Novel Benzothiazole-Based Ureas as 17β-HSD10 Inhibitors, A Potential Alzheimer’s Disease Treatment

It has long been established that mitochondrial dysfunction in Alzheimer’s disease (AD) patients can trigger pathological changes in cell metabolism by altering metabolic enzymes such as the mitochondrial 17β-hydroxysteroid dehydrogenase type 10 (17β-HSD10), also known as amyloid-binding alcohol dehydrogenase (ABAD). We and others have shown that frentizole and riluzole derivatives can inhibit 17β-HSD10 and that this inhibition is beneficial and holds therapeutic merit for the treatment of AD. Here we evaluate several novel series based on benzothiazolylurea scaffold evaluating key structural and activity relationships required for the inhibition of 17β-HSD10. Results show that the most promising of these compounds have markedly increased potency on our previously published inhibitors, with the most promising exhibiting advantageous features like low cytotoxicity and target engagement in living cells.


Introduction
There is a strong, well-documented connection between Alzheimer's disease (AD) and mitochondrial dysfunction [1][2][3]. Mitochondrial changes in AD patients are an early event, preceding the onset of amyloid plaque formation, and include morphology abnormalities and changes in metabolism stemming from alterations in the complexes of the electron transport chain, the enzymes in the tricarboxylic acid cycle (TCA), and changes in components of the mitochondrial membrane involved in import/export flux. Mitochondria are key to the production of adenosine triphosphate (ATP) via the metabolism of glucose and fatty acids. Mitochondrial dysfunction in AD includes activity changes in many enzymes involved in these processes and contributes to the reduction in energy metabolism in AD [4]. Mitochondrial dysfunction is also exacerbated by the presence of amyloid beta peptide (Aβ) within mitochondria [5]. One mitochondrial enzyme affected in AD is 17β-HSD10 (17β-hydroxysteroid dehydrogenase type 10, also known as amyloid-β binding alcohol dehydrogenase (ABAD) or 3-Hydroxyacyl-CoA dehydrogenase). 17β-HSD10's primary role is to HSD10 (17β-hydroxysteroid dehydrogenase type 10, also known as amyloid-β binding alcohol dehydrogenase (ABAD) or 3-Hydroxyacyl-CoA dehydrogenase). 17β-HSD10's primary role is to utilise several substrates to produce energy in the β-fatty acid oxidation pathway, the energy source when glucose levels are low, playing a prominent role in AD, where glucose metabolism is significantly decreased [6]. Importantly, we and others have shown that inhibition of this enzyme is beneficial in both in vitro and in vivo AD models in its own right and also protects against Aβ toxicity in both cellular and transgenic mouse models of AD [7][8][9][10][11][12]. A current working hypothesis is that by inhibiting the enzyme activity of 17β-HSD10 (a contributor to the β-fatty acid oxidation pathway) this can help re-balance alterations in glucose metabolism observed in AD (Aitken unpublished data).
17β-HSD10 was first identified as an Aβ binding protein in 1997 [13], a finding which has subsequently been confirmed using a number of techniques [5,13,14]. 17β-HSD10 is known to interact with the two major plaque forming isoforms of Aβ, namely Aβ(1-40) and Aβ , leading to distortion of the enzyme structure and inhibition of its normal function as an energy provider for cells [15,16]. In vitro experiments have shown that the interaction between 17β-HSD10 and Aβ is cytotoxic and 17β-HSD10's function is altered with a build-up of reactive oxygen species (ROS) and toxins leading to mitochondrial dysfunction. Using site-directed mutagenesis and surface plasmon resonance protein interaction assays (SPR), Lustbader et al. identified the LD loop of the 17β-HSD10 protein as the binding site for Aβ and subsequently synthesised a 28-amino acid peptide encompassing this region, which was termed the 17β-HSD10 decoy peptide [5]. Using SPR assays it has been shown that this 17β-HSD10 decoy peptide can prevent the binding of 17β-HSD10 to Aβ  and Aβ . Significantly, inhibition of the interaction between 17β-HSD10 and Aβ by the 17β-HSD10-decoy peptide was shown to translate into a cytoprotective effect in cell culture experiments. Cortical neurons exposed to Aβ(1-42) showed a significant increase in cell death, as measured by cytochrome-c release, whilst those pre-incubated with the 17β-HSD10 decoy peptide did not. Critically, for the first time, this work demonstrated that inhibition of the 17β-HSD10-Aβ interaction may target potential disease-relevant mechanisms.
Other than the disruption of the 17β-HSD10/Aβ interaction, there is a second approach which may hold merit in treating AD: the direct modulation of 17β-HSD10 enzyme activity. In vitro experiments with neuronal-like SHSY-5Y cells exposed to the 17β-HSD10 inhibitor AG18051, showed a reduction in mitochondrial dysfunction and oxidative stress associated with the interaction between 17β-HSD10 and Aβ and protected the cells from Aβ-mediated cytotoxicity [7,8]. This proved that inhibiting 17β-HSD10 activity may also be a viable therapeutic approach for the treatment of AD.
In our previously published work [9] we discuss the rationale behind utilising analogues of the FDA-approved drugs frentizole and riluzole as inhibitors of 17β-HSD10 for potential therapeutics in AD. Briefly, many benzothiazole analogues have been shown to possess various biological activities in the central nervous system, with riluzole itself highlighted as neuroprotective. Thus, we focused on generating potent 17β-HSD10 inhibitors based on the benzothiazole scaffold, identifying several potent inhibitors (Figure 1) [9]. These compounds highlighted key structural features required for 17β-HSD10 inhibition with the 6-trifluromethoxy and 6-halogen substitution of the benzothiazole moiety and 3-chloro, 4hydroxy substitution of the phenyl moiety proving the most favourable, however, with limited solubility the compounds were not optimal for cellular evaluation. The aim of this study was not only to generate benzothiazole urea scaffolds which would show improved potency, but to also generate These compounds highlighted key structural features required for 17β-HSD10 inhibition with the 6-trifluromethoxy and 6-halogen substitution of the benzothiazole moiety and 3-chloro, 4-hydroxy substitution of the phenyl moiety proving the most favourable, however, with limited solubility the compounds were not optimal for cellular evaluation. The aim of this study was not only to generate benzothiazole urea scaffolds which would show improved potency, but to also generate compounds with improved tolerance and less cytotoxicity within our cellular assays, i.e., better pharmacokinetic parameters. To that end, four series of compounds have been synthesised targeting the key areas of benzothiazole moiety, phenyl ring and urea linker ( Figure 2).
Molecules 2019, 24, x 3 of 24 compounds with improved tolerance and less cytotoxicity within our cellular assays, i.e., better pharmacokinetic parameters. To that end, four series of compounds have been synthesised targeting the key areas of benzothiazole moiety, phenyl ring and urea linker ( Figure 2).

