the SAR of 6-aryl-5-cyano-pyrimidine LSD1 inhibitors and shed light onto the effect
of the different substituents on the inhibitory activity of the compounds.
Moreover, combination of available structural data for LSD1, docking studies and
medicinal chemistry approaches were frequently used to guide the optimization of
novel LSD1 inhibitors [146–148]. For instance, Wu et al. reported on the structurebased design of a series of 3-(piperidin-4-ylmethoxy)pyridine derivatives, such as
compound 33 (IC 50 29 nM, Fig. 7), which exhibited high inhibitory activity against
LSD1, high selectivity over MAO-A/MAO-B, cellular activity, and inhibition of
tumor cells proliferation. Very recently, the structure-based optimization of a novel
chemical series of thieno[3,2-b]pyrrole-5-carboxamide LSD1 inhibitors, which had
been initially identified by an HTS campaign, was also reported. Structural optimization culminated in the discovery of the thieno[3,2-b]pyrrole-5-carboxamide 34
(IC 50 7.8 nM, Fig. 7) which was also successfully cocrystallized with LSD1 [147].
5.2 Jumonji Histone Demethylases
JmjC domain-containing proteins belong to the cupin superfamily of metalloenzymes
[149] and comprise about 20 (usually multidomain) proteins which are subdivided
into seven subfamilies (KDM2–KDM8) [129, 150]. JmjC domains characteristically
show a conserved antiparallel β-barrel fold which harbors the binding site for Fe
2+ and
α-KG cofactors. The catalytic Fe
2+ is coordinated in an octahedral fashion by three
conserved residues in the catalytic pocket (two histidine and one aspartate/glutamate),
besides the cofactor α-KG and a water molecule. Meanwhile, the substrate binding
pocket of the different JmjC subfamilies shows great structural variations and can thus
be addressed for the development of selective inhibitors [151]. Numerous crystal
structures of various JmjC domains in complex with inhibitors have appeared in the
last few years, which shed light onto the inhibitory mechanism of these compounds.
All inhibitors reported so far are metal chelators and bind competitively in the α-KG
binding pocket [129, 152]. Hence, many of them lack selectivity and inhibit different
JmjC protein subtypes as well as other Fe
2+ -/α-KG-dependent enzymes, such as
prolyl hydroxylase domain-containing proteins (PHDs) and factor inhibiting
hypoxia-inducible factor (FIH). While targeting the metal ion is obviously necessary
for the activity of the inhibitors, simultaneously addressing the substrate binding
pocket represents a promising strategy to achieve selectivity [153].
The first VS for JmjC proteins was reported by Chu et al. [154], who used an
available crystal structure of JMJD2A (KDM4A) and docked the NCI dataset
(around 240,000 compounds). Among the ten compounds selected for in vitro
screening, the 2,4-dinitrobenzene derivative 35 (Fig. 8) showed the highest inhibitory activity against JMJD2A (IC 50 6.4 μM) and JMJD2B (IC 50 9.3 μM) and
selectivity over JMJD2D and JMJD2E. However, the identified inhibitor does not
possess a classical metal ion-chelating group, and the binding mode of the molecule
is not clear, even if the docking study predicted an interaction between the nitro
group and the metal ion.
472
D. Robaa et al.
of the different substituents on the inhibitory activity of the compounds.
Moreover, combination of available structural data for LSD1, docking studies and
medicinal chemistry approaches were frequently used to guide the optimization of
novel LSD1 inhibitors [146–148]. For instance, Wu et al. reported on the structurebased design of a series of 3-(piperidin-4-ylmethoxy)pyridine derivatives, such as
compound 33 (IC 50 29 nM, Fig. 7), which exhibited high inhibitory activity against
LSD1, high selectivity over MAO-A/MAO-B, cellular activity, and inhibition of
tumor cells proliferation. Very recently, the structure-based optimization of a novel
chemical series of thieno[3,2-b]pyrrole-5-carboxamide LSD1 inhibitors, which had
been initially identified by an HTS campaign, was also reported. Structural optimization culminated in the discovery of the thieno[3,2-b]pyrrole-5-carboxamide 34
(IC 50 7.8 nM, Fig. 7) which was also successfully cocrystallized with LSD1 [147].
5.2 Jumonji Histone Demethylases
JmjC domain-containing proteins belong to the cupin superfamily of metalloenzymes
[149] and comprise about 20 (usually multidomain) proteins which are subdivided
into seven subfamilies (KDM2–KDM8) [129, 150]. JmjC domains characteristically
show a conserved antiparallel β-barrel fold which harbors the binding site for Fe
2+ and
α-KG cofactors. The catalytic Fe
2+ is coordinated in an octahedral fashion by three
conserved residues in the catalytic pocket (two histidine and one aspartate/glutamate),
besides the cofactor α-KG and a water molecule. Meanwhile, the substrate binding
pocket of the different JmjC subfamilies shows great structural variations and can thus
be addressed for the development of selective inhibitors [151]. Numerous crystal
structures of various JmjC domains in complex with inhibitors have appeared in the
last few years, which shed light onto the inhibitory mechanism of these compounds.
All inhibitors reported so far are metal chelators and bind competitively in the α-KG
binding pocket [129, 152]. Hence, many of them lack selectivity and inhibit different
JmjC protein subtypes as well as other Fe
2+ -/α-KG-dependent enzymes, such as
prolyl hydroxylase domain-containing proteins (PHDs) and factor inhibiting
hypoxia-inducible factor (FIH). While targeting the metal ion is obviously necessary
for the activity of the inhibitors, simultaneously addressing the substrate binding
pocket represents a promising strategy to achieve selectivity [153].
The first VS for JmjC proteins was reported by Chu et al. [154], who used an
available crystal structure of JMJD2A (KDM4A) and docked the NCI dataset
(around 240,000 compounds). Among the ten compounds selected for in vitro
screening, the 2,4-dinitrobenzene derivative 35 (Fig. 8) showed the highest inhibitory activity against JMJD2A (IC 50 6.4 μM) and JMJD2B (IC 50 9.3 μM) and
selectivity over JMJD2D and JMJD2E. However, the identified inhibitor does not
possess a classical metal ion-chelating group, and the binding mode of the molecule
is not clear, even if the docking study predicted an interaction between the nitro
group and the metal ion.
472
D. Robaa et al.