Screening of Big Pharma’s Library against Various in-house Biological Targets

Open innovation initiatives provide opportunities for collaboration and sharing of knowledge and experience between industry, academia, and government institutions. Through open innovation, Merck is offering a Mini Library of 80 carefully selected compounds from previous research and development projects to a broader scientific community for testing in academic drug discovery projects. These compounds are predominantly drug-like and cover a broad range of molecular targets. They could potentially interact with other enzymes, receptors, transporters, and ion channels of interest. The Mini Library was tested on seven in-house enzymes (bacterial MurA, MurC ligase, and DdlB enzyme, human MAO-A/B, human BChE, and murine AChE), and several hits were identified. A follow-up series of structural analogues provided by Merck gave a more detailed insight into the accessibility and the quality of the hit compounds. For example, sartan derivatives were moderate inhibitors of MurC, whereas bisarylureas were potent, selective, nanomolar inhibitors of hMAO-B. Importantly, 3-n-butyl-substituted indoles were identified as low nanomolar selective inhibitors of hBChE. All in all, the hit derivatives provide new starting points for the further exploration of the chemical space of high-quality enzyme inhibitors.


Introduction
Pharmaceutical companies and academic research groups often begin their small molecule drug discovery programs by screening compound libraries against biological targets to identify hit compounds that are, subsequently, optimized for potency, physicochemical, and ADMET properties [1][2][3]. Roughly, more than 16 million compounds are available from commercial vendors [4], and large pharmaceutical companies also have internal screening libraries assembled from several million compounds from previous projects [5]. Despite the development of several alternative approaches to drug design, such as virtual screening [6], the fragment-based hit-to-lead design [7], DNA-encoded libraries [8], and others, high-throughput screening (HTS) remains the most effective method for discovering hits and leads [5]. HTS has served as the starting point for numerous approved drugs [9], such as sitagliptin (DPP IV inhibitor) and rivaroxaban (factor Xa inhibitor). However, two features are critical to the success of a compound library screening: a robust and sensitive biochemical assay against the essential biological target and the quality of the compound library [10]. The most important features that determine the quality of the screening library are size, structural diversity, novelty, purity, and pharmaceutical properties, i.e., lead and/or drug-like properties [5]. In addition, problematic compounds, such as frequent hitters, reactive moieties, or redox-cycling compounds, should be identified and eliminated [10,11].
Drug repurposing is a promising approach to rapidly find new therapeutic options for existing drugs [12]. Hits derived from approved drugs have a significant advantage in the drug optimization process, as a considerable amount of biological data is already available.

Results and Discussion
Merck's Mini Library is part of the open innovation initiative, and contains drug-like compounds from former Merck Biopharma research and development projects. Most of these compounds have been well-characterized in vitro and in vivo against their primary targets and toxicologically evaluated to rule out toxic compounds. In addition, clinical data, such as human pharmacokinetics and human safety, are available for certain ligands.
Initially, 80 compounds from Merck's Mini Library were tested at a concentration of 100 µM against the bacterial enzymes MurA, MurC, and DdlB, and at a concentration of 10 µM against the remaining enzymes, i.e., hBChE, mAChE, hMAO-A, and hMAO-B (Supporting Information, Table S1). Results were expressed as residual activities (RAs) at the indicated compound concentration. Three compounds showed a moderate inhibition of MurA, three sartan derivatives inhibited MurC, four compounds were active against hBChE, three against mAChE, and only two inhibited hMAO-B. No active compounds were found against DdlB and hMAO-A. To determine the IC 50 values and to explore the chemical space around the active ligands, Merck secured sufficient amounts of active compounds and their structural analogues.
Salicylic acid analogues have been described as potent and selective inhibitors of glomerular epithelial protein-1 [25], which plays an important role in controlling the chemotaxis of various types of leukocytes. In our screening campaign, they were identified as inhibitors of the bacterial enzymes MurA, mAChE, and hBChE ( Figure 1, Tables 1 and S2). In general, the compounds inhibited mAChE in the nanomolar range, and MurA was inhibited with IC 50 values in the low micromolar range. The substitution of the hydroxyl group with fluorine did not affect the inhibitory potency on both MurA and mAChE (MS-ML24 vs. MS-ML26). The introduction of an amide bond also did not alter the inhibitory potencies to a significant extent (MS-ML24 vs. MS-ML25). The preferred substitution at the diphenylacetylene moiety was a linear alkyl chain, i.e., n-butyl, while the alkyl substituent at the amine/amide did not affect the observed activities. Unfortunately, these compounds did not show clear structure-activity relationships (SARs) against any of the targets, which was also reflected in higher Hill coefficients (>1.5), which could also indicate a nonspecific inhibition due to the highly lipophilic chemical structure, leading to poor solubility in the micromolar concentration range [26]. In addition, the presence of negatively charged moieties-i.e., carboxylic acids-is not optimal for permeation through the blood-brain barrier, which is a prerequisite for target enzymes located in the central nervous system (i.e., AChE and BChE), making salicylic acid analogues of no major interest for further development as cholinesterase inhibitors [27].
range [26]. In addition, the presence of negatively charged moieties-i.e., carboxylic acids-is not optimal for permeation through the blood-brain barrier, which is a prerequisite for target enzymes located in the central nervous system (i.e., AChE and BChE), making salicylic acid analogues of no major interest for further development as cholinesterase inhibitors [27].  4,5-Dihydro-4-oxo-3H-imidazo [4,5-c]pyridine derivatives are nonpeptide angiotensin II receptor antagonists. This class of compounds, named sartans, is widely used in clinics for the treatment of hypertension and heart failure [28][29][30]. The follow-up sartan derivatives were assayed at 100 µM against MurC, and the IC50 values were determined for compounds showing RAs below 50% (Table 2). Most sartan analogues inhibited MurC in a micromolar concentration range, with IC50 values of approximately 100 µM. Variously substituted benzyls were well-tolerated at position five of the 3,5dihydro-4H-imidazo [4,5-c]pyridin-4-one core, whereas the introduction of the acyclic tertbutyl moiety was not tolerated and resulted in the loss of MurC inhibition (compounds 15 and 17). To further explore the SARs of the sartan analogues against MurC, a larger number of compounds would need to be tested.  4,5-Dihydro-4-oxo-3H-imidazo[4,5-c]pyridine derivatives are nonpeptide angiotensin II receptor antagonists. This class of compounds, named sartans, is widely used in clinics for the treatment of hypertension and heart failure [28][29][30]. The follow-up sartan derivatives were assayed at 100 µM against MurC, and the IC 50 values were determined for compounds showing RAs below 50% (Table 2). Most sartan analogues inhibited MurC in a micromolar concentration range, with IC 50 values of approximately 100 µM. Variously substituted benzyls were well-tolerated at position five of the 3,5-dihydro-4H-imidazo [4,5-c]pyridin-4-one core, whereas the introduction of the acyclic tert-butyl moiety was not tolerated and resulted in the loss of MurC inhibition (compounds 15 and 17). To further explore the SARs of the sartan analogues against MurC, a larger number of compounds would need to be tested.
The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B (Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors.    The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B (Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table  S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors. The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B (Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table  S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors.
The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B ( Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table  S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors.