Structural Design and Chemical Synthesis
The first series of compounds is a continuation of our previously reported work [9]. While the benzothiazole scaffold and urea linker were kept intact, further substitution changes into the distal phenyl ring were introduced, mainly at position 3 (Table 1). Methoxy substitutions at position 6 of benzothiazole were selected and based on the comparable inhibitory activity with halogenated analogues and due to the availability of starting material and improved physical chemical properties. Benzothiazolylureas were formed using the two-step reaction process subsequently described. Initially, 6-methoxybenzo[d]thiazol-2-amine was activated with 1,1'-carbonyldiimidazole (CDI; Scheme 1a). Subsequently, intermediate 1 was reacted with corresponding substituted aniline (resp. 5-aminopyridin-2-ol for final product 8) to give final di-substituted ureas (2)(3)(4)(5)(6)(7)(8)(9)(10)(11). To obtain compounds 7, 9 and 10, N-Boc protective group was cleaved under acidic conditions (Scheme 1c) as the final step of their synthesis.

Structural Design and Chemical Synthesis
The first series of compounds is a continuation of our previously reported work [9]. While the benzothiazole scaffold and urea linker were kept intact, further substitution changes into the distal phenyl ring were introduced, mainly at position 3 (Table 1). Methoxy substitutions at position 6 of benzothiazole were selected and based on the comparable inhibitory activity with halogenated analogues and due to the availability of starting material and improved physical chemical properties. Table 1. First series of prepared compounds (2)(3)(4)(5)(6)(7)(8)(9)(10)(11).
Molecules 2019, 24, x compounds with improved tolerance and less cytotoxicity within our cellular assays, pharmacokinetic parameters. To that end, four series of compounds have been synthesised the key areas of benzothiazole moiety, phenyl ring and urea linker ( Figure 2).

