Synthesis and biological characterization of a 17β hydroxysteroid dehydrogenase type 10 (17β-HSD10) inhibitor

Abstract

Alzheimer's disease (AD) is estimated to affect over 55 million people across the world. Small molecule treatment options are limited to symptom management with no impact on disease progression. The need for new protein targets and small molecule hit compounds is unmet and urgent. Hydroxysteroid 17-β dehydrogenase type 10 (17β-HSD10) is a mitochondrial enzyme known to bind amyloid beta, a hallmark of AD, and potentiate its toxicity to neurons. Identification of small molecules capable of interacting with 17β-HSD10 may drive drug discovery efforts for AD. The screening compound BCC0100281 (1), was previously identified as an inhibitor of 17β-HSD10. Herein we report the first synthetic access to the hit compound following a convergent pathway starting from simple heterocyclic building blocks. The compound was found to be toxic to ‘neuron-like’ cells, specifically those of neuroblastoma origin, providing a potential hit compound for cancer drug discovery, wherein the protein is known to be overexpressed. However, assay of synthetic intermediates identified novel scaffolds with effect to rescue amyloid beta-induced cytotoxicity, showcasing the power of organic synthesis and medicinal chemistry to optimize hit compounds.

The mitochondrial enzyme 17β-HSD10 has been implicated in Alzheimer's disease. Modulating this protein using a small molecule has the potential to offer protective effect.

Hydroxysteroid 17β dehydrogenase type 10 (17β-HSD10) is a mitochondrial enzyme implicated in multiple disease types including, but not limited to; Alzheimer's disease (AD) and cancer. The enzyme is a multifunctional nicotinamide adenine dinucleotide (NAD⁺)-dependent homotetramer protein complex with many reported roles including the degradation of isoleucine, metabolism of neurosteroids and the processing of mitochondrial tRNA transcripts.^1–4 The crystal structure of 17β-HSD10 indicates a flexible active center that is directed outwards resulting in its wide substrate specificity encompassing steroids, simple alcohols, fatty acids and amino acid metabolites.^5,6

Amyloid beta (Aβ), the accumulation of which is one of the three pathological hallmarks of AD, has been shown to bind 17β-HSD10, leading to potentiation of Aβ toxicity.^7–10 The formation of a protein–protein interaction between Aβ and 17β-HSD10 disrupts normal enzymatic function, leading to increased toxicity and mitochondrial dysfunction that is not experienced in the presence of either protein without the other.¹¹ The observed toxicity can be attributed to the dysregulation of pathways that under normal conditions provide protective effects within the brain,^12–18 resulting in increased reactive oxygen species generation and mitochondrial dysfunction.^19–22 Multiple substrates, including acetoacetyl-CoA and estradiol, which are essential for neuronal survival, are modulated by 17β-HSD10, and dysfunction leads to aberrant signaling cascades. Additionally, 17β-HSD10 acts as a key energy regulator and plays a role in regulating brain metabolic homeostasis.^12,23–25 Levels of 17β-HSD10 have been found to be elevated in regions of the brain affected by AD pathology.²⁶ The 17β-HSD10–Aβ protein–protein interaction in both AD patients and mouse models has been linked to disturbances in the balance of estradiol, peroxiredoxin-2, and endophilin-1 in the brain. Estradiol exhibits neuroprotective effects within the brain through many pathways.^{12,13,15,16,27–30} The reduction of estradiol levels, due to the 17β-HSD10–Aβ interaction, increases brain macrophage reactivity, Aβ accumulation, and increases the risk of hypermetabolism in the brain. Peroxiredoxin-2, an antioxidant, is found to be elevated in AD.³¹ Despite the peroxiredoxin-2 expression being elevated, as a biological means to protect against peroxides in the brain, the increased calcium in the cytosol, due to the reduced estradiol levels, results in the phosphorylation of CDK5 leading to peroxiredoxin-2 deactivation.³² Similarly, endophilin-1 is responsible for proper synaptic vesicle endocytosis and, its levels become elevated in AD due to the 17β-HSD10–Aβ interaction. The increased endophilin-1 levels result in neuroinflammation, another pathological hallmark of AD, the activation of c-Jun N-terminal kinase and, subsequently, apoptosis.^33–35 The inactivation of peroxiredoxin-2, upregulation of endophilin-1, and reduced estradiol levels are all linked to the dysregulation of 17β-HSD10 due to its it's interaction with Aβ.⁵ Therefore, inhibition of the Aβ–17β-HSD10 protein–protein interaction through a small molecule intervention or direct modulation of the enzyme has the potential to be exploited as a therapy for AD.^36–39

Testosterone and 5α-dihydrotestosterone (DHT) activate the androgen receptor (AR), which contributes to the survival and growth of prostate cancer.⁴⁰ It is suggested that 17β-HSD10 may play a role in the survival of prostate cancer cells in those patients who are undergoing androgen deprivation therapy or are post-castration. This is due to 17β-HSD10 possessing the ability to catalyze the reaction of 5α-androstane-3α,17β-diol into DHT, thus activating the AR in the absence of testosterone.^41,42 Therefore, the elevated levels of 17β-HSD10 found in prostate cancer correlate with increased DHT levels and proliferation.^42–45 Additionally the upregulation of the 17β-HSD10 gene correlates to poor treatment response in osteosarcoma.⁴³ Thus, the targeted inhibition of 17β-HSD10 could provide new treatment options for prostate and other cancers where the enzyme is upregulated. Hence, identifying diverse molecules with the ability to modulate the activity of 17β-HSD10 is of interest to address multiple areas of unmet medical need. For a more detailed review of ways in which 17β-HSD10 is linked to health and disease, we refer the reader to an excellent review.⁴⁶

The screening library constituent BCC0100281 1 (Fig. 1), was one of 16 small molecules identified as inhibitors of 17β-HSD10 in a high-throughput screen of 6759 compounds.⁴⁷ Due to the favorable physiochemical properties and moderate inhibitory activity (pIC₅₀ 4.6), we identified 1 as a compound of interest with ‘drug-like’ potential. To confirm the activity of 1 and access analogues of it for medicinal chemistry optimization studies, synthetic access was required. Surprisingly, no synthetic route to obtain 1 or similar analogues has been disclosed in the literature to the best of our knowledge. Herein we report the total synthesis of BCC0100281 (1) suitable for analogue generation, conduct in silico and biochemical binding studies to confirm 17β-HSD10 target engagement, and evaluate the biological activity of selected intermediate derivatives as chemical probes.

Fig. 1. Structure of BCC0100281 (1).

Retrosynthetic analysis (Scheme 1) identified that 1 could be obtained through two key intermediates, a suitably functionalized guanidine and an α,β-unsaturated ketone. We proposed that these intermediates could be obtained from commercially available building blocks in two linear sequences which converge finally to afford 1 through a condensation reaction.

Scheme 1. Retrosynthetic analysis of 1.

