Enzyme kinetic studies of histone demethylases KDM4C and KDM6A: Towards understanding selectivity of inhibitors targeting oncogenic histone demethylases

Catalysis by KDM6 Histone Demethylases - A Synergy between the Non-Heme Iron(II) Center, Second Coordination Sphere, and Long-Range Interactions

KDM6A (UTX) and KDM6B (JMJD3) are human non-heme Fe(II) and 2-oxoglutarate (2OG) dependent JmjC oxygenases that catalyze the demethylation of trimethylated lysine 27 in the N-terminal tail of histone H3, a post-translational modification that regulates transcription. A Combined Quantum Mechanics/ Molecular Mechanics (QM/MM) and Molecular Dynamics (MD) study on the catalytic mechanism of KDM6A/B reveals that the transition state for the rate-limiting hydrogen atom transfer (HAT) reaction in KDM6A catalysis is stabilized by polar (Asn217) and aromatic (Trp369)/non-polar (Pro274) residues in contrast to KDM4, KDM6B and KDM7 demethylases where charged residues (Glu, Arg, Asp) are involved. KDM6A employs both σ- and π-electron transfer pathways for HAT, whereas KDM6B employs the σ-electron pathway. Differences in hydrogen bonding of the Fe-chelating Glu252(KDM6B) contribute to the lower energy barriers in KDM6B vs. KDM6A. The study reveals a dependence of the activation barrier of the rebound hydroxylation on the Fe-O-C angle in the transition state of KDM6A. Anti-correlation of the Zn-binding domain with the active site residues is a key factor distinguishing KDM6A/B from KDM7/4s. The results reveal the importance of communication between the Fe center, second coordination sphere, and long-range interactions in catalysis by KDMs and, by implication, other 2OG oxygenases.

Tunable, division-independent control of gene activation timing by a polycomb switch

During development, progenitors often differentiate many cell generations after receiving signals. These delays must be robust yet tunable for precise population size control. Polycomb repressive mechanisms, involving histone H3 lysine-27 trimethylation (H3K27me3), restrain the expression of lineage-specifying genes in progenitors and may delay their activation and ensuing differentiation. Here, we elucidate an epigenetic switch controlling the T cell commitment gene Bcl11b that holds its locus in a heritable inactive state for multiple cell generations before activation. Integrating experiments and modeling, we identify a mechanism where H3K27me3 levels at Bcl11b, regulated by methyltransferase and demethylase activities, set the time delay at which the locus switches from a compacted, silent state to an extended, active state. This activation delay robustly spans many cell generations, is tunable by chromatin modifiers and transcription factors, and is independent of cell division. With their regulatory flexibility, such timed epigenetic switches may broadly control timing in development.

KDM5B is a master regulator of the H3K4-methylome in stem cells, development and cancer

Epigenetic regulation of chromatin plays a critical role in controlling stem cell function and tumorigenesis. The histone lysine demethylase, KDM5B, which catalyzes the demethylation of histone 3 lysine 4 (H3K4), is important for embryonic stem (ES) cell differentiation, and is a critical regulator of the H3K4-methylome during early mouse embryonic pre-implantation stage development. KDM5B is also overexpressed, amplified, or mutated in many cancer types. In cancer cells, KDM5B regulates expression of oncogenes and tumor suppressors by modulating H3K4 methylation levels. In this review, we examine how KDM5B regulates gene expression and cellular fates of stem cells and cancer cells by temporally and spatially controlling H3K4 methylation levels.

Peptides Derived from Histone 3 and Modified at Position 18 Inhibit Histone Demethylase KDM6 Enzymes

The KDM6 subfamily of histone lysine demethylases has recently been implicated as a putative target in the treatment of a number of diseases; this makes the availability of potent and selective inhibitors important. Due to high sequence similarity of the catalytic domain of Jumonji C histone demethylases, the development of small-molecule, family-specific inhibitors has, however, proven challenging. One approach to achieve the selective inhibition of these enzymes is the use of peptides derived from the substrate, the histone 3 C terminus. Here we used computational methods to optimize such inhibitors of the KDM6 family. Through natural amino acid substitution, it is shown that a K18I variant of a histone H3 derived peptide significantly increases affinity towards the KDM6 enzymes. The crystal structure of KDM6B in complex with a histone 3 derived K18I peptide reveals a tighter fit of the isoleucine side chain, compared with that of the arginine. As a consequence, the peptide R17 residue also has increased hydrophilic interactions. These interactions of the optimized peptide are likely to be responsible for the increased affinity to the KDM6 enzymes.