Structural Design and Chemical Synthesis
The first series of compounds is a continuation of our previously reported work [9]. benzothiazole scaffold and urea linker were kept intact, further substitution changes into phenyl ring were introduced, mainly at position 3 (Table 1). Methoxy substitutions at po benzothiazole were selected and based on the comparable inhibitory activity with ha analogues and due to the availability of starting material and improved physical chemical p

11
-Cl -C a Substitution pattern replacement of whole distal phenyl ring.
-Cl -CH 2 OH a Substitution pattern replacement of whole distal phenyl ring.
Most aniline derivatives were commercially available, but in several cases the aniline intermediates had to be prepared as further described: In general, reduction of substituted nitrobenzenes into the corresponding anilines (e.g., 12) was achieved with palladium on activated carbon (Pd/C) catalysed hydrogenation (Scheme 2).
Benzothiazolylureas were formed using the two-step reaction process subsequently described. Initially, 6-methoxybenzo[d]thiazol-2-amine was activated with 1,1'-carbonyldiimidazole (CDI; Scheme 1a). Subsequently, intermediate 1 was reacted with corresponding substituted aniline (resp. 5-aminopyridin-2-ol for final product 8) to give final di-substituted ureas (2)(3)(4)(5)(6)(7)(8)(9)(10)(11). To obtain compounds 7, 9 and 10, N-Boc protective group was cleaved under acidic conditions (Scheme 1c) as the final step of their synthesis. Most aniline derivatives were commercially available, but in several cases the aniline intermediates had to be prepared as further described: In general, reduction of substituted nitrobenzenes into the corresponding anilines (e.g., 12) was achieved with palladium on activated carbon (Pd/C) catalysed hydrogenation (Scheme 2). 2-(Tert-butyl)phenol was selected as a starting material for introduction of tert-butyl group into the meta position of distal phenyl ring. Firstly, nitration was achieved with nitric acid in the presence of acetic acid as reaction solvent (Scheme 3a) to obtain intermediate 13. Secondly, the introduced nitro group was reduced to 4-amino-2-(tert-butyl)phenol (14). Initially, the reduction was attempted with Pd/C catalysed hydrogenation (Scheme 2). However, a complex mixture of decomposed starting material was received. Thus, reduction was accomplished using iron powder and ammonium chloride (Scheme 3b) to successfully obtain intermediate 14. N-Boc (de)protection had to be performed in order to obtain final compounds with the free amine group on the distal phenyl ring. Firstly, the amine group of nitroaniline was protected with ditert-butyl dicarbonate (Scheme 5a) to yield intermediates 16-18. Secondly, the nitro group was reduced with Pd/C catalysed hydrogenation to obtain intermediates 19-21 (Scheme 5b). Final N-Boc acidic deprotection was performed after the urea formation step (Scheme 1).  2-(Tert-butyl)phenol was selected as a starting material for introduction of tert-butyl group into the meta position of distal phenyl ring. Firstly, nitration was achieved with nitric acid in the presence of acetic acid as reaction solvent (Scheme 3a) to obtain intermediate 13. Secondly, the introduced nitro group was reduced to 4-amino-2-(tert-butyl)phenol (14). Initially, the reduction was attempted with Pd/C catalysed hydrogenation (Scheme 2). However, a complex mixture of decomposed starting material was received. Thus, reduction was accomplished using iron powder and ammonium chloride (Scheme 3b) to successfully obtain intermediate 14. Most aniline derivatives were commercially available, but in several cases the aniline intermediates had to be prepared as further described: In general, reduction of substituted nitrobenzenes into the corresponding anilines (e.g., 12) was achieved with palladium on activated carbon (Pd/C) catalysed hydrogenation (Scheme 2). 2-(Tert-butyl)phenol was selected as a starting material for introduction of tert-butyl group into the meta position of distal phenyl ring. Firstly, nitration was achieved with nitric acid in the presence of acetic acid as reaction solvent (Scheme 3a) to obtain intermediate 13. Secondly, the introduced nitro group was reduced to 4-amino-2-(tert-butyl)phenol (14). Initially, the reduction was attempted with Pd/C catalysed hydrogenation (Scheme 2). However, a complex mixture of decomposed starting material was received. Thus, reduction was accomplished using iron powder and ammonium chloride (Scheme 3b) to successfully obtain intermediate 14. N-Boc (de)protection had to be performed in order to obtain final compounds with the free amine group on the distal phenyl ring. Firstly, the amine group of nitroaniline was protected with ditert-butyl dicarbonate (Scheme 5a) to yield intermediates 16-18. Secondly, the nitro group was reduced with Pd/C catalysed hydrogenation to obtain intermediates 19-21 (Scheme 5b). Final N-Boc acidic deprotection was performed after the urea formation step (Scheme 1).   Most aniline derivatives were commercially available, but in several cases the aniline intermediates had to be prepared as further described: In general, reduction of substituted nitrobenzenes into the corresponding anilines (e.g., 12) was achieved with palladium on activated carbon (Pd/C) catalysed hydrogenation (Scheme 2). 2-(Tert-butyl)phenol was selected as a starting material for introduction of tert-butyl group into the meta position of distal phenyl ring. Firstly, nitration was achieved with nitric acid in the presence of acetic acid as reaction solvent (Scheme 3a) to obtain intermediate 13. Secondly, the introduced nitro group was reduced to 4-amino-2-(tert-butyl)phenol (14). Initially, the reduction was attempted with Pd/C catalysed hydrogenation (Scheme 2). However, a complex mixture of decomposed starting material was received. Thus, reduction was accomplished using iron powder and ammonium chloride (Scheme 3b) to successfully obtain intermediate 14. N-Boc (de)protection had to be performed in order to obtain final compounds with the free amine group on the distal phenyl ring. Firstly, the amine group of nitroaniline was protected with ditert-butyl dicarbonate (Scheme 5a) to yield intermediates 16-18. Secondly, the nitro group was reduced with Pd/C catalysed hydrogenation to obtain intermediates 19-21 (Scheme 5b). Final N-Boc acidic deprotection was performed after the urea formation step (Scheme 1).
Most aniline derivatives were commercially available, but in several cases the aniline intermediates had to be prepared as further described: In general, reduction of substituted nitrobenzenes into the corresponding anilines (e.g., 12) was achieved with palladium on activated carbon (Pd/C) catalysed hydrogenation (Scheme 2). 2-(Tert-butyl)phenol was selected as a starting material for introduction of tert-butyl group into the meta position of distal phenyl ring. Firstly, nitration was achieved with nitric acid in the presence of acetic acid as reaction solvent (Scheme 3a) to obtain intermediate 13. Secondly, the introduced nitro group was reduced to 4-amino-2-(tert-butyl)phenol (14). Initially, the reduction was attempted with Pd/C catalysed hydrogenation (Scheme 2). However, a complex mixture of decomposed starting material was received. Thus, reduction was accomplished using iron powder and ammonium chloride (Scheme 3b) to successfully obtain intermediate 14. N-Boc (de)protection had to be performed in order to obtain final compounds with the free amine group on the distal phenyl ring. Firstly, the amine group of nitroaniline was protected with ditert-butyl dicarbonate (Scheme 5a) to yield intermediates 16-18. Secondly, the nitro group was reduced with Pd/C catalysed hydrogenation to obtain intermediates 19-21 (Scheme 5b). Final N-Boc acidic deprotection was performed after the urea formation step (Scheme 1). Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties.  Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties. Aniline analogue with primary alcohol group in the para position (22) was generated via reduction of corresponding carboxylic acid with lithium aluminium hydride (Scheme 6). The next series was focused on selected modifications in the linker region of the scaffold, while the original distal phenyl ring substitution (3-chlorine-4-hydroxy) was selected in combination with either 6-methoxy, 6-chlorine or unsubstituted benzothiazole ring (Table 2). Additionally, to compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties.

