Benzothiazolyl Ureas are Low Micromolar and Uncompetitive Inhibitors of 17β-HSD10 with Implications to Alzheimer’s Disease Treatment

Human 17β-hydroxysteroid dehydrogenase type 10 is a multifunctional protein involved in many enzymatic and structural processes within mitochondria. This enzyme was suggested to be involved in several neurological diseases, e.g., mental retardation, Parkinson’s disease, or Alzheimer’s disease, in which it was shown to interact with the amyloid-beta peptide. We prepared approximately 60 new compounds based on a benzothiazolyl scaffold and evaluated their inhibitory ability and mechanism of action. The most potent inhibitors contained 3-chloro and 4-hydroxy substitution on the phenyl ring moiety, a small substituent at position 6 on the benzothiazole moiety, and the two moieties were connected via a urea linker (4at, 4bb, and 4bg). These compounds exhibited IC50 values of 1–2 μM and showed an uncompetitive mechanism of action with respect to the substrate, acetoacetyl-CoA. These uncompetitive benzothiazolyl inhibitors of 17β-hydroxysteroid dehydrogenase type 10 are promising compounds for potential drugs for neurodegenerative diseases that warrant further research and development.


Introduction
Human 17β-hydroxysteroid dehydrogenase type 10 (17β-HSD10) is an enzyme belonging to the superfamily of short-chain dehydrogenases/reductases (SDR) (SDR5C1 in SDR nomenclature) ( Figure 1). The compound designated AG18051 ((1-azepan-1-yl-2-phenyl-2-(4-thioxo-1,4dihydropyrazolo [3-d] pyrimidin-5-yl)-ethanone) was found to be a very potent inhibitor of the 17β-HSD10 enzyme, with an IC50 value of 92 nM, that functions irreversibly [19]. Other novel 17β-HSD10 enzyme inhibitors were developed with the aim to treating androgen-and estrogen-dependent diseases. The best hit compound, a steroidal inhibitor designated as RM-532-46 (Figure 1), had an IC50 value of 0.55 μM in vitro using an HEK-293 cell line stably overexpressing 17β-HSD10, and this derivative was hypothesized to be a reversible inhibitor of the 17β-HSD10 enzyme [20]. A subsequent study of the steroid inhibitor RM-532-46 and non-steroidal benzothiazole phosphonate derivatives that used purified enzyme and stably transfected HEK-293 cells revealed that the transformation of 17β-estradiol to estrone was inhibited with an IC50 of 1.7 μM [21]. Moreover, this steroidal inhibitor was also active against the 17β-HSD3 enzyme; thus, this inhibitor is more likely to be a promising drug candidate for the treatment of prostate cancer [20]. Recently, several publications came out focusing on development of 17β-HSD10 inhibitors that were structurally derived from benzothiazolyl urea scaffold of frentizole (Figure 1), however, the mode of inhibition has not been established in these publications, except of the study by Aitken et al. [22][23][24][25][26]. In 2016 Hroch et al. published a series of benzothiazolyl ureas with the best compounds showing 60% decrease in 17β-HSD10 activity in 25 μM screening ( Figure 1; compounds 13, 14) [22]. In 2017 Hroch et al. published a series of indolyl urea compounds that were designed as dual inhibitors of 17β-HSD10 and monoamine oxidase enzymes. However, there was no improvement in 17β-HSD10 inhibition compared to the previous publication [23]. In 2017 Benek et al. published series of frentizole analogues with urea linker replaced with thiourea or guanidine. The most promising compound with guanidine linker showed good inhibition activity with IC50 3.06 μM [24]. The latest publication from 2019 by Aitken et al. presents large structure-activity relationship (SAR) study addressing changes to benzothiazole moiety, urea linker and phenyl moiety of frentizole and previously identified inhibitors 13 and 14 ( Figure 1). The most promising compounds identified in this study showed IC50 values around 1-2 μM with mixed type of inhibition mechanism [26]. In this work, we aimed to design, synthesize, and evaluate new compounds based on benzothiazolyl urea scaffold with enhanced inhibitory ability against the human 17β-HSD10 enzyme. As a starting point for the design of novel compounds, the most potent benzothiazolyl urea-based 17β-HSD10 inhibitors, published by Hroch et al. [22], were employed ( Figure 1; compounds 13 and  14). The structural scaffold was divided into four separate parts and each part was modified to identify the structure-activity relationship (SAR) for a particular molecular fragment ( Figure 2). In this work, we aimed to design, synthesize, and evaluate new compounds based on benzothiazolyl urea scaffold with enhanced inhibitory ability against the human 17β-HSD10 enzyme. As a starting point for the design of novel compounds, the most potent benzothiazolyl urea-based 17β-HSD10 inhibitors, published by Hroch et al. [22], were employed ( Figure 1; compounds 13 and 14). The structural scaffold was divided into four separate parts and each part was modified to identify the structure-activity relationship (SAR) for a particular molecular fragment ( Figure 2).