To synthesize the first required intermediate, guanidine 8 was obtained in six steps from alcohol 2 (Scheme 2). Alcohol 2 can be accessed in two steps from 2,3-dihydrofuran^48,49 and is commercially available. Chlorination of 2 was accomplished in quantitative yield upon exposure to thionyl chloride to afford 3 which was subsequently treated with biphasic conditions to facilitate tert-butyloxycarbonyl (Boc) protection of the pyrazole nitrogen resulting in chloride 4 in 44% yield. Chloride 4 proved an effective electrophilic substrate for the alkylation of methylbenzylamine to afford tertiary amine 5b with 11% yield. The low yield of 5b was attributed to a complex mixture of side products obtained from the reaction (Table 1), with Boc deprotection occurring in situ to afford both 3 and 5a alongside elimination product 9. Reaction condition screening revealed that acetonitrile was required as solvent, and a temperature of 90 °C most effectively suppressed formation of the elimination product 9 (7%) whilst obtaining 5a with a 62% yield, removing the need for the Boc deprotection step. Furthermore, chloride 3 could be isolated upon purification and progressed back through the synthetic scheme, while Boc product 5b (9%) can undergo deprotection to give 5a quantitatively, affording an overall yield of 5b based on recovered starting material of 71%.

Scheme 2. Synthesis of intermediate 8.

Reaction trials for the synthesis of 5a/b.

Entry	Base	Solvent	Temp. °C	4	5a	5b	3	9
1	NaH	THF	rt	aNo reaction
2	K2CO4	DMF	70	bNo reaction
3	NaHCO3	DMF	70	bNo reaction
c4	NaHCO3	MeCN	70	43	11	34	9	3
c5	NaHCO3	MeCN	90	6	54	12	21	7
c6	NaHCO3	MeCN	110	0	55	0	15	30

^aReaction analysis (TLC/HPLC) showed only starting material.

^bReaction analysis showed no starting material remaining, NMR of isolated material showed no evidence of compounds 3, 4, 5a, 5b, 9, 10.

^cRatio quantification of 5a, 5b, 3, and 9 achieved by HPLC. Isolated yields of 5a, and 5b after purification are provided in Scheme 2 with difference in HPLC ratio and yield accounted for due to variation in compound UV absorption.

Due to the Boc deprotection occurring in situ, the reaction was attempted with the unprotected pyrazole 3 but this proved unsuccessful. Alternative pyrazole-protecting groups (Table 2) were introduced to improve ease of purification. The benzyl, p-methoxybenzyl, and tosyl protecting groups were selected given their increased stability to withstand the harsher reaction conditions, including high temperatures. The synthesis of 5c–e was achieved following a longer route through reaction with potassium phthalimide followed by deprotection, reductive amination, and methylation (ESI†). This method proved successful in obtaining 5c–e with no premature elimination or deprotection. However, this strategy proved futile as deprotection of all three pyrazole-protecting groups under multiple conditions was unsuccessful (Table 2). Treating the benzyl (5c) and p-methoxybenzyl (5d) pyrazole compounds with hydrogen and palladium on carbon afforded only deprotection of the aliphatic amine (to afford 6c and 6d respectively). Increased temperature up to 90 °C and pressure up to 7 bar (entries 1–5) proved ineffective for pyrazole deprotection. Alternative conditions for the PMB deprotection (entries 6–9) were attempted, including those used for the deprotection of PMB-ethers. Despite acidic and oxidative conditions, refluxing in TFA or reaction with DDQ, (entries 7 & 8) showing evidence of product by MS, isolation was unsuccessful, suggesting only trace amounts of desired product 5a being formed. Deprotection of tosyl pyrazole 5e also proved unsuccessful under both acidic and basic conditions (entries 10 & 11).

Alternative protecting groups and deprotection conditions.

Entry	Substrate	Conditions	Product
1	5c	H2 (1 bar), Pd/C (20%), MeOH (2 mL), rt	c6c
2	5c	H2 (1 bar), Pd/C (20%), MeOH (2 mL), 90 °C	c6c
3	5c	H2 (7 bar), Pd/C (20%), MeOH (2 mL), rt	c6c
4	5c	H2 (7 bar), Pd/C (20%), MeOH (2 mL), 90 °C	c6c
5	5d	H2 (7 bar), Pd/C (20%), MeOH (2 mL), 90 °C	c6d
6	5d	TFA (10 equiv.), Pd/C (20%), MeOH (2 mL), 70 °C	d6d
7	5d	TFA (47 equiv. or 1 mL per 100 mg), 100 °C	a6a
8	5d	DDQ (1.2 equiv.), DCM : H2O (20 : 4) (5.25 mL), rt	a6a
9	5d	Ceric ammonium nitrate (4 equiv.), MeCN : H2O (3 : 1) (4 mL), rt	e6d
10b	5e	AcOH : H2SO4 (1 : 1) (5 mL)	No product
11b	5e	NaOH (4 M) : MeOH : THF 2 : 1 : 4 (3.5 mL) 80 °C	No product

^aProduct detected by HRMS however no product obtained from purification.

^bNo starting material remaining after 4 h, no product mass or peak observed.

^cProduct obtained quantitatively after filtration of reaction through celite and concentration in vacuo.

^dProduct obtained with 93% yield after neutralization by addition of Et₃N (10 equiv.), followed by filtration of reaction through celite and concentration in vacuo.

^eReaction on 100 μM scale, product identification achieved by HRMS.

Hydrogenation of benzylamine 5a using palladium on carbon afforded secondary amine 6a in quantitative yield. Reaction of 6a with bis-Boc thiourea in the presence of mercury(ii) chloride and triethylamine afforded bis-Boc-protected guanidine 7 which subsequently underwent Boc deprotection when exposed to trifluoroacetic acid, as described by Kim et al.⁵⁰ to afford desired intermediate 8.

The second intermediate, α,β-unsaturated ketone 12, required for the synthesis of 1, was obtained in two steps from commercially available isoxazole 10 (Scheme 3). First described by Brunelle,⁵¹ the lithiation of 10 with lithium diisopropylamide (LDA) at low temperature (−78 °C) occurs regiospecifically at the C5-methyl group. Exploiting this regioselectivity and following a similar procedure reported by Del Bello et al.,⁵²11 was obtained in 51% yield initially. However, we identified that reversing the addition of reagents, i.e. adding a solution of 10 in THF to pre-cooled LDA, improved this yield to 92%. The subsequent reaction of ketone 11 and N,N-dimethylformamide dimethyl acetal (DMF-DMA) occurs through refluxing in toluene overnight to afford desired intermediate 12.

Scheme 3. Synthesis of intermediate 12.

The final step in the synthesis of 1 combines intermediates 8 and 12 (Scheme 4) in the presence of potassium carbonate to form the 2,4,5-substituted pyrimidine ring of 1. The cyclization step was optimized using a model reaction composed of commercially available guanidine and urea-like building blocks (Table 3) prior to attempt with 8. Alteration of the base and solvent from sodium ethoxide and methanol to their non-nucleophilic counterparts' potassium carbonate and dioxane (additionally acetonitrile can also be used as the solvent) improved the overall yield of this cyclization up to 90% after purification, and in some cases, no purification is required. However, reaction of 8 and 12 afforded a less than desirable isolated yield of 14%. Unlike most of the model reactions conducted (Table 3) this reaction produced multiple impurities and was sluggish in nature, a situation also experienced with 13e which afforded a similar yield.

