Novel selective indole based histone deacetylase 10 inhibitors as anticancer therapeutics

Abstract

Histone deacetylase (HDAC) inhibitors represent a promising class of anti-cancer agents that play a key role in both epigenetic and non-epigenetic regulation, leading to cancer cell death, apoptosis, and cell cycle arrest. This study synthesized novel bicyclic hydroxamic acid derivatives and evaluated their inhibitory and selectivity activity against class I and IIb HDACs. Our findings demonstrate that Compound 2e specifically inhibits HDAC10 with high selectivity over HDAC6, while shows no significant impact on class I HDACs. Compound 2a exhibited the most potant inhibitory activity against HDAC10, with IC₅₀ 0.41 ± 0.02 nM. In contrast, Compound 2f revealed a preference toward HDAC6, with an IC₅₀ value of 2.5 ± 0.3 nM. Compounds 2c and 2d demonstrated high selectivity toward class IIb over class I HDACs. Docking and molecular dynamics studies revealed that compound 2a fits well into the active site of HDAC10, forming stable and strong interactions with key residues F204, D94, W205, and E274 in HDAC10. In addition, we assessed the anti-proliferative activity these compounds against a panel of four human solid tumor cell lines. To evaluate their selectivity, non-cancerous kidney cell lines (LLC-PK1 and VERO) were employed to determine the effects of these compounds on normal cell proliferation.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-02774-6.

Introduction

Histones play a crucial role in regulating gene expression, the process by which the coded information in genes is converted into functional structures within the cell. Histone acetylation (HA) neutralizes the positive charge on histones by converting amine groups into amide groups¹. This modification reduces the affinity of histone for DNA, leading to chromatin relaxation (chromatin expansion), which permotes genetic transcription^2,3. Histone deacetylases (HDACs) are enzymes that remove the acetyl group from acetylated lysine residues on both histone and non-histone proteins. This removal restores the positive charge on histones, increasing their affinity for DNA and resulting in chromatin condensation, which suppresses gene expression. Histones are basic proteins present in the chromatin of eukaryotic cell nuclei and are essential for organizing DNA into structural units called nucleosomes, the fundamental repeating units of chromatin^4,5.

There are 18 subtypes of HDACs encoded in the human genome. These enzymes are categorized into two main groups; the first group comprisesis zinc-dependent HDACs, which are further divided into the following sub-classes: class Ia (HDAC1, HDAC2), class Ib (HDAC3), and class Ic (HDAC8), class IIa (HDAC4, HDAC5, HDAC7, and HDAC9), class IIb (HDAC6 and HDAC10), and class IV (HDAC11). The second group consists of nicotinamide adenine dinucleotide (NAD⁺)-dependent HDACs, known as class III HDACs or sirtuins^6,7. Accumulating evidence has domonstrated the crucial role of HDACs in various cellular processes, including autophagy, cell cycle control, and apoptosis^8–11. Recent studies indicate the protective role of HDACs function against DNA damage, making them promising anti-tumor therapies¹¹.

Several HDAC inhibitors have been identified, and some (Fig. 1) were approved by FDA. For instance, vorinostat (SAHA) was approved for the treatment of cutaneous T cell lymphoma (CTCL)¹², Recently, the introduction of an amino group (“aza-scan”) into the hexyl linker moiety of vorinostat (SAHA) yielded a selective chemical probe, DKFZ-748, targeting histone deacetylase 10 (HDAC10). This modification conferred unprecedented selectivity for HDAC10 over other HDAC isozymes, highlighting its potential as a highly specific therapeutic tool¹³. While belinostat (PXD-101) and romidepsin (FK-228) were approved for the treatment of peripheral T-cell lymphoma (PTCL)^14,15. Romidepsin has also shown efficacy in treating bipolar disorders and migraine. In addition, panobinostat (LBH-589) was approved for the treatment of multiple myeloma. These inhibitors induce growth arrest, differentiation, and apoptosis of many transformed cells, thereby eliminating cancer cells¹⁶. The catalytic domains of zinc-dependent HDACs share a high degree of homology. However, the lack of isoform specific selectivity in approved HDAC inhibitors often leads to potential side effects^17–19. Non-selective HDAC inhibitors, such as vorinostat and panobinostat modulate multiple pathways and factors in MDA-MB-231, 4T1, and BT-549 cell lines. this effect includes the suppression of growth factor FOXA1²⁰, upregulation of tumor suppressor factors p21 and p27, and downregulation of the anti-apoptotic protein Bcl-2²¹. In addition, both inhibitors can suppqress matrix metalloproteinase (MMP9) activity²².