Figure 2.
Design of novel 17β-HSD10 inhibitors. The original scaffold was divided into for parts (depicted in violet, blue, green, and red) and each part was modified separately.

Chemical Synthesis
Generally, the compounds were prepared in a two-step process. Firstly, the corresponding benzothiazole-2-amine, benzoxazole-2-amine, or benzimidazole-2-amine (1) was treated with 1,1′carbonyldiimidazole to form an imidazolyl intermediate (2), which was then treated with the corresponding aniline derivative (3) to yield the 1,3-disubstituted urea (4) as a final product (Scheme 1) [22,27]. The thiourea analogue (4ak) was prepared in a similar way by just using 1,1′thiocarbonyldiimidazole instead of 1,1′-carbonyldiimidazole in the first reaction step. An inverse reaction setup was used for the synthesis of the final products 4aj and 4bb, i.e., CDI was first reacted with the corresponding aniline derivative, and the resulting intermediate was treated with 6-hydroxybenzothiazole-2-amine (1f) to give the final product 4aj. O-Demethylation of compound 4aj then yielded the final product 4bb (Scheme 2).

Figure 2.
Design of novel 17β-HSD10 inhibitors. The original scaffold was divided into for parts (depicted in violet, blue, green, and red) and each part was modified separately.

Chemical Synthesis
Generally, the compounds were prepared in a two-step process. Firstly, the corresponding benzothiazole-2-amine, benzoxazole-2-amine, or benzimidazole-2-amine (1) was treated with 1,1′carbonyldiimidazole to form an imidazolyl intermediate (2), which was then treated with the corresponding aniline derivative (3) to yield the 1,3-disubstituted urea (4) as a final product (Scheme 1) [22,27]. The thiourea analogue (4ak) was prepared in a similar way by just using 1,1′thiocarbonyldiimidazole instead of 1,1′-carbonyldiimidazole in the first reaction step. An inverse reaction setup was used for the synthesis of the final products 4aj and 4bb, i.e., CDI was first reacted with the corresponding aniline derivative, and the resulting intermediate was treated with 6-hydroxybenzothiazole-2-amine (1f) to give the final product 4aj. O-Demethylation of compound 4aj then yielded the final product 4bb (Scheme 2). An inverse reaction setup was used for the synthesis of the final products 4aj and 4bb, i.e., CDI was first reacted with the corresponding aniline derivative, and the resulting intermediate was treated with 6-hydroxybenzothiazole-2-amine (1f) to give the final product 4aj. O-Demethylation of compound 4aj then yielded the final product 4bb (Scheme 2).