Scheme 4. Synthesis of 1 from intermediates 8 and 12.

Optimization of conditions and substrates for the cyclization of guanidine and urea building blocks with 12.

Entry	Substrate	Base	Solvent	Temperature	Product	Yield
1	a13a	NaOMe	MeOH	65 °C	14a	32%
2	13b	K2CO3	Dioxane	100 °C	14b	11%
3	13c	K2CO3	Dioxane	100 °C	14a	b100%
4	13d	K2CO3	MeCN	100 °C	14d	72%
5	13e	K2CO3	Dioxane	100 °C	14e	17%
6	13e	K2CO3	MeCN	100 °C	14e	22%

^a13a (2.4 equiv.), base (3.4 equiv.).

^bNo purification of product required.

With a synthetic route suitable for access to hit compound 1 and generation of derivatives in hand, we next determined the ability of 1 to bind to the 17β-HSD10 protein target. In parallel to our synthetic efforts, we conducted in silico binding studies of 1 using the published crystal structure of the NADPH 17β-HSD10 homotetramer in complex with the inhibitor AG18051 (PDB ID: 1U7T).⁵³ Employing SeeSAR 12.1 software (BioSolveIT), the AG18051-NADPH ligand was removed and the binding site of 61 amino acids defined. Compound 1 was prepared and docked using HYDE to determine predicted binding interactions (Fig. 2). The pyrazole ring was predicted to form hydrogen bonds with amino acid residues valine 65 and alanine 63, the pyrimidine nitrogen distal to the cyclopropyl ring was predicted to from a hydrogen bond with glycine 93. The oxygen of the isoxazole ring was predicted to form a hydrogen bond with a bridging water molecule to amino acid residues glycine 23 and glycine 17 while the nitrogen is predicted to form a hydrogen bond with leucine 22. Thus, in silico modelling supports compound 1 binding 17β-HSD10.

Fig. 2. Docking of 1 into the X-ray structure of 17β-HSD10 (PDB ID: 1U7T). A) Structure of 1 in gold, green dotted lines indicated predicted strong hydrogen bonds: B) overlay of NADH co-factor in magenta.

To confirm binding of 1 to 17β-HSD10 we conducted a protein thermal shift assay. Upon titration of 1 to 17β-HSD10, the melting temperature of the protein decreased (Table 4). This change in melting temperature is concentration-dependent with increasing concentration of 1, ranging from 42.29 °C with 5 μM treatment of 1, a change of −0.37 °C from the native protein, to 40.46 °C with 100 μM treatment of 1, a change of −2.2 °C from the native protein (ESI† Fig. S1A). Solubility of 1 hindered testing at higher concentrations. The reduction of melting temperature suggests that 1–17β-HSD10 complex formation decreases the thermal stability of the protein and confirms binding. This is consistent with published studies where structurally-dynamic loops forming part of the 17β-HSD10 active site undergo a conformational change upon substrate binding.^53–55 These data suggest that binding of compound 1 to 17β-HSD10 may stimulate structural changes resulting in thermal destabilization and that the conformation of 17β-HSD10 in complex with 1 is less stable than the apo-protein conformation. As a control, the same experiment was performed using a compound known to bind 17β-HSD10. Frentizole, a benzothiazole urea/thioflavin T derivative which inhibits Aβ–17β-HSD10 activity with an IC₅₀ of approximately 200 μM, was chosen as the control ligand.⁵⁶ As expected, frentizole altered the thermal stability of 17β-HSD10 showing a dose-dependent increase as high as 5 °C at 100 mM concentration (ESI† Fig. S1B), which is consistent with protein-inhibitor binding. In contrast to the negative thermal shift induced by compound 1, frentizole is unlikely to elicit a conformational change in 17β-HSD10.

Melting temperature (T_m) of 17β-HSD10 in the presence of indicated concentration of 1 or frentizole, derived from thermal denaturation curves.

	Conc. (μM)	T m		ΔTm mean
		Mean	Std. error
DMSO		42.78	0.27	—
Frentizole	5	44.28	0.25	1.49
	10	44.93	0.26	2.15
	25	46.42	0.21	3.64
	50	46.61	0.05	3.83
	100	47.78	0.36	4.99
1	5	42.32	0.10	−0.42
	10	42.33	0.12	−0.41
	25	41.97	0.19	−0.77
	50	41.67	0.15	−1.07
	100	40.64	0.28	−2.1

With high confidence that 1 is interacting with 17β-HSD10, we next assessed the toxicity of compound 1 in three cell lines known to overexpress 17β-HSD10.^57,58 The SH-SY5Y neuroblastoma cell line was employed due to exhibiting neurite structure and expressing immature neuronal markers and as such would serve as an in vitro model for proposed AD studies. The compound was found to exhibit toxicity in SH-SY5Y cells at concentrations greater than 10 μM (Fig. 3A), with 31%, 22%, and 10% loss of cell viability observed at concentrations of 100 μM, 50 μM, and 10 μM, respectively. We then sought to assess the compound toxicity in a range of malignant (PANC-1, LNCap, SK-N-MC, SK-N-AS, CHLA255, SK-N-BE2) and non-malignant (WPMY-1, PnT₂C₂, HPNE) cell lines (Fig. 3B) to identify if 1 exhibited specific toxicity towards the SH-SY5Y cell line or has indiscriminate toxicity. This small screen identified that 1 shows specific toxicity towards neuroblastoma (NB) cell lines SK-N-MC, SK-N-AS, CHLA255, and the proto-oncogene MYCN amplified cell line SK-N-BE2. However, minimal (≤10%) or no decrease in cell viability was observed for all other cell lines at all tested concentrations (100–0.1 μM).

Fig. 3. Assessment of cell viability on treatment of A) SH-SY5Y cell line, B) malignant and non-malignant cell lines with varying concentrations of 1. For relevant controls see ESI† Fig. S2. Data represents mean of n = 3 ± SEM, experiments run in triplicate. Unpaired t-test; 95% confidence interval; ****, P < 0.0001.

This toxicity data suggests the potential of 1 to be used as a starting point for development of novel compounds for neuroblastoma treatment. The >80% reduction observed at 100 μM for MYCN gene amplified cell line SK-N-BE2 is especially promising. Amplification of the MYCN gene in neuroblastoma patients is associated with high-risk NB and, thus, poor prognosis.^59–61 Despite extensive research into NB resulting in therapeutic advances there still remains a challenge, with high-risk NB patients having a 5-year survival rate of just 50%. Therefore, the need to identify new therapeutics is paramount to improving patient outcomes.⁶²

Modulators of 17β-HSD10 have the potential to disrupt binding of Aβ and, therefore, provide protective effects against Aβ toxicity. We assessed the ability of 1 to ameliorate Aβ_1–42 toxicity in SH-SY5Y cells. The co-treatment of cells with Aβ and 1 showed that this compound provides no amelioration of Aβ_1–42 toxic effects. (Fig. 4).

Fig. 4. A) Co-treatment of Aβ and 1 in SH-SY5Y cells after 48 h incubation. B) Pre-treatment of 1 in SH-SY5Y cells for 24 h followed by addition of Aβ (no wash step). Data represents mean of n = 3 ± SEM. Unpaired t-test; 95% confidence interval; *, p < 0.05; **, p < 0.001.