Fig. 1.

Example of approved HDAC inhibitors by the Food and Drug Administration (FDA).

Recently, HDAC10 has been shown to play a unique role in neuroblastoma cells, where its inhibition lead to the accumulation of autolysosomes. This result indicates that HDAC10 may be potential therapeutic target for neuroblastoma treatment²³. Conversely, HDAC10 has been found to suppress tumorigenesis in cervical cancer by downregulating miR-233 expression, which subsequently targeting EPB41L3²⁴. Consequently, there is a critical need for HDAC10-selective chemical probes to validate its role in both physiological and pathological processes. Such tools are essential to establish a mechanistic connection between HDAC10’s function as a polyamine deacetylase and the phenotypic outcomes observed following genetic manipulation of HDAC10. Additionally, HDAC10 inhibitors (HDAC10i) hold substantial therapeutic promise, particularly in oncology and as immunomodulatory agents^12–22.

Investigating the biological relevance of HDAC10 and evaluating its pharmacological role in cancer cell lines highlight the need for selective HDAC10 inhibitors. To date, only few inhibitors targeting HDAC10 have been identified²⁵, such as tubastatin A and its derivative (1b), which are selective for HDAC subfamily IIb²⁶. Recently, TH34 was reported as a class I/IIb selective HDAC inhibitor²⁷. Additionally, compounds including 2-(oxazol-2-yl)phenol moiety have been identified as HDAC1/6/10 inhibitors²⁸. The importance of a basic moiety in the linker group has also been demonstrated in a recently developed series of piperidine-based HDAC10 inhibitors^29,30. A major challenge lies in introducing HDAC10 inhibitors with high selectivity over the class IIb isozyme member HDAC6.

Most HDAC inhibitors exhibit a conserved pharmacophoric features, including a ‘cap’ group and hydrophobic chain (linker); both of which influence selectivity. Additionally, the zinc-binding group (ZBG) is critical for the inhibitor’s potency³¹. SAHA¹², a pan-HDAC inhibitor, holds the distinction of being the first HDAC inhibitor approved for clinical use (Fig. 1). Its structure features a hexyl linker (black), which serves as a mimic of the lysine side chain. This linker connects the hydroxamic acid zinc-binding group (red) to the anilide “cap” group (green), forming the core architecture of the molecule. We hypothesized that mimicking the structural framework of SAHA¹² and Panobinostat (LBH-589)¹⁶ inhibitors by incorporating hydrophobic moieties, such as indole or quinoline—privileged structures commonly found in various anticancer drugs. Additionally, the inclusion of an amino group and a hydroxamic acid would introduce high polarity, while the use of indole or quinoline as cap groups could improve their drug-like properties. On the other hand, to optimize physicochemical properties and minimize the number of rotatable bonds, we proposed the use of rigid linkers, such as piperidine and aromatic moeity. The significance of incorporating a basic moiety within the linker group has been further supported by recently developed series of HDAC10 inhibitors^29,30. Herein, we report efforts to optimize indole and quinoline ligands as HDAC10-selective inhibitors. Several modified derivatives that were developed exhibited significant HDAC10 inhibition and selectivity.