Recombinant Enzyme Production
The human recombinant 17β-HSD10 enzyme was expressed and purified using standard molecular biological and chromatographic techniques. Briefly, the human 17β-HSD10 enzyme was expressed as polyhistidine-tagged fusion protein using bacterial expression system and autoinduction culture medium (detailed procedures are presented in the Material and Methods section). Immobilized metal affinity chromatography was used for polyhistidine-tagged recombinant 17β-HSD10 purification. The results of purification procedures were confirmed using SDS-PAGE analysis, followed by Western blotting analysis, resulting in a monomeric molecular mass of approximately 27 kDa (Figure 3). The typical yield obtained after a single purification procedure, starting from 50 mL of overexpressed bacterial culture prepared in autoinduction culture medium, was 1.5 mg, and the purity was over 95%. All synthesized final products (4a-4bg) were sufficiently characterized by 1 H NMR, 13 C NMR spectroscopy and high-resolution mass spectrometry (HRMS) analyses. Intermediate products were generally characterized only by 1 H NMR (see Supplementary Materials). In case of fluorinated final products (4a-4u and 4az) 19 F NMR spectra were recorded. To remove potential aluminum residues (demethylation using AlCl 3 ) or mercury residues (guanidine formation using HgO) corresponding final products and/or their intermediates were purified by suitable procedures, commonly used for this purpose, e.g., extraction followed by column chromatography or filtration through Celite followed by precipitation and filtration of the product (detailed synthetic procedures can be found in Supplementary Materials).

Recombinant Enzyme Production
The human recombinant 17β-HSD10 enzyme was expressed and purified using standard molecular biological and chromatographic techniques. Briefly, the human 17β-HSD10 enzyme was expressed as polyhistidine-tagged fusion protein using bacterial expression system and autoinduction culture medium (detailed procedures are presented in the Material and Methods Section). Immobilized metal affinity chromatography was used for polyhistidine-tagged recombinant 17β-HSD10 purification. The results of purification procedures were confirmed using SDS-PAGE analysis, followed by Western blotting analysis, resulting in a monomeric molecular mass of approximately 27 kDa (Figure 3). The typical yield obtained after a single purification procedure, starting from 50 mL of overexpressed bacterial culture prepared in autoinduction culture medium, was 1.5 mg, and the purity was over 95%.

17β-HSD10 Reductase Assay and Enzyme Kinetics
The enzyme activity of the human recombinant 17β-HSD10 was determined from previously published methods using acetoacetyl-CoA (AAC) as a substrate [4,19,38,39]. The method was based on the decrease in nicotinamide adenine dinucleotide (NADH) cofactor absorbance at 340 nm during its utilization by the enzyme. To establish the reaction properties correctly, the enzyme kinetic parameters were determined. The Michaelis constant (Km) obtained for acetoacetyl-CoA as a substrate, 79.2 ± 15.0 μM at pH 7.4, corresponded to the previously reported values of 117 ± 28 μM [38] and 89 ± 5.4 μM [3] for the same substrate. The maximum reaction rate, Vmax, for the enzyme, was estimated to be 27.7 ± 1.8 μM·min −1 . Based on known Km values, the substrate concentration for compound screening and IC50 evaluation was chosen as 320 μM (4 × the Km value).

Screening Inhibitory Effects of Novel Compounds
Initially, all newly synthesized compounds were screened to determine inhibitory potency against 17β-HSD10 enzyme at 10 μM. Previously identified benzothiazolyl urea inhibitors 13 and 14 [22], together with the irreversible inhibitor AG18051 [19], and the template structure frentizole [16], were assayed as the standards (Table 1).

17β-HSD10 Reductase Assay and Enzyme Kinetics
The enzyme activity of the human recombinant 17β-HSD10 was determined from previously published methods using acetoacetyl-CoA (AAC) as a substrate [4,19,38,39]. The method was based on the decrease in nicotinamide adenine dinucleotide (NADH) cofactor absorbance at 340 nm during its utilization by the enzyme. To establish the reaction properties correctly, the enzyme kinetic parameters were determined. The Michaelis constant (K m ) obtained for acetoacetyl-CoA as a substrate, 79.2 ± 15.0 µM at pH 7.4, corresponded to the previously reported values of 117 ± 28 µM [38] and 89 ± 5.4 µM [3] for the same substrate. The maximum reaction rate, V max , for the enzyme, was estimated to be 27.7 ± 1.8 µM·min −1 . Based on known K m values, the substrate concentration for compound screening and IC 50 evaluation was chosen as 320 µM (4 × the K m value).