While compound 1 provides some protective effects to ameliorate Aβ-induced toxicity on pre-treatment, the pattern of cell viability closely matches the toxicity pattern seen with the compound in SH-SY5Y cells. Thus, greater protection of Aβ toxicity may be masked by this innate toxic effect. Gratifyingly, the compound is not indiscriminately toxic. We thus investigated the potential of selected synthetic derivatives, that showed no innate toxicity, to ameliorate Aβ-induced toxicity. Compounds 14b–e represent a discrete series for determination of an initial structure–activity relationship (SAR) (Fig. 5), differing by para-substituent of the central pyrimidine ring. All four compounds show amelioration of Aβ-induced toxicity at either or both 1 μM and 0.1 μM concentrations (Fig. 5A–D). These compounds appear to show an inverse dose dependency, with lower concentrations providing protection. Protein thermal shift studies on each of the compounds at a 100 μM concentration did not show a significant change in 17β-HSD10 melting temperature (ESI† Table S1 and Fig. S3). Therefore, there is the potential that although these compounds do not initially appear to be bind 17β-HSD10 at high concentrations, at lower concentrations, they may act by some alternative mechanism to provide protection against Aβ-induced toxicity as is the case for other inhibitory molecules,^63–65 and enzyme activators.⁶⁵ Interestingly, the tolerance to a range of para-functional groups suggests that a hydrogen bond donor at this position is not required. Further assessment will be required to determine the need for a hydrogen bond acceptor or polarity at this position. Such work is ongoing in our laboratory.

Fig. 5. Pre-treatment of 14b, 14c, 14d, and 14e, respectively, for A, B, C, and D in SH-SY5Y cells for 24 h, followed by addition of Aβ (no wash step). Data represents mean of n = 2 ± SEM experiments run in triplicate. Unpaired t-test; 95% confidence interval; *, p < 0.05; **, p < 0.001.

In conclusion, we have reported the first synthetic route to access screening library compound BCC0100281 (1) in nine steps with the longest linear sequence of seven steps with an overall yield of 0.7%. The route is suitable for the generation of intermediates and parent compound derivatives for SAR exploration. Furthermore, we have undertaken preliminary biological assessments to confirm the interaction of 1 with 17β-HSD10. Unfortunately, the compound proves toxic to SH-SY5Y cells. Interestingly, 1 proved to be consistently toxic across a range of neuroblastoma cell lines including the MYCN gene amplified cell line SK-N-BE2 providing an alternative utility for compounds of this chemotype for a disease requiring further therapeutic advances. Intermediates containing the isoxazole ring show significant protection of cell viability against Aβ-induced toxicity. A discrete series reveals a tolerance for various functionality which may be exploited in later compounds to control compound physicochemical properties. Thus, we report the identification of a novel chemotype for further exploration in two diseases with critical need.

Experimental

Chemistry

General chemistry procedures

All reactions were carried out in oven-dried glassware under positive nitrogen pressure unless otherwise noted. Reaction progress was monitored by thin-layer chromatography carried out on silica gel plates (2.5 cm × 7.5 cm, 200 μm thick, 60 F254) and visualized by using UV (254 nm) or by the use of phosphomolybdic acid, vanillin, or potassium permanganate solution as indicator. Flash column chromatography was performed with silica gel (40–63 μm, 60 Å) using the mobile phase indicated. Commercial grade solvents and reagents were purchased from Fisher Scientific (Houston, TX) or Sigma-Aldrich (Milwaukee, WI) and were used without further purification. Anhydrous solvents were purchased from Across Organics and stored under an atmosphere of dry nitrogen over molecular sieves.

¹H and ¹³C NMR spectra were recorded in the indicated solvent on a Bruker 400 MHz Advance III spectrometer at 400 MHz (100 MHz for ¹³C) or Bruker Advance III HD spectrometer at 500 or 600 MHz (125 or 150 MHz respectively for ¹³C). Multiplicities are indicated by s (single), d (doublet), t (triplet), m (multiplet), and br (broad). Chemical shifts (δ) are reported in parts per million (ppm) and coupling constants (J), in hertz.

High-resolution mass spectra (HRMS) were recorded with an Agilent 6230 LC/TOF spectrometer using an ESI source coupled to an Agilent Infinity 1260 system running in reverse phase with a ZORBAX RRHT Extend-C18 (80 Å, 2.1 × 50 mm, 1.8 μm) column using solvent A (water with 0.1% formic acid), solvent B (acetonitrile with 0.1% formic acid), and a flow rate of 0.4 mL min⁻¹ starting a mixture of 50% A and 50% B. Solvent B is gradually increased to 95% at 8 min.

The purity analysis of final compounds was determined ≥95% pure using an Agilent 1260 Infinity High-performance liquid chromatography (HPLC) System with diode-array detection (DAD) detector (254 nm) coupled with an Agilent 6230 liquid chromatography/time-of-flight/mass spectrometry (LC/TOF) mass spectrometer (MS) and Agilent Mass Hunter software was used for analysis. The chromatography was performed on an Eclipse plus C8 3.5μ column (4.6 × 150 mm) using solvent A (water with 0.1% formic acid), solvent B (acetonitrile with 0.1% formic acid), and a flow rate of 1.0 mL min⁻¹ for 20 min, the % of B was varied from 5 to 95 over 8 minutes.

Synthesis of 4-(2-chloroethyl)-1H-pyrazole hydrochloride (3)

Thionyl chloride (2.0 mL, 26.8 mmol, 3 equiv.) was added to 2-(1H-pyrazol-4-yl)ethan-1-ol (1.00 g, 8.92 mmol, 1 equiv.) and the reaction was heated at 70 °C in an open flask for 15 minutes after which the reaction was cooled to room temperature and concentrated in vacuo. The solid was azeotrope with methyl tert-butyl ether (5 × 20 mL) to afford 4-(2-chloroethyl)-1H-pyrazole hydrochloride 3 (1.35 g, 8.08 mmol, 91%) as an orange solid.

¹H NMR (400 MHz, DMSO-D₆): δ_H 2.91 (t, J = 7.00 Hz, 2H), 3.77 (t, J = 7.00 Hz, 2H), 7.75 (s, 2H), 8.04 (br. s, 2H).

¹³C NMR (100 MHz, DMSO-D₆): δ_C 27.70, 45.66, 117.35, 133.19.

Synthesis of tert-butyl 4-(2-chloroethyl)-1H-pyrazole-1-carboxylate (4)

To a suspension of 4-(2-chloroethyl)-1H-pyrazole (1.79 g, 13.7 mmol, 1 equiv.) in acetonitrile (20 mL) was added di-tert-butyl dicarbonate (3.80 mL, 16.4 mmol, 1.2 equiv.) followed by dropwise addition of triethylamine (7.64 mL, 54.8 mmol, 4 equiv.) the reaction was stirred at room temperature for 72 h. The reaction was quenched with water (2.0 mL) concentrated in vacuo, the residue was partitioned between ethyl acetate (20 mL) and water (10 mL), the phases were separated, and the aqueous extracted with ethyl acetate (20 mL), the combined organics were dried (MgSO₄) and concentrated in vacuo. The crude material was purified by normal phase chromatography 0–15% ethyl acetate/hexane to afford tert-butyl 4-(2-chloroethyl)-1H-pyrazole-1-carboxylate (1.39 g, 6.03 mmol, 44%) as a yellow oil.