Results and discussion

Chemistry

The compounds in this study were synthesized as outlined in Schemes 1 and 2. The starting building block, 1H-indole-3-carbaldehyde, was treated with 4-methylbenzene-1-sulfonyl chloride in anhydrous N, N’-Dimethylformamide (DMF) using sodium hydride as a base, this intermediate was then coupled with various amine compounds via a reductive amination reaction using sodium triacetoxyborohydride ((CH₃COO)₃BHNa). Finally, a condensation reaction with hydroxylamine and potassium hydroxide KOH to afford final targeted compounds 2c-e, as shown in Scheme 1.

Scheme 1.

Reagent and condition: (a) DMF, NaH, TsCl, ice bath, 15 min. (b) CH₂Cl₂, (CH₃COO)₃BHNa, r.t, overnight. (c) CH₂Cl₂/MeOH (1:2), hydroxylamine (50 wt % in H₂O), NaOH, r.t, 24 h.

Scheme 2.

Reagent and condition: (a) CH₂Cl₂, (CH₃COO)₃BHNa, r.t, overnight. (b) CH₂Cl₂/MeOH (1:2), hydroxylamine (50 wt % in H2O), NaOH, r.t, 24 h.

Similarly, indole-3-carboxaldehyde and quinolone were treated in the same manner as indole-3-carboxaldehyde to produce compounds 2a-b, and 2f-g Schemes 1 and 2.

Pharmacology/biology

Synthesis and in vitro testing of novel inhibitors

Recently, piperidine-4-acryl hydroxamates (PZ45 and PZ48) were reported as potent HDAC10 inhibitors with high specificity for HDAC10 over HDAC6 and with no significant impact on class I HDACs³⁰. The selectivity of these newly discovered ligands PZ45/48 was further evaluated in acute myeloid leukemia (AML) cells harboring the FLT3-ITD oncogene, demonstrating promising activity.

In the current work, we extended the structure-activity relationships of HDAC10 inhibitors by synthesizing a series of benzhyldroxamic acid derivatives featuring a zinc-binding group (ZBG) and a bicyclic aromatic moiety as a capping group. These derivatives were designed with a methylene spacer linked to a basic amine group, maintaining the common pharmacophoric features essential for HDAC inhibition as previously described. All prepared derivatives were evaluated in vitro for their inhibitory activity against HDAC10 (from zebrafish) and human HDACs 1, 6, and 8 (Table 1). Compound 2a, which incorporates an indole methylene capping group, demonstrated potent inhibition of HDAC10 with an IC₅₀ of 0.41 ± 0.02 nM (Table 1). The tosylatated indole analogue, compound 2c, also exhibtied nanomolar HDAC10 inhibition with an IC₅₀ of 4.5 ± 0.3 nM. Figure 3, shows that hydrophobic tosyl moiety does not interact with the rim of the HDAC10 binding site and instead remains fully exposed to the solvent. However, the removal of the methylene spacer between the primary amine and the benzhydroxamic acid moiety in analogs 2d and 2f result in a significant reduction in the inhibitory activity, with IC₅₀ of 290 ± 60 nM and 110 ± 10 nM, respectively. The loss of salt bridge interactions with D94 and E274, along with the cation-π interaction with W205 in HDAC10, which were present in the other compounds reported in this study, is evident. As anticipated, only hydrogen bond interactions between the NH-group and D94 are observed (Figure S1b and Figure S1c).

Table 1.

Chemical structure and IC₅₀ values for synthesized compounds against HDAC10.

ID	Structure	IC50 drHDAC10 ± SD [nM]
2a		0.41 ± 0.02
2b		2.0 ± 0.1
2c		4.5 ± 0.3
2d		290 ± 60
2e		75 ± 12
2f		110 ± 10
2g		11 ± 1.0
Tubastatin A		50 ± 3

Fig. 2.