Screening Inhibitory Effects of Novel Compounds
Initially, all newly synthesized compounds were screened to determine inhibitory potency against 17β-HSD10 enzyme at 10 µM. Previously identified benzothiazolyl urea inhibitors 13 and 14 [22], together with the irreversible inhibitor AG18051 [19], and the template structure frentizole [16], were assayed as the standards ( Table 1).
The 13 most potent compounds (residual activity < 55%) showed similar or even stronger inhibitory effects on 17β-HSD10 compared to the previously published compounds 13 and 14 (Figure 1) [22]. All these compounds were used for the determination of IC 50 values, which ranged from~1 to 7 µM ( Table 1). The irreversible 17β-HSD10 inhibitor AG18051 was used as a control showing more than 96% of enzyme inhibition at 10 µM and the IC 50 value for this compound was determined as 97 nM, which is consistent with the published data (92 nM; [19]).
The most potent inhibitors 4at, 4bb, and 4bg (IC 50 < 2 µM) were further used for kinetic experiments to determine the type of inhibition. The acquired data were linearized by using Lineweaver-Burk and Hanes-Wolf plots ( Figure 4A,B). K m and V max values in presence of inhibitors were assessed with respect to the uninhibited 17β-HSD10 enzymatic reaction ( Figure 4C). All three inhibitors decreased both the maximum reaction velocity and the Michaelis constant relative to the uninhibited enzyme ( Figure 4C). The performed kinetic experiments indicate that the selected inhibitors (4at, 4bb, and 4bg) act via an uncompetitive mode of action with respect to AAC as a substrate of the reaction. Such inhibitors are only able to bind the enzyme-substrate complex and the substrate concentration enhances the inhibitory ability (in contrast to competitive inhibition).

Screening Inhibitory Effects of Novel Compounds
Initially, all newly synthesized compounds were screened to determine inhibitory potency against 17β-HSD10 enzyme at 10 μM. Previously identified benzothiazolyl urea inhibitors 13 and 14 [22], together with the irreversible inhibitor AG18051 [19], and the template structure frentizole [16], were assayed as the standards (Table 1). The 13 most potent compounds (residual activity < 55%) showed similar or even stronger inhibitory effects on 17β-HSD10 compared to the previously published compounds 13 and 14 ( Figure  1) [22]. All these compounds were used for the determination of IC50 values, which ranged from ~1 to 7 μM ( Table 1). The irreversible 17β-HSD10 inhibitor AG18051 was used as a control showing more than 96% of enzyme inhibition at 10 μM and the IC50 value for this compound was determined as 97 nM, which is consistent with the published data (92 nM; [19]).
The most potent inhibitors 4at, 4bb, and 4bg (IC50 < 2 μM) were further used for kinetic experiments to determine the type of inhibition. The acquired data were linearized by using Lineweaver-Burk and Hanes-Wolf plots ( Figure 4A,B). Km and Vmax values in presence of inhibitors were assessed with respect to the uninhibited 17β-HSD10 enzymatic reaction ( Figure 4C). All three inhibitors decreased both the maximum reaction velocity and the Michaelis constant relative to the uninhibited enzyme ( Figure 4C). The performed kinetic experiments indicate that the selected inhibitors (4at, 4bb, and 4bg) act via an uncompetitive mode of action with respect to AAC as a substrate of the reaction. Such inhibitors are only able to bind the enzyme-substrate complex and the substrate concentration enhances the inhibitory ability (in contrast to competitive inhibition).