¹H NMR (500 MHz, DMSO-D₆): δ_H 1.57 (s, 9H), 2.90 (t, J = 6.70 Hz, 2H), 3.81 (t, J = 6.80 Hz, 2H), 7.75 (s, 1H), 8.16 (s, 1H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 26.91, 27.45, 44.60, 84.62, 120.83, 129.40, 144.39, 147.03.

Synthesis of N-benzyl-N-methyl-2-(1H-pyrazol-4-yl)ethan-1-amine (5a)

Tert-Butyl 4-(2-chloroethyl)-1H-pyrazole-1-carboxylate (8.52 g, 37.0 mmol, 1 equiv.), sodium bicarbonate (9.31 g, 111 mmol, 3 equiv.), N-methyl-1-phenylmethanamine (7.15 mL, 55.4 mmol, 1.5 equiv.) and acetonitrile (80 mL) were combined and split between 4 vials (of volume 40 mL), sealed, and heated at 90 °C for 18 h. The reactions were combined and diluted with ethyl acetate (40 mL) and water (60 mL), the phases were separated, and the aqueous extracted with ethyl acetate (20 mL); the combined organics were washed with sat. brine (10 mL), dried (MgSO₄) and concentrated in vacuo. The crude material was purified by normal phase chromatography 5 : 0.5 methanol : ammonia/dichloromethane followed by reverse phase chromatography 5–95% acetonitrile/water (0.1% ammonia) to afford N-benzyl-N-methyl-2-(1H-pyrazol-4-yl)ethan-1-amine (3.98 g, 18.5 mmol, 53%) as a colorless oil.

¹H NMR (400 MHz, MeOH-D₄): δ_H 2.26 (s, 3H), 2.56–2.62 (m, 2H), 2.69–2.76 (m, 2H), 3.57 (s, 2H), 7.21–7.34 (m, 5H), 7.41 (br. s, 2H).

¹³C NMR (100 MHz, MeOH-D₄): δ_C 22.64, 42.24, 59.54, 62.97, 119.57, 128.34, 129.29, 130.60, 139.15.

Synthesis of tert-butyl 4-(2-(benzyl(methyl)amino)ethyl)-1H-pyrazole-1-carboxylate (5b)

¹H NMR (400 MHz, MeOH-D₄): δ_H 2.26 (s, 3H), 2.58–2.66 (m, 2H), 2.67–2.75 (m, 2H), 3.56 (s, 2H), 7.21–7.34 (m, 5H), 7.63 (s, 1H), 7.98 (s, J = 0.8 Hz, 1H).

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₈H₂₅N₃O₂, 316.2020; found, 316.200.

Synthesis of N-methyl-2-(1H-pyrazol-4-yl)ethan-1-amine (6a)

To a solution of 4-(2-(benzyl(methyl)amino)ethyl)-1H-pyrazol-1-ium (4.50 g, 20.8 mmol, 1 equiv.) in methanol (40 mL) was added palladium on carbon (221 mg, 2.08 mmol, 0.1 equiv.) the reaction was evacuated and backfilled with hydrogen 5 times and stirred at room temperature for 18 h. The reaction was filtered through a pad of celite and concentrated in vacuo. The crude material was purified by normal phase chromatography 4 : 0.4 methanol : ammonia/dichloromethane to afford product N-methyl-2-(1H-pyrazol-4-yl)ethan-1-amine (583 mg, 4.62 mmol, 22.2%) as a colorless oil.

¹H NMR (500 MHz, DMSO-D₆): δ_H 1.58 (s, 3H), 1.84–1.91 (m, 2H), 1.92–1.97 (m, 2H), 4.12–4.46 (br. s, 2H) 6.65 (s, 2H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 15.26, 26.37, 44.08, 109.59, 124.43.

Synthesis of di-tert-butyloxycarbonyl-1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine (7)

To a solution of 1H-pyrazol-5-amine (3.00 g, 36.1 mmol, 1.0 equiv.), 3,3-di-tert-butyloxycarbonyl-2-methylisothiourea (11.5 g, 39.7 mmol, 1.1 equiv.) and triethylamine (20.1 mL, 144 mmol, 4 equiv.) in dimethylformamide (42 mL) was added mercury(ii)chloride (10.8 g 39.7 mmol, 1.1 equiv.) the reaction was stirred at room temperature for 5 h. The reaction was filtered, and the precipitate washed with ethyl acetate (30 mL). The solution was quenched into water (500 mL) and extracted with toluene (4 × 100 mL); the combined organics were washed with water (100 mL) and sat. brine (100 mL), dried (MgSO₄), and concentrated in vacuo. The crude material was purified by normal phase chromatography 0–50% acetone/hexane, the material was repurified by normal phase chromatography 0–50% acetone/toluene to afford di-tert-butyloxycarbonyl-1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine (3.11 g, 7.80 mmol, 24%) as a white solid.

¹H NMR (500 MHz, DMSO-D₆): δ_H 1.36 (s, 9H), 1.41 (s, 9H), 2.66 (t, J = 7.85 Hz, 2H), 2.88 (s, 3H), 3.47 (t, J = 7.30 Hz, 2H), 7.42 (br. s, 2H), 9.38 (s, 1H), 12.58 (s, 1H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 21.04, 27.94, 28.06, 35.98, 76.93, 79.89, 125.31, 128.20, 128.90.

Synthesis of 1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine-tetra-2,2,2-trifluoroacetate (1/4) (8)

To a suspension of di-tert-butyloxycarbonyl-1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine (3.11 g, 7.79 mmol, 1 equiv.) in dichloromethane (20 mL) was added trifluoroacetic acid (24.0 mL, 311 mmol, 40 equiv.) the reaction was stirred at room temperature open to the atmosphere for 24 h. The reaction was concentrated in vacuo to afford 1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine-2,2,2-trifluoroacetate (5.41 g, quant.) as a brown oil.

¹H NMR (500 MHz, DMSO-D₆): δ_H 2.69 (t, J = 7.85 Hz, 2H), 2.92 (s, 3H), 3.46 (t, J = 7.50 Hz, 2H), 7.23 (br. s, 4H), 7.55 (s, 2H), 13.70 (br. s, 4H).

ESI-HRMS (m/z): [M + H]⁺ calc. for C₇H₁₃N₅, 168.1244; found, 168.1244.