Predicted binding modes of compound 1 in different HDAC isoforms: (a) 2a (teal sticks) in drHDAC10 (PDB ID 6UHU), (b) 2a (violet sticks) in HDAC6 (PDB ID 5EDU), (c) 2a (orange sticks) in HDAC8 (PDB ID 2V5X) (d) 2a (magenta sticks) in HDAC1 (PDB ID 5ICN). The surface of the proteins is colored according to lipophilicity; green for hydrophobic and magenta for hydrophilic. Side chains of binding site residues are shown as white sticks and the catalytic zinc ion as an orange sphere. H-bond interactions are depicted as blue-dashed lines, salt bridge interactions as magenta-dashed lines, π-π interactions as orange-dashed lines, cation-π interactions as green-dashed lines and zinc ion coordinationby the ligand as yellow-dashed lines.

The second series of compounds comprise of 4-(1-piperazinyl)benzhydroxamic acid derivatives that bear a bicyclic aromatic moiety as a capping group. Compound 2b exhibited potent HDAC10 inhibitory activity with an IC₅₀ = 2.0 ± 0.1 nM. In contrast the methyltosylated indole derivative, compound 2e, showed a decrease in the HDAC10 inhibitory activity with an IC₅₀ of 75 ± 12 nM (Table 2). Meanwhile, replacing the indole moiety in compound 2b with aquinoline moiety yielded compound 2g, which was slightly less potent, displaying an IC₅₀ value of 11 ± 1 nM.

Table 2.

In vitro selectivity of newly synthesized HDAC10 inhibitors.

ID	hHDAC1 IC50 [nM]	hHDAC6 IC50 [nM]	hHDAC8 IC50 [nM]
2a	4500 ± 200	37 ± 2	350 ± 20
2b	48.4%@1 µM 90.6% @10 µM	73 ± 3	74 ± 6
2c	22.2%@1 µM 72.8% @10 µM	51 ± 5	65.1%@1 µM 97.3% @10 µM
2d	1.3%@1 µM 38.9% @10 µM	53 ± 5	49.2%@1 µM 98.9% @10 µM
2e	7.3%@1 µM 67.2% @10 µM	35.3%@1 µM 95% @10 µM	26.9%@1 µM 87.9% @10 µM
2f	51.7%@1 µM 91.4% @10 µM	2.5 ± 0.3	210 ± 10
2g	32.1%@1 µM 79.4% @10 µM	130 ± 10	300 ± 20
TubastatinA	7.6% @1 µM 29.9%@10 µM	19 ± 1	46.4% @1 µM 84% @10 µM

In conclusion of the structure-activity relationship (SAR) analysis, the absence of a strong basic amino group was found to reduce the inhibitory activity against HDAC10. Additionally, modifying the capping group with a hydrophobic tosyl moiety also led to a decrease in inhibitory activity.

The preliminary structure-activity relationship (SAR) analysis revealed that the incorporation of bicyclic aromatic moiety enhanced HDAC10 inhibitory activity. We designed and synthesized seven compounds featuring various substituents at the cap position and linker. Among these, the benzhydroxamic acid derivative bearing an indole ring, compound 2a, stood out as the most potent HDAC10 inhibitor, demonstrating exceptional enzyme inhibitory activity and selectivity toward HDAC10. Herein we introduce highly selective ligand 2e for HDAC10 with nanomolar inhibitory activity.

Enzymatic in vitro testing

All the synthesized compounds were evaluated in vitro against zebrafish HDAC10 (drHDAC10), and human HDAC 1, 6, and 8 (details see Methods section). DrHDAC10 was chosen as the close homolog of human HDAC10 since it was found to be more stable and easier to express compared to the human HDAC10^30–32. It’s worth mentioning that none of the compounds displayed potent inhibitory against class I HDACs. Data in Table 2 indicate that all compounds retained activity on HDAC10 comparable or even higher than the reference HDAC6/10 inhibitor Tubastatin A. We reported that the indole ring significantly affected the potency and selectivity towards HDAC10³⁰. Compounds 2a bearing a benzhydroxamic acid displays a highly potent inhibitory activity with IC₅₀ of 0.41 ± 0.02 nM towards HDAC10, but also potent against HDAC6 and moderately potent against HDAC8 with IC₅₀ value of 37 ± 2 and 350 ± 20 nM, respectively.