Discussion
In our previous work, compounds harboring the 3-chloro-4-hydroxy substitution on the phenyl ring alongside the 6-substitution (fluorine, chlorine, trifluoromethyl) on the benzothiazole moiety ( Figure 2; compounds 13 and 14) showed good inhibitory activities towards 17β-HSD10, resulting in less than 40% residual enzymatic activity at 25 µM, however, their mechanism of inhibition has not been determined [22]. Within this study we made different modification into the structure of previous hits (compounds 13 and 14) in order to establish the SAR and find more potent 17β-HSD10 inhibitors, and further we performed kinetic experiments to establish the mode of inhibition.
Firstly, a series was prepared that combined different patterns of fluorine, hydroxy, or methoxy substitutions of the phenyl ring, together with fluorine, chlorine, or methoxy substituents in position 6 of the benzothiazole moiety (4a-4u). Fluorine substitution of the phenyl ring was used instead of the original chlorine to improve physico-chemical properties, and methoxy substitution of the benzothiazole moiety instead of the original halogen substituent was introduced for the same reason. Most compounds within this series were found to be less active thanthe parental compounds 13 and 14. Any deviation from the original 3-chloro, 4-hydroxy substitution of the phenyl ring resulted in decrease in inhibitory activity except for 3,5-fluoro, 4-hydroxy substitution pattern (4g-4i), which resulted in inhibition comparable to the parental molecules 13 and 14. There was no significant difference in activity between compounds substituted with fluorine, chlorine of methoxy group in the position 6 of benzothiazole moiety.
Secondly, we explored the SAR of the phenyl ring by changing the position and amount of hydroxy and chlorine substituents and, further, by introducing methoxy and carboxy groups, while the original 6-chlorobenzothiazole moiety was fixed along with the urea linker (4v-4aj). However, none of modification led to stronger inhibitory ability compared to the parental compound 14.
We also focused on the urea linker, which was replaced with thiourea or guanidine. These substitutions (4ak, 4al) did not lead to enhanced inhibitory ability and resulted in more than 70% of residual enzyme activity.
In the next series, the original 6-chlorine substitution of the benzothiazole moiety was moved to alternative positions 4, 5, and 7 (4am, 4an, and 4ao). Derivatives with chlorine in position 5 or 6 showed inhibition similar to parental compound 14 (~55% residual activity). While the substitution in position 7 (4ao) increased the inhibition (38% residual activity).
In the next step, we have selected 13 compounds (4an-4ap, 4ar-4au, 4ay-4bb, 4bf, and 4bg), which showed similar or even stronger inhibitory effects on 17β-HSD10 than the previously published compounds 13 and 14, to determine their IC 50 values. All these compounds shared a common phenyl part, with 3-chlorine, 4-hydroxy substitutions, together with the urea linker, but they differed in the substitution of the benzothiazole, benzoxazole, or benzimidazole moiety. The obtained IC 50 values ranged from 1 to 7 µM (Table 1).
The three most potent inhibitors 4at, 4bb, and 4bg with IC 50 < 2 µM were further used for kinetic experiments to determine the type of inhibition. From this point of view, all the selected inhibitors were found to act via an uncompetitive mode of action with respect to acetoacetyl-CoA as a substrate of the reaction. This is in contrast to the mixed inhibition mode described for similar type of inhibitors by Aitken et al. [26]. However, when looking closely at the previously published data, the mode of inhibition seems to be identical. This is because the uncompetitive inhibition can be considered as a subtype of mixed inhibition and deeper analysis of the previous kinetic experiment would probably yield the same conclusion, i.e., uncompetitive mode of inhibition. In drug research and development, uncompetitive inhibitors are relatively rare, although they can provide unique physiological consequences. The inhibitors can bind the target only if the substrate is present and this mechanism of action should be favorable for enzymes with cellular substrate concentration higher than the enzyme K m value [40]. Several drugs act in an uncompetitive manner towards dehydrogenases or reductases with a NAD + /NADH pair as their enzyme cofactor: e.g., mycophenolic acid, reversible inhibitor of inosine 5 -monophosphate dehydrogenase [41,42]; ononetin, an inhibitor of dihydrofolate reductase [43]; or epristeride, an inhibitor of human steroid 5α-reductase [44].
Development of uncompetitive inhibitors is a very valuable strategy in drug discovery in terms of specificity to the target when such inhibitors can only recognize the enzyme-substrate complex and therefore should not inhibit other enzymes with a similar structure. For this reason, the presented uncompetitive inhibitors of 17β-HSD10 appear to be promising molecules for further research and development.