Synthesis of 1-cyclopropyl-2-(3-methylisoxazol-5-yl)ethan-1-one (11)

Lithium diisopropylamide (2 M in tetrahydrofuran/hexanes) (75.0 mL, 150 mmol, 2.5 equiv.) was cooled to −75 °C after which a solution of 3,5-dimethylisoxazole (9.78 mL, 99.9 mmol, 2 equiv.) in tetrahydrofuran (100 mL) was added dropwise via a dropping funnel over 45 minutes; the reaction was stirred at −78 °C for 1 h. A solution of methyl cyclopropanecarboxylate (5.09 mL, 49.9 mmol, 1 equiv.) in tetrahydrofuran (50 mL) was added dropwise over 20 minutes. The reaction was then stirred at −78 °C under nitrogen for 2 h followed by warming to room temperature over 18 h. The reaction was quenched with sat. brine (100 mL) and diluted with water (20 mL); the phases were separated and the aqueous were extracted with ethyl acetate (2 × 100 mL). The combined organics were dried (MgSO₄) and concentrated in vacuo. The crude material was purified using a silica plug eluting with 40% ethyl acetate/hexanes to afford 1-cyclopropyl-2-(3-methylisoxazol-5-yl)ethan-1-one (7.64 g, 46.2 mmol, 93%) as a yellow oil.

¹H NMR (400 MHz, CCl₃): δ_H 0.92–0.99 (m, 2H), 1.08–1.14 (m, 2H), 1.96–2.05 (m, 1H), 2.29 (s, 3H), 3.96 (s, 2H), 6.07 (s, 1H).

¹³C NMR (100 MHz, CDCl₃): δ_C 11.58, 12.05, 20.56, 41.31, 104.35, 160.24, 165.50, 203.79.

Synthesis of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (12)

To a solution of 1-cyclopropyl-2-(3-methylisoxazol-5-yl)ethan-1-one (410 mg, 2.50 mmol, 1 equiv.) in toluene (15 mL) was added N,N-dimethylformamide dimethyl acetal (0.40 mL, 3.0 mmol, 1.2 equiv.) the reaction was heated at reflux under nitrogen for 18 h. The reaction was concentrated in vacuo, and the crude material was purified by normal phase chromatography 0–60% ethyl acetate/hexane to afford (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (431 mg, 1.96 mmol, 79%) as a yellow solid.

¹H NMR (400 MHz, CDCl₃): δ_H 0.84–0.69 (m, 2H), 1.05–1.02 (m, 2H), 1.76–1.70 (m, 1H), 2.34 (s, 3H), 2.90 (br. s, 6H), 6.10 (s, 1H), 7.78 (s, 1H).

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₂H₁₆N₂O₂, 221.1285; found, 221.1288.

Synthesis of N-(2-(1H-pyrazol-4-yl)ethyl)-4-cyclopropyl-N-methyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (1) (BCC0100281)

To a solution of 1-(2-(1H-pyrazol-4-yl)ethyl)-1-methylguanidine-2,2,2-trifluoroacetate (1/4) (5.42 g, 8.69 mmol, 1 equiv.) and potassium carbonate (8.41 g, 60.9 mmol, 7 equiv.) in dioxane (20 mL) was added (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (2.39 g, 10.9 mmol, 1.25 equiv.) and, the reaction was heated at reflux under nitrogen for 18 h. The reaction was concentrated in vacuo, and the residue was diluted with ethyl acetate (50 mL) and water (50 mL). The phases were separated and the aqueous extracted with ethyl acetate (2 × 50 mL); the combined organics were washed with sat. brine (50 mL), dried (MgSO₄) and concentrated in vacuo. The crude material was purified by normal phase chromatography 0–30% acetone/hexane. The solid obtained was triturated with dichloromethane (10 mL), acetonitrile (10 mL) and ethanol (10 mL) to afford N-(2-(1H-pyrazol-4-yl)ethyl)-4-cyclopropyl-N-methyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (144 mg, 444 μmol, 5.11%) as a pale peach solid.

¹H NMR (500 MHz, DMSO-D₆, 330 K): δ_H 1.00–1.08 (m, 2H), 1.11–1.16 (m, 2H), 2.26–2.34 (m, 4H), 2.71 (t, J = 7.5 Hz, 2H), 3.10 (s, 2H), 3.18 (s, 2H), 3.75 (t, J = 7.5 Hz, 2H), 6.58 (s, 1H), 7.33 (br. s, 1H), 7.48 (br. s, 1H), 8.44 (s, 1H), 12.45 (s, 1H).

¹³C NMR (100 MHz, DMSO-D₆, 330 K): δ_C 10.47, 10.80, 14.07, 18.31, 21.48, 34.76, 49.97, 55.98, 102.49, 109.01, 116.50, 156.54, 159.74, 160.66, 166.51, 168.17.

HPLC: retention time, 7.299 min; purity, 97.6%.

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₇H₂₀N₆O, 325.1771; found, 325.1742.

Synthesis of 5-(4-cyclopropyl-2-(methylthio)pyrimidin-5-yl)-3-methylisoxazole (14a)

To a solution of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (500 mg, 2.27 mmol, 1 equiv.) in methanol (10 mL) was added S-methylisothiourea hemisulfate (1.03 g, 5.45 mmol, 2.4 equiv.) and sodium methoxide (417 mg, 7.72 mmol, 2.4 equiv.) the reaction was heated at 65 °C under nitrogen for 18 h. The reaction was cooled to room temperature, filtered through a pad of celite, and the solution concentrated in vacuo. The material was dissolved in dichloromethane, the precipitant was filtered, the solution concentrated in vacuo and purified by normal phase chromatography 0–20% ethyl acetate/hexane to afford 5-(4-cyclopropyl-2-(methylthio)pyrimidin-5-yl)-3-methylisoxazole (180 mg, 0.73 mmol, 32%) as a cream solid.

¹H NMR (500 MHz, DMSO-D₆): δ_H 1.31–1.42 (m, 4H), 2.58 (m, 1H), 2.68 (s, 6H), 7.06 (s, 1H), 8.88 (s, 1H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 11.48, 12.46, 14.03, 14.95, 105.62, 116.60, 156.16, 160.73, 165.24, 168.72, 172.70.

Synthesis of 5-(4-cyclopropyl-2-methoxypyrimidin-5-yl)-3-methylisoxazole (14b)

To a solution of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (1.49 g, 6.76 mmol, 1 equiv.) and methyl carbamimidate hydrochloride (0.82 g, 7.44 mmol, 1.1 equiv.) in dioxane (10 mL) was added potassium carbonate (2.80 g, 20.28 mmol, 3 equiv.). The reaction was heated at 100 °C for 48 h. The reaction was concentrated in vacuo, and the residue partitioned between ethyl acetate (50 mL) and water (30 mL); the phases were separated, and the aqueous extracted with ethyl acetate (2 × 30 mL). The combined organics were washed with sat. brine (20 mL), dried (MgSO₄) and concentrated in vacuo. The crude material was purified by normal phase chromatography 0–50% ethyl acetate/hexane and re-purified by normal phase chromatography 0–15% acetone/hexane to afford 5-(4-cyclopropyl-2-methoxypyrimidin-5-yl)-3-methylisoxazole (170 mg, 0.73 mmol, 11%) as a colorless solid.

¹H NMR (500 MHz, DMSO-D₆): δ_H 1.15–1.21 (m, 4H), 2.31 (s, 3H), 2.35–2.40 (m, 1H), 3.91 (s, 3H), 6.81 (s, 1H), 8.69 (s, 1H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 11.01, 11.89, 14.49, 54.67, 104.63, 114.95, 158.45, 160.18, 165.02, 165.13, 171.16.

HPLC: retention time, 7.623 min; purity, 96.2%.

