Table of Contents:
In vitro 17β-hydroxysteroid dehydrogenases (17β HSDs) can catalyze the oxidation and the reduction of the hydroxyl or keto group in position 17 of steroid hormones. Because of their ability to modulate the concentration of active and inactive steroids in the cell, these enzymes can be described as prereceptor molecular switches. So far, fourteen isoforms have been identified, of which twelve can be found in human tissues. In the recent years 17β HSD types 1-3 gained major interest as potential drug targets for the treatment of sex steroid hormone-related diseases for which different “proof of concepts” were described, whereas the other 17β HSDs are known to be multifunctional.
17β-HSD14 is the latest isoform discovered and it belongs to the short-chain dehydrogenase-reductase family. In vitro the enzyme can catalyze the oxidation of estradiol in the presence of the cofactor NAD+. However, in vivo, the natural substrate is still unknown. This enzyme is predominantly expressed in the brain, liver, placenta and in the kidney and it was revealed to be cytosolic.
The goal of the project was to further characterize the protein to get a deeper insight into the function of 17β-HSD14. Inhibitors are useful chemical tools to elucidate the binding site of an enzyme and even more to disclose the physiological role of the latter upon in vivo administration. As no inhibitor has been reported for this enzyme by that time, the aim of this work was to identify the first nonsteroidal 17β-HSD14 inhibitors.
From the existing 3D-structure of the target enzyme no information about the protein/ligand interactions could be gained. Even more, it was not evident whether the crystallized enzyme, showing an extensive and open site cleft, was present in its active conformation. Therefore, no structure-based inhibitor design and no docking studies could be applied. Instead, a ligand-based approach was utilized starting from a small library of 17β-HSD1 and HSD2 inhibitors, which were tested for type 14 inhibitory activity. The initial hit identified in this preliminary screen (Ki around 250 nM) was optimized to result in six highly active compounds with Ki < 15 nM, showing good physicochemical properties, which should be associated with a promising bioavailability profile. Furthermore, these compounds, bearing a 2,6-pyridine ketone motife, also possess a good selectivity profile with respect to 17β-HSD1 and -2. In a close collaboration with NICOLE BERTOLETTI (KLEBE group, Philipps-University Marburg) it was possible to obtain crystal structures of five inhibitors in ternary complex with the target enzyme, which provided together with the biological results the basis for the understanding of an initial first structure-activity relationship of 17β-HSD14. It appears that an inhibitor’s acidic phenol group, being in close H-bond distance with Tyr154 and Ser141 of the catalytic triad, is essential to achieve high potency. Thereby the strong interaction with Tyr is stabilized through an extensive H-bonding network.
The determined crystal structures give important insights to characterize the target protein and even more they provide the basis for a structure-based inhibitor modification. The scaffold of a promising 2,6-pyridine ketone derivative was used as starting point to improve the structure of the inhibitor for optimal binding to 17β-HSD14. In a first optimization strategy, the focus was set on the carbonyl group, which only serves to induce a V shape of the inhibitor. As no direct interaction of this group was observed in the crystal structure, this motif was replaced by different substituents. Even though a ten-fold higher binding affinity toward 17β-HSD14 was observed for one compound bearing a methylene group, all the newly synthesized compounds showed a less satisfactory selectivity profile toward the types 1 and 2 compared to the initial inhibitor. In a second optimization round the pyridine core was extended to address a hydrophobic pocket observed in the crystal structures. A quinoline-based inhibitor was designed and modeled into the active site of the target protein prior to its synthesis. The predicted binding mode was verified later on with a co-crystal structure. The resulting highly active quinoline (Ki = 12 nM) was decorated with different substituents showing a varying hydrophobicity profile. Four crystal structures of different qunioline-based inhibitors in ternary complex with 17β HSD14 were obtained, revealing the variation of one inhibitor’s binding mode due to a small change in its substituents. In the new binding mode the previously mentioned phenol group is no longer in H-bond distance to Tyr154, which was described to be essential to achieve a good affinity toward the target. Among others, Tyr154 is forming instead an H-bond network via a water molecule with the carbonyl group of the ligand. This different binding mode was not related to a significant change in binding affinity for this compound (Ki = 12 nM) compared to its starting structure. Including the initial pyridine ketone derivative of this study, three quinoline-based inhibitors show a good selectivity profile toward the 17β-HSD enzymes 1, 2 and 10. Furthermore, they display very low cytotoxicity and they have no effect on the multi-drug resistance protein P-gp. Together with their calculated physicochemical properties, two ligands can be highlighted as promising tool compounds to investigate the physiological role of 17β HSD14 after in vivo administration.
In the last part of this thesis it was the goal to discover new inhibitor scaffolds. Two fragments showing very low inhibitory potency toward 17β-HSD14 were identified in a fragment-based lead discovery (FBLD) campaign by Nicole Bertoletti. In a fragment-growing approach compounds with a higher target affinity could be obtained, highlighting this method as a promising strategy for the identification of new lead structures.