General Chemistry
Solvents and reagents were purchased from Fluka and Sigma-Aldrich (Prague, Czech Republic) and used without further purification. Reactions were monitored by thin layer chromatography performed on aluminum sheets pre-coated with silica gel 60 F 254 (Merck, Prague, Czech Republic) and detected under 254 nm UV light. Column chromatography was performed on a silica gel 60 column (230 mesh). Melting points were measured by using a Stuart SMP30 melting point apparatus (Bibby Scientific Limited, Staffordshire, UK) and were uncorrected. 1 H and 13 C NMR spectra were recorded at Varian Gemini 300 ( 1 H 300 MHz, 13 C 75 MHz, Palo Alto CA, USA) or Varian S500 ( 1 H 500 MHz, 13 C 126 MHz, Palo Alto CA, USA). In all cases, the chemical shift values for 1 H spectra were reported in ppm (δ) relative to residual CHD 2 SO 2 CD 3 (δ 2.50) or CDCl 3 (δ 7.27); shift values for 13 C spectra are reported in ppm (δ) relative to the solvent peak for dimethylsulfoxide-d 6 (δ 39.52) or CDCl 3 (δ 77.2). Proton decoupled 19 F NMR spectra were recorded on a Bruker AVANCE III HD 500 spectrometer (Billerica, MA, USA) operating at 470.55 MHz for fluorine using 5 mm broadband tunable probe. Samples were dissolved in dimethylsulfoxide-d 6 and 19 F chemical shifts were referred to internal CFCl 3 .
For HRMS determination, a Dionex UltiMate 3000 analytical LC-MS system coupled with a Q Exactive Plus hybrid quadrupole-orbitrap spectrometer (both produced by ThermoFisher Scientific, Bremen, Germany) was used. The LC-MS system consisted of a binary pump HPG-3400RS connected to a vacuum degasser, a heated column compartment TCC-3000, an autosampler WTS-3000 equipped with a 25 µL loop, and a VWD-3000 ultraviolet detector. A Waters Atlantis dC18 100Å (2.1 × 100 mm/ 3 µm) column was used as the stationary phase. The analytical column was protected against mechanical particles by an in-line filter (Vici Jour) with a frit with 0.5 µm pores. Water (MFA) and acetonitrile (MFB) used in the analyses were acidified with 0.1% (v/v) formic acid. Ions for mass spectrometry were generated by heated electro-spray ionization source (HESI) working in positive mode, with the following settings: sheath gas flow rate 40, aux gas flow rate 10, sweep gas flow rate 2, spray voltage 3.2 kV, capillary temperature 350 • C, aux gas temperature 300 • C, S-lens RF level 50, microscans 1, maximal injection time 35 ms, resolution 140,000. The full-scan MS analyses monitored ions within m/z range 100-1500. The studied compounds were dissolved in methanol, and 1 µL of the solution was injected into the LC-MS system. For elution, following ramp-gradient program was used: 0-1 min: 10% MFB, 1-4 min: 10-100% MFB, 4-5 min: 100% MFB, 5-7.5 min: 10% MFB. The flow-rate in the gradient elution was set to 0.4 mL/min. To increase the accuracy of HRMS, internal lock-mass calibration was employed using polysiloxane traces of m/z = 445.12003 ([M+H] + , [C 2 H 6 SiO] 6 ) present in the mobile phases. The chromatograms and mass spectra were processed in Chromeleon 6.80 and Xcalibur 3.0.63 software, respectively (both produced by ThermoFisher Scientific, Bremen, Germany).
Novelty of prepared final products was checked using Reaxys database (www.reaxys.com). Three final products were found not to be novel structures (4v, 4w and 4af). Two of those compounds, 4w [45] and 4af [16], were previously mentioned in scientific articles and compound 4v is indexed within Pubchem database (https://pubchem.ncbi.nlm.nih.gov) and can be supplied by commercial vendors. However, none of those compounds has ever been tested for inhibition of 17β-HSD10 enzyme.  The general reaction mixture for the reductase activity measurement consisted of 320 µM AAC, 320 µM NADH as the enzyme cofactor and 0.15 µg of recombinant enzyme in assay buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM dithiothreitol, 0.001% Tween 20 and 0.01% bovine serum albumin) in 100 µL of reaction buffer. The process was as follows: the reaction mixture of enzyme and cofactor in assay buffer was preincubated for 5 min at 37 • C prior to substrate addition. The reaction was performed in 96-well polystyrene plates at 37 • C (Tecan Spark 10M, Männedorf, Switzerland), and enzyme activity was recorded as the change in absorbance at 340 nm over 1 min and calculated using the molar extinction coefficient of NADH (ε = 6220 M −1 ·cm −1 ).