Lysine demethylase inhibition protects pancreatic β cells from apoptosis and improves β-cell function

Transcriptional changes control β-cell survival in response to inflammatory stress. Posttranslational modifications of histone and non-histone transcriptional regulators activate or repress gene transcription, but the link to cell-fate signaling is unclear. Inhibition of lysine deacetylases (KDACs) protects β cells from cytokine-induced apoptosis and reduces type 1 diabetes incidence in animals. We hypothesized that also lysine demethylases (KDMs) regulate β-cell fate in response to inflammatory stress. Expression of the demethylase Kdm6B was upregulated by proinflammatory cytokines suggesting a possible role in inflammation-induced β-cell destruction. Inhibition of KDM6 demethylases using the selective inhibitor GSK-J4 protected insulin-producing cells and human and mouse islets from cytokine-induced apoptosis by blunting nuclear factor (NF)-κB signaling and endoplasmic reticulum (ER) stress response gene expression. GSK-J4 furthermore increased expression of insulin gene and glucose-stimulated insulin secretion. Expression of genes regulating purinergic and cytokine ligand-receptor interactions was downregulated following GSK-J4 exposure, while expression of genes involved in cell maintenance and survival was upregulated. These data suggest that KDMs are important regulators of inflammation-induced β-cell dysfunction and death.

Structural Basis of Histone Demethylase KDM6B Histone 3 Lysine 27 Specificity

KDM subfamily 6 enzymes KDM6A and KDM6B specifically catalyze demethylation of di- and trimethylated lysine on histone 3 lysine 27 (H3K27me3/2) and play an important role in repression of developmental genes. Despite identical amino acid sequence in the immediate surroundings of H3K9me3/2 (ARKS), the enzymes do not catalyze demethylation of this general marker of repression. To address this question for KDM6B, we used computational methods to identify H3(17-33)-derived peptides with improved binding affinity that would allow co-crystallization with the catalytic core of human KDM6B (ccKDM6B). A total of five peptides were identified, and their IC₅₀ values were determined in a matrix-assisted laser desorption ionization time-of-flight-based assay. Despite none of the peptides showing affinity significantly higher than that of the H3(17-33) peptide, it was possible to co-crystallize ccKDM6B with a H3(17-33)A21M peptide. This structure reveals the interactions between the KDM6B zinc binding domain and the H3(17-23) region. Although KDM6A and KDM6B differ in primary sequence, particularly in the H3L20 binding pocket of the zinc binding domains, where two histidines in KDM6A have been replaced by a glutamate and a tyrosine, they bind H3(17-23) in a very similar fashion. This structure shows that KDM6B, in analogy with KDM6A, also uses the zinc binding domain to achieve H3K27me3/me2 specificity. The histidine to glutamine substitution at amino acid position 1564 in the KDM6B zinc binding domain can further explain why KDM6B binds substrates with an affinity higher than that of KDM6A.

Structure-Based Design of a New Scaffold for Cell-Penetrating Peptidic Inhibitors of the Histone Demethylase PHF8

The histone demethylase PHF8 catalyzes demethylation of mono- and di-methylated Lys9 on histone H3 (H3K9me1/2), and is a transcriptional activator involved in the development and cancer. Affinity and specificity of PHF8 towards H3K9me2 is affected by interaction with both the catalytic domain and a PHD reader domain. The latter specifically recognizes tri-methylated Ly4 on histone H3. A fragment of the histone H3 tail with tri-methylated Lys4 was used as a template for the structure-based design of a cyclic, cell-penetrating peptide that exhibits micromolar binding affinity to PHF8 in biochemical assays. The inhibitor has significantly lower affinity towards KDM2 enzymes (the phylogenetically closest subfamily), and to KDM3 and KDM6 subfamilies. Selectivity is only marginal towards an enzyme from the KDM4 family, which shares histone tail specificity with PHF8. It is a substrate of KDM5B, thus implying that the free N terminus is not part of the KDM5 enzyme substrate recognition machinery. The cyclic peptide's ability to penetrate cells is achieved by incorporation of a sequence derived from HIV Tat. The derived cyclic peptide can be used as a starting compound in the search for potent and selective PHF8 inhibitors.

Analyses of publicly available genomics resources define FGF-2-expressing bladder carcinomas as EMT-prone, proliferative tumors with low mutation rates and high expression of CTLA-4, PD-1 and PD-L1

Fibroblast growth factor 2 (FGF-2) is overexpressed in a subset of invasive bladder carcinomas and its overexpression correlates with poor prognosis. Analyses of publicly available databases addressing the molecular mechanisms that may be responsible for the poor prognosis of these tumors, revealed that FGF-2 expression correlates positively with the expression of epithelial to mesenchymal transition (EMT)-promoting transcription factors and with changes in gene expression that are characteristic of EMT. The same analyses also revealed that FGF-2 correlates negatively with the expression, mutation and copy number variations of FGFR-3, all of which are associated with noninvasive bladder carcinomas. Finally, they showed that FGF-2 expression correlates with the expression of FGFR-1, the expression of the IIIc variant of FGFR-2 and with the expression of Akt3. The latter observation is significant because our earlier studies had shown that Akt3 regulates FGFR-2 alternative splicing, shifting the balance toward the IIIc relative to the IIIb FGFR-2 splice variant. As the IIIc variant is recognized by FGF-2, while the IIIb variant is not, we conclude that Akt3 may facilitate the FGF-2 response. FGF-2 is known to promote the expression of KDM2B, which functions in concert with EZH2 to repress the EZH2-targeting microRNA miR-101, activating a switch, which stably upregulates EZH2. The cancer genome atlas (TCGA) data showing a correlation between KDM2B and EZH2 expression and Oncomine data, showing a correlation between KDM2B and tumor progression, strongly support the role of the FGF-2/KDM2B/miR-101/EZH2 pathway in bladder cancer. These observations combined, suggest a model according to which FGF-2 induces EMT, cell proliferation and cancer stem cell self-renewal by coupling the Akt3 and KDM2B-controlled pathways outlined above, in bladder carcinomas. Further analyses of publicly available databases, revealed that FGF-2-expressing bladder carcinomas carry fewer genetic alterations and they tend to express high levels of CTLA-4, PD-1 and PD-L1, which suggests immune blockade by checkpoint activation. EMT, enhanced proliferation and immune checkpoint activation combined, may be responsible for the poor prognosis of FGF-2-expressing bladder carcinomas.