MS-ML80
The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B (Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table  S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors. The Mini Library screening identified two structurally distinct classes of hMAO-B inhibitors (Tables 3 and S3). Heterocyclic-substituted bisarylureas were patented by Merck as kinase inhibitors; notably, these compounds inhibit the VEGF-stimulated mitogenesis of human vascular endothelial cells [31]. In addition, triazaindolizines have been patented as inhibitors of methionine aminopeptidase 2 [32]. Several bisarylureas inhibited hMAO-B in the low micromolar to submicromolar range, and small structural changes resulted in large differences in inhibitory potencies. For example, the introduction of N-methylacetamide on the pyridine ring resulted in the complete loss of activity, whereas the replacement of the pyridine moiety with aminopurine (22) or the complete removal of the pyridine to obtain p-aniline analogue 23 resulted in equipotent inhibitors. The introduction of substituents on the pyridine ring, e.g., N,N-dimethylethyl (27) or piperidine-1-ethyl (28), resulted in active compounds. Only triazaindolizine MS-ML31 with the phenyloxypyridinyl fragment inhibited hMAO-B (Table 3). The replacement of phenyloxypyridine with substituted quinolines, (tetrahydro)isoquinolines, and phenyls completely abolished hMAO-B inhibition (Table  S3). All compounds were inactive against hMAO-A, indicating that bisarylureas were selective hMAO-B inhibitors.  A number of indoles were described by Merck as dopaminergic receptor agonists [33], and these compounds were found to be potent hBChE inhibitors ( A number of indoles were described by Merck as dopaminergic receptor agonists [33], and these compounds were found to be potent hBChE inhibitors (Table 4). The most potent inhibitors were analogues with unsubstituted or substituted indoles via an n-butyl linker to 4-phenyl-1,2,3,6-tetrahydropyridine. Bulky groups such as naphthene (30) or 6-methoxy-2methyl-2,3,4,9-tetrahydro-1H-pyrido [3,4-b]indole (37) linked to 1,2,3,6-tetrahydropyridine reduced hBChE inhibition. All compounds were also assayed for their mAChE inhibition, and most of them were selective hBChE inhibitors. Indole and tryptophan derivatives are known hBChE inhibitors and have also been extensively studied in our group as promising pharmacological tool molecules and lead compounds for the treatment of Alzheimer's disease [34,35].
a successful screening of Merck's Mini Library on seven in-house enzymes, revealing several hit compounds. Particular attention should be paid to validate the activity detected in the screening phase and not to pursue hits with questionable activity. Salicylic acid analogues fell into the latter category, while sartans, as moderate MurC inhibitors, needed to be validated in secondary assays to confirm the clear SARs of these hits. On the other hand, bisarylureas and indoles as selective, nanomolar hMAO-B and hBChE inhibitors, respectively, in combination with Merck's data on these derivatives (e.g., selectivity, preclinical pharmacokinetics, and toxicology) [39][40][41][42][43][44][45][46][47][48][49], provided a solid foundation for the further exploration of their activities. Further cell-based and in vivo rodent models of neurodegenerative diseases should be used to reveal the true potential of hit compounds as potential agents against neurodegenerative disorders.  Funding: This research was funded by the Slovenian Research Agency (core research funding P1-0208, and projects Z1-1859 and J1-2484).

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author, M.H., upon reasonable request.