Compounds 2b and 2g showed a preference for HDAC10 (IC₅₀ 2.0 ± 0.1 and 11 ± 1 nM, respectively) compared with HDAC8 and HDAC6. As can be expected, compound 2f, which only contains a weakly basic aromatic amine moiety, showed a significant decrease in the HDAC10 inhibitory activity and high preference and potency to HDAC6 with an IC₅₀ value of 2.5 ± 0.3 nM, which again highlights the importance of the protonated amine group to achieve high HDAC10 potency and selectivity.

Meanwhile, compounds 2c and 2d showed a preference for class IIb enzymes (HDAC10 and 6) with no significant inhibition on HDAC1/8. HDAC6 is well documented, it plays a crucial role in microtubule deacetylation and regulation of PDL1 and further targets related to cancer immunotherapy^33,34.

In combination with the observed in vitro potency and selectivity of compounds 2c and 2d, both represent promising hits for further optimization. The lack of specific and potent HDAC10 inhibitors led to limited knowledge about the biological function of HDAC10³⁰. A more distinguished result showed by compound 2e examined the inhibitory of compound 2e revealed high selectivity with a novel nanomolar inhibitor toward HDAC10 (IC₅₀ 75 ± 12 nM).

Cytotoxic activity

All the synthesized compounds (Table 3) were screened at three different concentrations for their cytotoxicity towards a panel of four human solid tumor cell lines: melanoma (SK-MEL), epidermal carcinoma (KB), breast carcinoma (BT-549), and ovarian carcinoma (SK-OV-3). Moreover, non-cancer kidney cell lines (LLC-PK1 and VERO) were also included in the study.

Table 3.

Cytotoxicity of compounds towards a panel of cell lines.

ID	Cancer cell lines IC50 µM				Kidney cells IC50 µM
	SK-MEL	KB	BT-549	SK-OV-3	LLC-PK1	Vero
2a	NC	72.80 ± 2.40	NC	60.95 ± 14.36	NC	62.64 ± 2.39
2b	25.97 ± 3.62	12.13 ± 1.41	16.84 ± 3.63	11.42 ± 4.04	13.27 ± 1.41	5.99 ± 0.40
2c	15.12 ± 0.94	20.25 ± 0.32	16.46 ± 0.32	21.36 ± 1.89	16.57 ± 3.93	21.36 ± 1.27
2d	8.38 ± 2.11	13.32 ± 0.33	11.37 ± 0.16	12.63 ± 0.98	9.99 ± 0.49	50.52 ± 9.74
2e	8.82 ± 0.70	10.01 ± 0.70	9.71 ± 0.28	8.62 ± 0.42	8.82 ± 1.26	13.18 ± 1.82
2f	15.17 ± 1.21	9.72 ± 0.73	14.83 ± 2.65	10.91 ± 2.89	9.38 ± 0.25	3.92 ± 0.24
2g	> 69.03	> 69.03	> 69.03	> 69.03	> 69.07	51.04 ± 1.95
Doxorubicin	2.52	2.30	3.44	2.01	2.34	> 11

NC: No cytotoxicity up to 80 µM. Values are represented as mean ± SD (n = 3).

In the first screening, all compounds were tested at a concentration of 80 µM by the MTT assay. The result showed all compounds had cytotoxicity at 80 µM, and hence they were carried out for the second round of screening at a lower concentration. Data showed that all compounds displayed moderate cytotoxicity than the control compound, doxorubicin. Observing the dataset noted that all HDAC10 inhibitors were weakly toxic for LLC-PK1 and VERO kidney cell lines.

Molecular docking

Molecular modelling studies

Docking studies have been conducted to elucidate the binding mode of the synthesized compounds in various HDAC isoforms, including HDAC1 (PDB ID: 5ICN), HDAC6 (PDB ID: 5EDU), HDAC8 (PDB ID: 2V5X) and drHDAC10 (PDB ID: 6UHU). To test the plausibility of the docking results, we compared them with several crystal structures of structurally related hydroxamic acids, including those we recently reported for drHDAC10.