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₂H₁₃N₃O₂, 232.1081; found, 232.1769.

Synthesis of 4-cyclopropyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (14c)

To a suspension of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (200 mg, 0.91 mmol, 1 equiv.) and guanidine hydrochloride (104 mg, 1.09 mmol, 1.2 equiv.) in dioxane (10 mL) was added potassium carbonate (376 mg, 2.72 mmol, 3 equiv.) the reaction was heated at 100 °C under nitrogen for 18 h. The reaction was cooled to room temperature, filtered and the solution concentrated in vacuo. The residue was partitioned between ethyl acetate (30 mL) and water (10 mL); the phases were separated, the organic dried (MgSO₄) and concentrated in vacuo to afford 4-cyclopropyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (198 mg, 0.92 mmol, quant.) as a pale yellow solid.

¹H NMR (500 MHz, DMSO-D₆): δ_H 0.95–1.09 (m, 4H), 2.20–2.29 (m, 4H), 3.33 (s, 3H), 6.59 (s, 1H), 6.94 (S, 2H), 8.34 (S, 1H).

¹³C NMR (125 MHz, DMSO-D₆): δ_C 10.90, 11.50, 14.53, 103.15, 110.42, 157.65, 160.33, 163.78, 167.13, 169.22.

HPLC: retention time, 6.115 min; purity, 98.9%.

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₁H₁₂N₄O, 217.1084; found, 217.1750.

Synthesis of 4-cyclopropyl-N-methyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (14d)

To a suspension of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (2.92 g, 13.3 mmol, 1 equiv.) and N-methyl-guanidine (1.45 g, 13.3 mmol, 1 equiv.) in acetonitrile (40 mL) was added potassium carbonate (5.50 g, 39.8 mmol, 3 equiv.) the reaction was heated at 100 °C under nitrogen for 18 h. The reaction was filtered, and the solution was concentrated in vacuo. The crude residue was purified by normal phase chromatography 0–50% ethyl acetate/hexane to afford 4-cyclopropyl-N-methyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-amine (1.23 g, 5.34 mmol, 40%) as a yellow solid.

¹H NMR (400 MHz, MeOH-D₄): δ_H 0.98–1.08 (m, 2H), 1.16–1.27 (m, 2H), 2.24–2.36 (m, 4H), 2.91 (s, 3H), 6.46 (s, 1H), 8.30 (s, 1H).

¹³C NMR (100 MHz, MeOH-D₄): δ_C 10.66, 11.08, 14.26, 27.72, 102.62, 109.51, 109.90, 156.98, 160.01, 162.26, 166.77, 168.62.

HPLC: retention time, 7.115 min; purity, 98.1%.

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₂H₁₄N₄O, 231.1240; found, 231.1929.

Synthesis of tert-butyl (4-cyclopropyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-yl)carbamate (14e)

To a suspension of (Z)-1-cyclopropyl-3-(dimethylamino)-2-(3-methylisoxazol-5-yl)prop-2-en-1-one (300 mg, 1.36 mmol, 1 equiv.) and N-boc-guanidine (260 mg, 1.63 mmol, 1.2 equiv.) in dioxane (10 mL) was added potassium carbonate (565 mg, 4.09 mmol, 3 equiv.) the reaction was heated at 100 °C under nitrogen for 18 h. The reaction was concentrated in vacuo, and the residue partitioned between ethyl acetate (20 mL) and water (40 mL); the phases were separated, and the aqueous extracted with ethyl acetate (10 mL) and dichloromethane (10 mL). The combined organics were dried (MgSO₄) and concentrated in vacuo. The crude material was dissolved in dichloromethane, the solid removed by filtration, and the solution purified by chromatography 0–40% ethyl acetate/hexanes to afford tert-butyl (4-cyclopropyl-5-(3-methylisoxazol-5-yl)pyrimidin-2-yl)carbamate (72 mg, 0.23 mmol, 17%) as a white solid.

¹H NMR (400 MHz, DMSO-D₆): δ_H 1.10–1.21 (m, 4H), 1.46 (s, 9H), 2.31 (s, 3H), 2.33–2.39 (m, 1H), 6.81 (s, 1H), 8.66 (s, 1H), 10.10 (s, 1H).

¹³C NMR (100 MHz, DMSO-D₆): δ_C 10.98, 11.49, 14.50, 27.90, 79.73, 104.46, 114.93, 150.66, 156.79, 158.06, 160.12, 165.14, 170.27.

HPLC: retention time, 8.113 min; purity, 97.0%.

ESI-HRMS (m/z): [M + H]⁺ calc. for C₁₆H₂₀N₄O₃, 317.1608; found, 317.2413.

Biology

Cell culture

The cell line SH-SY5Y (ATCC catalog number CRL-2266) was cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 (fisher scientific, catalog number 10131035) supplemented with 10% fetal bovine serum (FBS) (corning, catalog number 35-01-CV) and 1% penicillin–streptomycin (Pen–Strep) (Corning, catalog number 30-002-CI). The cell lines LNCap (ATCC catalog number CRL-1740), HPNE (gifted from Dr. Mohs Lab of UNMC, 11-June-24) and PnT₂C₂ (Sigma Aldrich, catalog number 95012613-DNA-5UG) were cultured in Roswell Park Memorial Institute (RPMI) 1640 Medium (Corning, catalog number MT10041CV) supplemented with 10% FBS (corning, catalog number 35-01-CV) and 1% Pen–Strep (Corning, catalog number 30-002-CI). The cell lines PANC-1 (ATCC catalog number CRL-1469) and WPMY1 (ATCC catalog number CRL-2854) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS (corning, catalog number 35-01-CV) and 1% Pen–Strep (Corning, catalog number 30-002-CI). The cell line CHLA-255 (gifted from Dr. Challagundla, 7-May-24) was cultured in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% FBS, 1% Pen–Strep and 1% MEM non-essential amino acids (MEM-NEAA) (Gibco, cat# 11140050). The cell line SK-N-AS (ATCC catalog number CRL-2137) was cultured in DMEM supplemented with 10% FBS, 1% Pen–Strep and 1% MEM-NEAA. The cell like SK-N-BE2 (ATCC catalog number CRL-2271) was cultured in RPMI-1640 supplemented with 10% FBS, 1% Pen–Strep and 1% MEM-NEAA. All cell lines were maintained as monolayer cultures in a humidified atmosphere containing 5% CO₂ at 37 °C. All compounds were diluted to 100 mM and serially diluted to 50, 10, 1, and 0.1 mM solutions in dimethyl sulfoxide and stored at −20 °C prior to use. These solutions were diluted 500-fold (cell viability, Aβ co-treatment) and 250-fold (Aβ pre-treatment) in the specified cell culture media on the day of dosing for the addition of 50 μL per well.