Compound ID
compliment recently published work [8,17], dimethyl phosphonate analogues were prepared as standards (34)(35)(36) for comparison between inter-workgroup biological evaluations along with the most promising 3-chloro, 4-hydroxy substitution pattern. Finally, methylation of either one or both nitrogen atoms of the urea linker was conducted with the aim of constraining the conjugation between the two aromatic moieties.  Compounds 28 and 29 were prepared using the general procedure for urea linker synthesis in reaction with CDI (Scheme 8). In case of compound 28 synthesis, the corresponding benzylamine Compounds 28 and 29 were prepared using the general procedure for urea linker synthesis in reaction with CDI (Scheme 8). In case of compound 28 synthesis, the corresponding benzylamine intermediate (27) was first prepared from its methoxy analogue by demethylation using AlCl 3 (Scheme 8). While the originally used reaction conditions proved to be troublesome to produce the desired compounds [17], dimethyl phosphonates (34-37) were instead prepared in a two-step process. Firstly, 6-methoxybenzo[d]thiazol-2-amine and corresponding aldehyde were coupled at reflux conditions to obtain imines 30-33, which were subsequently treated with dimethyl phosphite and 1,1,3,3tetramethylguanidine to generate the final products in satisfactory yields (Scheme 9). Compound 41 was prepared in four steps (Scheme 10). The benzothiazole moiety (38) was prepared from 2-iodoaniline in reaction with methylisothiocynate and tetrabutylammonium bromide catalysed by copper (I) chloride [18]. 3-chloro-4-methoxyaniline was treated with triphosgene to give the isocyanate intermediate (39), which was then reacted with the benzothiazole moiety and the resulting methoxy derivative (40) was demethylated using AlCl3 to give compound 41.  While the originally used reaction conditions proved to be troublesome to produce the desired compounds [17], dimethyl phosphonates (34-37) were instead prepared in a two-step process. Firstly, 6-methoxybenzo[d]thiazol-2-amine and corresponding aldehyde were coupled at reflux conditions to obtain imines 30-33, which were subsequently treated with dimethyl phosphite and 1,1,3,3-tetramethylguanidine to generate the final products in satisfactory yields (Scheme 9). While the originally used reaction conditions proved to be troublesome to produce the desired compounds [17], dimethyl phosphonates (34-37) were instead prepared in a two-step process. Firstly, 6-methoxybenzo[d]thiazol-2-amine and corresponding aldehyde were coupled at reflux conditions to obtain imines 30-33, which were subsequently treated with dimethyl phosphite and 1,1,3,3tetramethylguanidine to generate the final products in satisfactory yields (Scheme 9). Compound 41 was prepared in four steps (Scheme 10). The benzothiazole moiety (38) was prepared from 2-iodoaniline in reaction with methylisothiocynate and tetrabutylammonium bromide catalysed by copper (I) chloride [18]. 3-chloro-4-methoxyaniline was treated with triphosgene to give the isocyanate intermediate (39), which was then reacted with the benzothiazole moiety and the resulting methoxy derivative (40) was demethylated using AlCl3 to give compound 41. Compound 41 was prepared in four steps (Scheme 10). The benzothiazole moiety (38) was prepared from 2-iodoaniline in reaction with methylisothiocynate and tetrabutylammonium bromide catalysed by copper (I) chloride [18]. 3-chloro-4-methoxyaniline was treated with triphosgene to give the isocyanate intermediate (39), which was then reacted with the benzothiazole moiety and the resulting methoxy derivative (40) was demethylated using AlCl 3 to give compound 41. While the originally used reaction conditions proved to be troublesome to produce the desired compounds [17], dimethyl phosphonates (34-37) were instead prepared in a two-step process. Firstly, 6-methoxybenzo[d]thiazol-2-amine and corresponding aldehyde were coupled at reflux conditions to obtain imines 30-33, which were subsequently treated with dimethyl phosphite and 1,1,3,3tetramethylguanidine to generate the final products in satisfactory yields (Scheme 9). Compound 41 was prepared in four steps (Scheme 10). The benzothiazole moiety (38) was prepared from 2-iodoaniline in reaction with methylisothiocynate and tetrabutylammonium bromide catalysed by copper (I) chloride [18]. 3-chloro-4-methoxyaniline was treated with triphosgene to give the isocyanate intermediate (39), which was then reacted with the benzothiazole moiety and the resulting methoxy derivative (40) was demethylated using AlCl3 to give compound 41.  