Tubastatin A, a known inhibitor of HDAC6 and HDAC10²⁹, along with the newly synthesized inhibitors described here, were subjected to docking studies within the crystal structures of various HDACs. The binding modes were visually analyzed and the docking studies revealed that tubastatin A and the newly synthesized compounds can chelate the zinc ion in a bidentate manner in HDAC1, HDAC8, and HDAC10. On the other hand, in HDAC6, zinc chelation was observed to occur via the hydroxyl oxygen of the hydroxamate moiety in a monodentate fashion that has also been reported for other benzhydroxamic acids³⁵. Docking solutions in all investigated HDAC isoforms (Figs. 2, 3, and Supporting information S1: Figures a-d) showed that the phenyl moiety of the linker was embedded in the hydrophobic lysine binding pocket. Furthermore, hydrogen bond interactions with the conserved residues H136/142/140, H137/143/141, and Y307/306/303 were observed in HDAC10, HDAC8, and HDAC1, respectively. In contrast, in HDAC6, the hydroxyl oxygen formed a water-mediated hydrogen bond with the conserved residues H610 and H611.

Fig. 3.

Predicted binding modes of compounds, (a) 2b (peach sticks) and (b) 2e (teal sticks) in drHDAC10 (PDB ID 6UHU). The surface of the proteins is colored according to lipophilicity; green for hydrophobic and magenta for hydrophilic. Side chains of binding site residues are shown as white sticks and the catalytic zinc ion as orange spheres. H-bond interactions are depicted as blue-dashed lines, salt bridge interactions as magenta-dashed lines, π-π interactions as orange-dashed lines, cation-π interactions as green-dashed lines, and coordination of the zinc ion by the ligand as yellow-dashed lines.

The predicted binding mode of compound 2a in HDAC10 (Figs. 1a, 2a) provides insights into its significant inhibitory activity against HDAC10. In addition to the previously mentioned hydrogen bond interactions with conserved tyrosine and histidine residues, the capping group exhibits π-π interactions with F204. Moreover, the protonated amine forms electrostatic and salt-bridge interactions with D94 and the gatekeeper residue E274, along with cation-π interactions with W205 in HDAC10. These interactions resemble those observed in previously reported potent HDAC10 inhibitors and polyamine substrates^13,29,30,36. HDAC6 belongs to class IIa of HDACs. In comparison to HDAC10, the HDAC6 binding pocket exhibits specific differences, notably involving mutations D94/S568 and E274/L749. These variations in the binding pocket led to the docking results showing only a single hydrogen bond interaction between the protonated nitrogen of compound 2a and S568. This observation could potentially explain the approximately 90-fold selectivity of HDAC10 over HDAC6 (Fig. 2b). In class I HDAC members (HDAC1 and HDAC8), variations in the gatekeeper residue at the top of the lysine binding tunnel (E274 in HDAC10 is replaced by M274 and L271 in HDAC1 and − 8, respectively) led to the loss of one electrostatic interaction with the protonated nitrogen of 2a as well as the π-π interactions via the capping group, which were observed in HDAC10. The solvent-exposed capping groups together with the loss of electrostatic interaction with the gatekeeper are most likely the reason for the decrease in activity in both isoforms. Furthermore, the narrow binding pocket of HDAC1 due to the E274/L271 variation might lead to steric hindrance for the linker and capping groups of 2a which may explain the higher loss of activity in this isoform (Fig. 2c, d).

Similar results were observed for compounds 2b and 2g, which bear a piperazine linker instead of the aminomethyl linker of compound 2a. Here, the protonated piperazine-NH was able to undergo two electrostatic interactions with D94 and E274 and cation-π interactions with W205. Additionally, the capping groups showed similar π-π interaction with W205 in the HDAC10 binding pocket as observed for compound 2a (Fig. 3a and Figure S1d).