17β-HSD10 Enzyme Inhibition Screening and IC 50 Determination
The novel compounds were screened at 10 µM or 1 µM to determine their ability to inhibit 17β-HSD10 enzyme. The tested compounds were dissolved in dimethyl sulfoxide (DMSO) to a stock concentration of 10 mM and then diluted into deionized water to get the working concentration. The inhibitory screening was based on AAC reductase assay. The remaining enzyme activity was measured at 340 nm at 20 s intervals for 5 min and determined as the difference in absorbance between inhibited and non-inhibited enzymatic reaction. The DMSO was used as well as the inhibitor as the vehicle control. The potent known 17β-HSD10 inhibitor AG18051 [19] was used as a control for the enzymatic reaction.
For compounds with at least 50% 17β-HSD10 inhibition in the 10 µM screening assay, the IC 50 was determined. For this purpose, the AAC reductase assay was used and the remaining enzyme activity was measured as dose-response inhibition with at least six different inhibitor concentrations (ranging from 0 to 3 µM) at fixed substrate, enzyme, and cofactor concentrations. The data were analyzed by using GraphPad Prism 7 software in non-linear regression, and IC 50 values were calculated for each inhibitor from at least three independent measurements, all in triplicate.

Determination of Inhibition Type
The determination of the inhibition type for three most potent compounds was made by using an AAC reductase activity assay. The inhibitors were used at different concentrations (0-3 µM) in combination with different substrate concentrations (25-400 µM) and saturated NADH concentration to determine the cofactor-based absorbance change measured at 340 nm. The uninhibited enzymatic reaction contained the same DMSO concentration as the inhibited one and was used as the vehicle control. For data evaluation, the linearization of Lineweaver-Burk and Hanes-Wolf was used. To determine the binding mechanism more precisely, the K m and V max values of inhibited reactions were compared with uninhibited reactions.

Conclusions
In conclusion, we have prepared and in vitro evaluated over 50 novel compounds based on a benzothiazolyl urea scaffold acting as 17β-HSD10 inhibitors. SAR study revealed crucial structural aspects required for good inhibitory potency, namely 3-chlorine and 4-hydroxy substitution on the phenyl ring, urea linker and small substituent at position 6 of the benzothiazole moiety. The most potent inhibitors were able to inhibit 17β-HSD10 activity with IC 50 at low micromolar level, thus being ten times more potent compared to parental inhibitors. The novel compounds were found to be uncompetitive inhibitors of 17β-HSD10 with selectivity towards the enzyme-substrate complex, which makes them interesting for further research and development as potential drugs.