The first step in the synthesis of products 45, 46 and 49 was to prepare corresponding N-methylated phenyl moieties in one (N-methylation with methyl iodide) or actually two steps (Odemethylation using AlCl3) as shown in Scheme 11. The third series of compounds (Table 3) focused on evaluating substitutions within the benzothiazole ring, predominantly to exploit position 6, a key area highlighted previously [9]. Our previous findings indicated that a 6-trifluromethoxy moiety and a 6-halogen moiety, led to an increased inhibitory ability towards 17β-HSD10. If not commercially available, the 6-substituted benzothiazole-2-amines were prepared from the corresponding 4-substituted anilines in reaction with potassium isocyanate and bromine (50, 51) or potassium isocyanate and tetramethylammonium dichloroiodate (52). 6-thiocyanatobenzothiazol-2amine (53) was obtained as a by-product during preparation of 6-iodobenzo[d]thiazol-2-amine (Scheme 13). The synthesis proceeded according to the general procedure using CDI to give intermediates (54-60) and final products 61-67 (Scheme 13). The third series of compounds (Table 3) focused on evaluating substitutions within the benzothiazole ring, predominantly to exploit position 6, a key area highlighted previously [9]. Our previous findings indicated that a 6-trifluromethoxy moiety and a 6-halogen moiety, led to an increased inhibitory ability towards 17β-HSD10.  The third series of compounds (Table 3) focused on evaluating substitutions within benzothiazole ring, predominantly to exploit position 6, a key area highlighted previously [9]. previous findings indicated that a 6-trifluromethoxy moiety and a 6-halogen moiety, led to increased inhibitory ability towards 17β-HSD10. If not commercially available, the 6-substituted benzothiazole-2-amines were prepared from corresponding 4-substituted anilines in reaction with potassium isocyanate and bromine (50, 51 potassium isocyanate and tetramethylammonium dichloroiodate (52). 6-thiocyanatobenzothiazo amine (53) was obtained as a by-product during preparation of 6-iodobenzo[d]thiazol-2-am (Scheme 13). The synthesis proceeded according to the general procedure using CDI to intermediates (54-60) and final products 61-67 (Scheme 13). If not commercially available, the 6-substituted benzothiazole-2-amines were prepared from the corresponding 4-substituted anilines in reaction with potassium isocyanate and Molecules 2019, 24, 2757 9 of 23 bromine (50, 51) or potassium isocyanate and tetramethylammonium dichloroiodate (52). 6-thiocyanatobenzothiazol-2-amine (53) was obtained as a by-product during preparation of 6-iodobenzo[d]thiazol-2-amine (Scheme 13). The synthesis proceeded according to the general procedure using CDI to give intermediates (54-60) and final products 61-67 (Scheme 13). In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition.  In the fourth series the benzothiazole heterocycle itself became the subject of modifi indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyc (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety w and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 1 inhibition. The general procedure for synthesis of the urea molecules described earlier in the tex 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecar intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 a synthesis procedure had to be updated due to an increase in the solubility of imidazolecar intermediates, which did not allow for their simple isolation by filtration in satisfacto In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. In the fourth series the benzothiazole heterocycle itself became the subject of modifications (as indicated in Table 4). The benzene ring was replaced with a saturated cyclohexane (71), separated (73) or completely removed (74), and the thiazole ring was replaced with an aliphatic cyclopentane (72) or replaced with an ethylene bridge (75). Further, the whole benzothiazole moiety was flipped and attached to urea via carbon in position 6 of the heterocycle. Moreover, the symmetric derivative (78) was prepared to find out whether the dimerized phenyl moiety alone is sufficient for 17β-HSD10 inhibition. The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields.
The general procedure for synthesis of the urea molecules described earlier in the text (Scheme 1) was only suitable for compounds comprising the 2-aminothiazole core (imidazolecarboxamide intermediates 68-70 and final products 71, 73 and 74). Therefore, for compounds 72, 75 and 76, the synthesis procedure had to be updated due to an increase in the solubility of imidazolecarboxamide intermediates, which did not allow for their simple isolation by filtration in satisfactory yields. Consequently, after the activation of starting compound with CDI was completed, 4-amino-2-chlorophenol was added directly to the current reaction mixture (Scheme 14).