Human UTY(KDM6C) is a male-specific Nϵ-methyl lysyl demethylase

The Jumonji C lysine demethylases (KDMs) are 2-oxoglutarate- and Fe(II)-dependent oxygenases. KDM6A (UTX) and KDM6B (JMJD3) are KDM6 subfamily members that catalyze demethylation of N(ϵ)-methylated histone 3 lysine 27 (H3K27), a mark important for transcriptional repression. Despite reports stating that UTY(KDM6C) is inactive as a KDM, we demonstrate by biochemical studies, employing MS and NMR, that UTY(KDM6C) is an active KDM. Crystallographic analyses reveal that the UTY(KDM6C) active site is highly conserved with those of KDM6B and KDM6A. UTY(KDM6C) catalyzes demethylation of H3K27 peptides in vitro, analogously to KDM6B and KDM6A, but with reduced activity, due to point substitutions involved in substrate binding. The results expand the set of human KDMs and will be of use in developing selective KDM inhibitors.

The H3K27me3 demethylase UTX in normal development and disease

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.

Studies of H3K4me3 demethylation by KDM5B/Jarid1B/PLU1 reveals strong substrate recognition in vitro and identifies 2,4-pyridine-dicarboxylic acid as an in vitro and in cell inhibitor

Dynamic methylations and demethylations of histone lysine residues are important for gene regulation and are facilitated by histone methyltransferases and histone demethylases (HDMs). KDM5B/Jarid1B/PLU1 is an H3K4me3/me2-specific lysine demethylase belonging to the JmjC domain-containing family of histone demethylases (JHDMs). Several studies have linked KDM5B to breast, prostate and skin cancer, highlighting its potential as a drug target. However, most inhibitor studies have focused on other JHDMs, and inhibitors for KDM5B remain to be explored. Here, we report the expression, purification and characterization of the catalytic core of recombinant KDM5B (ccKDM5B, residues 1-769). We show that ccKDM5B, recombinantly expressed in insect cells, demethylates H3K4me3 and H3K4me2 in vitro. The kinetic characterization showed that ccKDM5B has an apparent Michaelis constant (K(m) (app) ) value of 0.5 μm for its trimethylated substrate H3(1-15)K4me3, a considerably increased apparent substrate affinity than reported for related HDMs. Despite the presence of a PHD domain, the catalytic activity was not affected by additional methylation at the H3K9 position, suggesting that in vitro chromatin cross-talk between H3K4 and H3K9 does not occur for ccKDM5B. Inhibition studies of ccKDM5B showed both in vitro and in cell inhibition of ccKDM5B by 2,4-pyridinedicarboxylic acid (2,4-PDCA) with a potency similar to that reported for the HDM KDM4C. Structure-guided sequence alignment indicated that the binding mode of 2,4-PDCA is conserved between KDM4A/C and KDM5B.

Identification of catechols as histone-lysine demethylase inhibitors

Identification of inhibitors of histone-lysine demethylase (HDM) enzymes is important because of their involvement in the development of cancer. An ELISA-based assay was developed for identification of inhibitors of the HDM KDM4C in a natural products library. Based on one of the hits with affinity in the low μM range (1, a catechol), a subset of structurally related compounds was selected and tested against a panel of HDMs. In this subset, two inhibitors (2 and 10) had comparable affinities towards KDM4C and KDM6A but no effect on PHF8. One inhibitor restored H3K9me3 levels in KDM4C transfected U2-OS cells.

What a difference an X or Y makes: sex chromosomes, gene dose, and epigenetics in sexual differentiation

A modern general theory of sex determination and sexual differentiation identifies the factors that cause sexual bias in gene networks, leading to sex differences in physiology and disease. The primary sex-biasing factors are those encoded on the sex chromosomes that are inherently different in the male and female zygotes. These factors, and downstream factors such as gonadal hormones, act directly on tissues to produce sex differences and antagonize each other to reduce sex differences. Recent studies of mouse models such as the four core genotypes have begun to distinguish between the direct effects of sex chromosome complement (XX vs. XY) and hormonal effects. Several lines of evidence implicate epigenetic processes in the control of sex differences, although a great deal of information is needed about sex differences in the epigenome.