell proliferation.¹⁸ In adenoid cystic carcinomas, the impact of the identified UTX missense mutations was evaluated by introduction of these UTX mutants in HEK293T cells, which led to an increase in cell growth and a decrease in H3K27me3 levels.⁸⁴ Finally, a Sleeping Beauty based insertional mutagenesis screen using a Kras driven pancreatic mouse model identified 543 candidate cancer driver genes with 10% of the genes involved in chromatin regulation including the Utx gene. This un-biased screening approach provides an independent in vivo confirmation of the potential role of Utx as a bona fide tumor suppressor gene in cancer biology.⁹¹

Targeting UTX by Small Molecule Inhibitors

The most selective compound against JMJD3 and UTX, GSK-J1, was discovered based on the structural insights of the human and mouse JMJD3 protein with an IC₅₀ of 60nM for inhibiting JMJD3. Furthermore, GSK-J1 is active against both H3K27me3 erasers UTX and JMJD3 but inactive against a panel of demethylases of the JmjC family. Notably, the cell-penetrating derivative GSK-J4 was able to inhibit the JMJD3-induced loss of total nuclear H3K27me3 levels and to enable specific inhibition of H3K27 demethylation at promoter regions of JMJD3 and UTX target genes.⁹²

Several general inhibitors of JmjC demethylases have been identified including the α-ketoglutaric acid mimics N-oxalylglycine,^93,94 methylstat⁹⁵ and 2,4-dicarboxypyridine; the iron-chelating agent deferoxamine⁹⁵; the pyridine hydrazone JIB-04⁹⁶ and catechols.⁹⁷ Unfortunately, these inhibitors are selective against some or all JmjC enzymes but are unable to target one specific JmjC member.

Conclusions

In this review, we have summarized and discussed the identification of the histone H3K27me2/3 erasers UTX and JMJD3, and the role of UTX in normal development and disease. UTX cooperates with a set of chromatin players, including the histone methyltransferase complex MLL2, the histone acetyltransferase CBP and the BRG1-containing SWI/SNF remodeling complex,^8-10,37 to trigger active gene transcription by H3K27me2/3 demethylation or H3K4 methylation.^{3-6,8-10,30,37} Up until now, the main direct target genes of UTX encompass a broad set of HOX genes important in animal body patterning^3,5 and a RB gene network implicated in cell fate control.⁴¹ Furthermore, UTX loss can lead to the development of the Kabuki syndrome⁶⁹ and can contribute to cancer pathogenesis.¹⁸

The third UTX family member UTY has no clear enzymatic demethylase activity despite the presence of a conserved JmjC domain and high amino acid conservation.^3,7,14 Surprisingly, UTY can partly compensate UTX functions including the cooperation with the MLL2 and BRG1-containing SWI/SNF complexes.¹⁴ Also, UTY can partially rescue embryonic lethality in UTX KO male mice.^11-15 Furthermore, UTX mutations are frequently found together with deletions encompassing the UTY gene locus in male cancer patients.^18,80 Hence, a set of overlapping functions between UTX and UTY in transcriptional regulation seems present despite the lack of H3K27me2/3 demethylase activity for UTY.

All together, the complexity of biological functions assigned to UTX is just starting to emerge and will trigger additional research in a broad range of developmental processes. Multiple studies have highlighted UTX as an important player in cancer biology, but its role as a bona fide tumor suppressor still needs further confirmation using genetically engineered in vivo models. Ultimately, these biological insights will clarify to what extend loss of UTX provides therapeutic opportunities for human disease.

Abbreviations:
CBP	=	CREB-binding protein
ChIP	=	chromatin immunoprecipitation
CMML	=	chronic myelomonocytic leukemia
E	=	embryonic day
ESC	=	embryonic stem cell
H3K27me2/3	=	histone 3 lysine 27 di- and tri-methylation
HOX	=	homeobox
IPA	=	Ingenuity Pathway Analysis
iPSC	=	induced pluripotent stem cell
JmjC	=	Jumonji C
JMJD3	=	Jumonji D3
KO	=	knockout
lincRNA	=	long intergenic non-coding RNA
LSD1	=	Lysine-Specific Demethylase 1
MEF	=	mouse embryonic fibroblast
PRC2	=	Polycomb Repressor Complex 2
TCGA	=	The Cancer Genome Atlas
TPR	=	tetratricopeptide repeat
Trr	=	Trithorax related
TSS	=	transcriptional start site
UTX	=	ubiquitously transcribed tetratricopeptide repeat on chromosome X
UTY	=	ubiquitously transcribed tetratricopeptide repeat on chromosome Y

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

This work is supported by the Fund for Scientific Research (FWO) Flanders (postdoctoral grant to P.V.V., PhD grant to J.V.d.M., Odysseus grant to P.V.V, project grants G.0198.08, G.0564.13N, G.0550.13N, G.0869.10N to F.S. and G.A001.13N to P.V.V.); the Flemish Liga against Cancer (VLK) (PhD grant to J.V.d.M.); the GOA-UGent (grant no. 12051203 to F.S.); the Cancer Plan from the Federal Public Service of Health (F.S.); the Children Cancer Fund Ghent (F.S.); the Belgian Program of Interuniversity Poles of Attraction IUAP (F.S.) and the Belgian Foundation Against Cancer (365O9110 to F.S.).

10.4161/epi.28298