Biochemical and Biophysical Evaluation
In order to reduce attrition rates and improve assay reproducibility we have developed a high throughput screening (HTS) pipeline ( Figure 3 [19]). In brief, compounds are screened in the recombinant 17β-HSD10 enzyme activity assay (Table 5, Figure 4). Our best previously published compounds have set the threshold of 40% remaining 17β-HSD10 activity as a minimum standard [9] and if compounds can better this threshold, they are further screened using our orthogonal counter assays, dose response assays and kinetic assessment (Table 5). Finally, if passing these criteria with favourable characteristics, the compounds progress into cellular evaluation through cytotoxicity testing and measuring 17β-HSD10 activity within cells (Table 6).

Biochemical and Biophysical Evaluation
In order to reduce attrition rates and improve assay reproducibility we have developed a high throughput screening (HTS) pipeline ( Figure 3 [19]). In brief, compounds are screened in the recombinant 17β-HSD10 enzyme activity assay (Table 5, Figure 4). Our best previously published compounds have set the threshold of 40% remaining 17β-HSD10 activity as a minimum standard [9] and if compounds can better this threshold, they are further screened using our orthogonal counter assays, dose response assays and kinetic assessment (Table 5). Finally, if passing these criteria with favourable characteristics, the compounds progress into cellular evaluation through cytotoxicity testing and measuring 17β-HSD10 activity within cells (Table 6).

Biochemical and Biophysical Evaluation
In order to reduce attrition rates and improve assay reproducibility we have developed a high throughput screening (HTS) pipeline ( Figure 3 [19]). In brief, compounds are screened in the recombinant 17β-HSD10 enzyme activity assay (Table 5, Figure 4). Our best previously published compounds have set the threshold of 40% remaining 17β-HSD10 activity as a minimum standard [9] and if compounds can better this threshold, they are further screened using our orthogonal counter assays, dose response assays and kinetic assessment (Table 5). Finally, if passing these criteria with favourable characteristics, the compounds progress into cellular evaluation through cytotoxicity testing and measuring 17β-HSD10 activity within cells (Table 6). The symmetric 1,3-bis(3-chloro-4-hydroxyphenyl)urea (78) was prepared in two steps (Scheme 15). First, 3-chloro-4-methoxyaniline was treated with CDI to give 1,3-bis(3-chloro-4methoxyphenyl)urea (77), which was then O-demethylated in reaction with AlCl3.