MTS cell viability assay

Cells were plated at a density of 50 000 cells per well for LNCaP, SH-SY5Y, SK-N-MC, CHLA255, SK-N-BE2 and SK-N-AS cell lines and 10 000 cells per well for PANC-1, WPMY-1, HPNE and PnT2C2 cell line in 50 μL per well of the specified cell culture media and allowed to adapt overnight in a humidified atmosphere containing 5% CO₂ at 37 °C. The cells were treated with a test compound by the addition of a compound in a growth medium (50 μL) at double the final concentration (final compound concentrations 100, 50, 10, 1, 0.1 μM). The plate was incubated in a humidified atmosphere containing 5% CO₂ at 37 °C for 24 h. MTS reagent (CellTiter 96 AQ_ueous one solution reagent, Promega, catalog number G3580) was added 20 μL per well, and the plate was incubated in a humidified atmosphere containing 5% CO₂ at 37 °C for 3–4 h. Absorbance was recorded using the Agilent BioTek Synergy LX Multi-Mode Reader at 490 nm.

Amyloid preparation

The peptide Aβ_1–42, referred to as “Aβ”, was obtained from rPeptide (catalog number A11662). The Aβ peptide was suspended in 100 μM NaOH (295 μL) at a concentration of 750 μM and stored at −80 °C. For use in cell cultures, Ab 750 μM (6.7 μL) was pipetted into PCR tubes (separate tubes for each well) and diluted to 100 μM and 200 μM by the addition of DMEM/F12 media (43.3 μL/18.3 μL). The Aβ solutions were incubated at 37 °C for 24 h prior to addition to wells. The Aβ was diluted to a final concentration of 50 μM by addition to wells.

Co-treatment

For co-treatment experiments with 1 and Aβ. Cells were plated at a density of 50 000 cells per well in 96 well plate and allowed to adapt overnight in a humidified atmosphere containing 5% CO₂ at 37 °C. A solution of 1 in DMSO was added to the 24 h incubated Aβ samples and aspirated to achieve double the final concentration. Cells were treated with compound/Aβ solutions, resulting in a total well volume of 100 μL (final Aβ concentration 50 μM, final DMSO concentration in cells <1%), and incubated for 48 h. MTS reagent (CellTiter 96 AQ_ueous one solution reagent, Promega, catalog number G3580) was added 20 μL per well, and the plate was incubated in a humidified atmosphere containing 5% CO₂ at 37 °C for 3–4 h. Absorbance was recorded at 490 nm.

Pre-treatment

For pre-treatment experiments with compound and Aβ. Cells were plated at a density of 50 000 cells per well in 96 well plate and allowed to adapt overnight in a humidified atmosphere containing 5% CO₂ at 37 °C. The cells were treated by the addition of a solution of compound (4 × final concentration) (25 μL) and incubated for 24 h. The Aβ solution 200 μM (25 μL), resulting in a total well volume of 100 μL (final Aβ concentration 50 μM, final DMSO concentration in cells <1%), and incubated for 48 h. MTS reagent (CellTiter 96 AQ_ueous one solution reagent, Promega, catalog number G3580) was added 20 μL per well, and the plate was incubated in a humidified atmosphere containing 5% CO₂ at 37 °C for 3–4 h. Absorbance was recorded at 490 nm.

Protein expression and purification

A synthetic E. coli codon optimized gene encoding human 17β-HSD10 was PCR amplified and incorporated at the PshA1 cut site of a derivatized pET32 plasmid (EMD Biosciences). The resulting protein possesses a cleavable N-terminal 6× histidine tag. The cloned expression plasmid was used to transform E. coli T7 expression cells (New England Biolabs). The cells were grown in Luria Broth (Research Products International) in the presence of 100 μg mL⁻¹ carbenicillin (Gold Biotechnology) at 37 °C and expression induced at the same temperature with 1 mM IPTG (Gold Biotechnology) for 3 h. The cells were centrifuged at 4000 rpm (Avanti J-15R Beckman Coulter) for 30 minutes and then resuspended in buffer A containing 50 mM Tris pH 7.5, 150 mM NaCl, 0.3 mM TCEP, 5 mM imidazole, and lysed by addition of 1 μM lysozyme (MP Biochemicals, LLC) and sonication (Sonicator 3000, Misonix). A final concentration of 0.1 μM DNase I (Gold Biotechnology) is also added. The lysate was centrifuged at 10 000 rpm (Allegra X-14R Centrifuge, Beckman Coulter) for 30 minutes. The supernatant was loaded onto a HisTrap™ TALON crude 5 mL column (GE Healthcare) and washed with buffer A for 15 column volumes. Histidine tagged protein bound to the column was eluted with buffer B containing 150 mM imidazole, 50 mM Tris pH 7.5, 150 mM NaCl and 0.3 mM TCEP. The N-terminal 6× histidine tag was cleaved by incubation with rhinovirus 3C protease overnight coincident with dialysis against buffer A. After running the sample over the HisTrap™ TALON crude column again, the flowthrough was collected, concentrated by ultrafiltration, and subjected to size exclusion chromatography on a Superdex™ 75 Increase 10/300 column (Cytiva) for further purification. The size exclusion chromatography mobile phase consisted of 20 mM Bis-Tris pH 6.5, 150 mM NaCl, and 0.3 mM TCEP. The only observed peak, corresponding to the 17β-HSD10 dimer, was subjected to SDS-PAGE and stained with Coomassie to assess purity.

Protein thermal shift assay

At a concentration of 10 μM 17β-HSD10 protein in 20 mM Bis-Tris pH 6.5, 150 mM NaCl and 0.3 mM TCEP buffer, different concentrations of the 1 compound including 5 μM, 10 μM, 25 μM, 50 μM, 100 μM were titrated. The final concentration of the DMSO carrier solvent was 1% in each reaction. A control experiment using 10 μM 17β-HSD10 and 1% DMSO but lacking 1 was carried out to compare the melting temperature changes upon binding of the compound. Protein Thermal Shift™ dye (ThermoFisher) was added into each sample at a final concentration of 1× from the 1000× stock solution. The temperature of the samples was elevated at a rate of 0.05 °C s⁻¹ starting from 25 °C to 95 °C using a QuantStudio™ 3 Real-Time PCR system (ThermoFisher). Applied Biosystems™ Protein Thermal Shift™ software v1.4 was used to obtain the derivative plots and determine the melting temperature of the different conditions. Experiments were performed in triplicate.

Molecular modelling

Docking experiments were performed with SeeSAR 12.1 software (BioSolveIT, Sankt Augustin, Germany). The crystal structure of the NADPH 17β-HSD10 homo tetramer in complex with the inhibitor AG18051 (PDBID: 1U7T) was imported into the binding site tool as a PDB file. The reference ligand was removed, and the binding site defined as a 61 amino acid residue pocket directly surrounding the template ligand. The default parameter settings of SeeSAR were maintained. Compounds were prepared and docked with FlexX, wherein fragments are placed into multiple places in the defined pocket and scored with a pre-scoring system.⁶⁶ FlexS⁶⁷ was used to generate compound/reference ligand superimpositioning to determine similarity between the test compound and the reference ligand. Binding poses were scored by hydrogen dehydration (HYDE)⁶⁷ and the top 20 scoring binding poses were imported and analyzed in SeeSAR. The top scoring pose was selected based on estimated affinity, ligand efficiency, and torsion energy. Individual compound docking figures are generated from a perspective to illustrate binding interactions in the most accessible way in a 2D figure, and therefore do not represent a fixed orientation or perspective.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

Compounds in this manuscript have been submitted for intellectual property protection.