Biochemical and Biophysical Evaluation
In order to reduce attrition rates and improve assay reproducibility we have developed a high throughput screening (HTS) pipeline ( Figure 3 [19]). In brief, compounds are screened in the recombinant 17β-HSD10 enzyme activity assay (Table 5, Figure 4). Our best previously published compounds have set the threshold of 40% remaining 17β-HSD10 activity as a minimum standard [9] and if compounds can better this threshold, they are further screened using our orthogonal counter assays, dose response assays and kinetic assessment (Table 5). Finally, if passing these criteria with favourable characteristics, the compounds progress into cellular evaluation through cytotoxicity testing and measuring 17β-HSD10 activity within cells (Table 6).

Primary Enzyme Assay Results
Full results for the primary nicotinamide adenine dinucleotide (NADH) assay screens are shown in Figure 4 including our four best previously published compounds for comparison ( Figure 1) [9].
Our first analogue series (2-11; Table 1 and Figure 4) focused on establishing how alterations to the 3 and 4 position on the distal (phenolic) ring affect inhibition potency. In this series, compounds 5 and 6 showed a huge improvement in potency with remaining 17β-HSD10 activity of 13.45% and 6.72%, respectively, at 25 µM. A significant finding from our previous work indicated that a p-hydroxy along m-chlorine substitution pattern displayed the most pronounced inhibitory activity [9], and a deviation from the 3-halogen and 4-hydroxyl pattern resulted in a dramatic decrease in 17β-HSD10 inhibition. Our findings in this series further support this, and establish that the bulkier, 3-bromo and 3-iodo substitutions are even more favourable at this position. Replacement of the phenolic hydroxyl with an amine or methylhydroxy group led to loss of activity, which further confirms the importance of the 4-positioned phenolic hydroxyl as was previously suggested [9].
The second analogue series (23-49; Table 2) focused on evaluating changes to the urea linker. Unfortunately, any variation from the urea linker resulted in a dramatic decrease of inhibitory activity ( Figure 4). This was further supported by the inclusion of compounds (34-36) previously published by Valasani et al. with the inclusion of our novel phosphonate compound (37) determining that a phosphonate linker did not increase 17β-HSD10 inhibition. Indeed, all linker variations resulted in negligible changes to 17β-HSD10 activity with the exception of compound 24. Although the secondary amide substitution (with amide nitrogen attached to benzothiazole moiety) in compound 24 appeared slightly more favourable than most, it was still not as potent as the original urea moiety and just outside of the threshold for further analysis at 47.19% 17β-HSD10 activity remaining at 25 µM (Figure 4). Mono and dimethylation of the urea linker to enforce sp 3 , rather than sp 2 character, showed a clear detrimental effect on the activity in compound 45-49.
The third series of compounds (61-67) focused on evaluating substitutions within the benzothiazole ring, predominantly to exploit position 6, a key area highlighted previously (Hroch et al. 2016). Our previous findings indicated that a 6-trifluromethoxy moiety and a 6-halogen moiety led to an increased inhibitory ability towards 17β-HSD10. This series appear to be the most promising displaying the largest decrease in 17β-HSD10 activity (indicated in Figure 4), in particular when bulky substitutions at position 6 were applied. This is particularly apparent in compounds 61 and 62 whereby, as the functional group size increases at position 6, 17β-HSD10 activity decreases with 6-t-butyl (62) inhibiting 17β-HSD10 by 78.36% and the 6-isopropyl substitution (61) inhibiting 17β-HSD10 by 77.37% at 25 µM.
In order to validate the importance of the benzothiazol-2-yl moiety, several structural analogues were prepared and evaluated in the fourth series (71-78; Table 4). Indeed, any deviation from the benzothiazol-2-yl moiety resulted in decreased biological activity and only the symmetrical compound (78) showed real inhibitory activity with 17β-HSD10 activity reduced to 40.69% at 25 µM ( Figure 4).

Orthogonal Counter Screens
During our enzymatic assay development, it was noted that the assay was susceptible to false positives through redox cycling and aggregation mechanism [19], therefore, two orthogonal counter screens have been implemented to validate the primary screen results. The addition of the detergent Triton X-100 to the assay buffer prevents the hydrophobic interactions required for aggregation, by which a reduction in inhibition in the presence of Triton X-100 indicates the undesirable inhibitory mode of action whereby, the compound could be potentially inhibiting the enzyme through the indirect sequestration of the protein. Results (Table 5) identify compound 63 as a potential aggregator as it showed a 61% increase in 17β-HSD10 activity in the presence of Triton-X100. Compounds 65 and 67 also showed some reduction in activity but this is much less pronounced. Given that these three compounds were part of a small series we decided to advance them into the next step of screening.
With the inclusion of the strong reducing agent dithiothreitol (DTT) in the assay buffer, compounds can appear as a false positive as DTT is capable of generating H 2 O 2 causing indirect enzyme inhibition and assay interference. The fluorescence change during the reduction of resazurin to resorufin can be measured as an indication of any redox cycling compounds. The results indicate that none of the compounds appear to be acting via this undesirable mode of action (Table 5).

Dose Response and Kinetic Evaluation
Our most promising compounds demonstrate reasonable IC 50 values of around 1-2 µM (Table 5 and graphs in Supplementary data). Significantly, these compounds all display a mixed mechanism of inhibition with respect to both substrate acetoacetyl-Coenzyme A and co-factor NADH whereby at low concentrations they appear to act in a competitive manner, but at high concentrations they are inhibiting in other sites (Table 5, Hanes-Woolf plots in Supplementary data). This is favourable over other previously published work [7,20] as the AG18051 compound irreversibly inhibits 17β-HSD10, forming a covalent adduct with NADH at the active site, thus introducing a potential specificity issue.

Cellular Screening
Compound toxicity and potency was also assessed using HEK293 mts17β-HSD10 cells; results are shown in Table 6. Our fluorogenic probe, (−)-CHANA, a 17β-HSD10 substrate [21] was used to calculate cellular IC 50 values (graphs in Supplementary data) with the exception of compounds 64, 66, 67 and 78, which were precipitating within the assay media and not able to effectively penetrate into cells. Compounds 61, 62 and 63 proved to be the most potent in our cellular assay with IC 50 values of 7.88, 3.77 and 2.29 µM, respectively.
Lactate dehydrogenase (LDH) is a colorimetric assay routinely used to quantitatively measure LDH released into the media from damaged cells as a biomarker for cellular cytotoxicity and cytolysis. HEK293 mts17β-HSD10 cells were treated with compound (100 and 25 µM) for 24 h before measurements were taken. At 25 µM compounds showed around 10-30% cytotoxicity, however, this concentration is substantially higher than the measured IC 50 values and as such is not a cause for concern. Compound 65 showed a remarkably higher IC 50 value in the (−)-CHANA assay which suggests that the uptake of compound by cells and 17β-HSD10 target engagement is not as favourable as others.

General Chemistry
All reagents and solvents were purchased from commercial sources (Sigma Aldrich, Prague, Czech Republic; Activate Scientific, Prien, Germany; Alfa Aesar, Kandel, Germany; Merck, Darmstadt, Germany; Penta Chemicals, Prague, Czech Republic and VWR, Stribrna Skalice, Czech Republic) and they were used without any further purification. Low boiling point (≥90% 40-60 • C) petroleum ether (PE) was used if not stated otherwise.
Thin-layer chromatography (TLC) for reaction monitoring was performed on Merck aluminium sheets, silica gel 60 F 254 (Darmstadt, Germany). Visualisation was performed either via UV (254 nm) or appropriate stain reagent solutions (alternatively in combination of both). Preparative column chromatography was performed on silica gel 60 (70-230 mesh, 63-200 µm, 60 Å pore size). Melting points were determined on a Stuart SMP30 melting point apparatus and are uncorrected.
Nuclear magnetic resonance (NMR) spectra were acquired at 500/126/202 MHz ( 1 H, 13 C and 31 P) on a Varian S500 spectrometer or at 300/75 MHz ( 1 H and 13 C) on a Varian Gemini 300 spectrometer (both produced by Palo Alto, CA, USA). Chemical shifts δ are given in ppm and referenced to the signal center of solvent peaks (DMSO-d 6 : δ 2.50 ppm and 39.52 ppm for 1 H and 13 C, respectively; Chloroform-d: δ 7.26 ppm and 77.16 ppm for 1 H and 13 C, respectively), thus indirectly correlated to TMS standard (δ 0 ppm). Chemical shifts δ for 31 P are given in ppm and referenced to the phosphoric acid standard (δ 0 ppm). Coupling constants are expressed in Hz.
High-resolution mass spectra (HRMS) were recorded by coupled LC-MS system consisting of Dionex UltiMate 3000 analytical LC system and Q Exactive Plus hybrid quadrupole-orbitrap spectrometer (both produced by ThermoFisher Scientific, Bremen, Germany). As an ion-source, heated electro-spray ionization (HESI) was utilised (setting: 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). Positive ions were monitored in the range of 100-1500 m/z with the resolution set to 140,000. Obtained mass spectra were processed in Xcalibur 3.0.63 software (ThermoFisher Scientific, Bremen, Germany).
Further synthetic information can be found in the Supplementary material.

Final Products Characterization
The purification method is specified here only when altered from the generally used method described in Supplementary information.  13 13 13