Compounds lacking a strong basic amino group including compounds 2d and 2f showed a significant decrease in the HDAC10 inhibitory activity. This can be attributed to the loss of the salt bridge interactions with D94 and E274, and the cation-π interaction with W205 in HDAC10, which were observed with the other compounds reported in this study. As expected, only hydrogen bond interactions between the NH-group and D94 are observed (Figure S1b and Figure S1c). These findings further emphasize the significance of the protonated nitrogen and its precise positioning in optimizing the binding interactions with HDAC10.

The in vitro data presented in this study demonstrate that substituting the capping group with an additional tosyl moiety result in a decrease in HDAC10 inhibitory activity. Compound 2c exhibits approximately a 10-fold reduction in HDAC10 inhibitory activity compared to compound 2a, while compound 2e shows an almost 38-fold decrease in activity compared to its unsubstituted counterpart, compound 2b. The predicted binding mode of compound 2e (Fig. 3b) was found to be similar to that of compound 2b (Fig. 3a). However, an additional hydrogen bond via the sulfonyl group with N207 was observed in compound 2e. Notably, the hydrophobic tosyl moiety does not interact with the rim of the HDAC10 binding site and remains fully exposed to the solvent. This observation could explain the decrease in HDAC10 inhibitory activity compared to compound 2b.

Molecular dynamics (MD) simulations

For the validation of the MD protocol, 100 ns MDs were applied to the available humanized drHDAC10 crystal structures (PDB IDs; 7U6B, 7U69, 7U6A, and 7U59, details in the Methods Section). Obtained RMSD plots from the molecular dynamic simulations revealed that the HDAC10-inhibitor structures (Figure S3c, Figure S4c, Figure S5c, and Figure S6c) are stable with an RMSD below 1.5 Å while the Zn ions generally show lower root mean square deviation (RMSD) values (Figure S3d, Figure S4d, Figure S5d and Figure S6d). Meanwhile, the ligand molecules tended to show higher fluctuations (Figure S3a, Figure S4a, Figure S5a, and Figure S6a) which was majorly attributed to the flexibility of the capping groups as observed in the ligand root mean square fluctuations (RMSF) plots (Figure S3b, Figure S4b, Figure S5b and Figure S6b).

We further performed molecular dynamics simulation studies on the obtained docking pose of the most active compound 2a in drHDAC10 to examine the stability of the predicted binding mode. The obtained protein-ligand complex was subjected to 100 ns MD simulation protocol three times using AMBER22, and the obtained trajectories were analyzed. RMSD plots showed that the protein structure and the zinc ion in the complex remained stable during the 100 ns simulation time in all three replicas with RMSD values below 1.5 Å (Figure S2a and Figure S2b). Meanwhile, the ligand RMSD plots (Fig. 4a) demonstrate that the ligand shows significant deviations from the predicted docking pose. Further analyses were performed to assess the stability of the obtained docking pose of compound 2a, and analyze the causes behind the observed fluctuation. Examination of the ligand RMSF plots of compound 2a (Fig. 4b) demonstrates that the zinc-binding hydroxamate group as well as the linker moiety are stable during the MD simulations with RMSF values < 1 Å. It’s worth noting that the interactions between the hydroxamate group and the zinc ion are maintained during the MD simulations as shown by the unchanged distances between the hydroxamate-O atoms and zinc ion (< 2.5 Å; Fig. 3c). Meanwhile, the indole-capping group displayed significantly high RMSF values that explains the strong deviations observed in the ligand RMSD plots. Clustering the obtained MD trajectories based on the RMSD of the ligand yielded three clusters with an occupancy > 10%. As can be demonstrated by the inspection of the obtained clusters (Fig. 4d), the indole capping group can occupy one of two different hydrophobic regions at the rim of the binding pocket: In two clusters (occupancy 55% and 15%, respectively), the capping group is undergoing hydrophobic interactions with I27, A28 and P29. In the third cluster (17% occupancy), the capping group is situated between F204 and W205 (Fig. 4C). The calculated data show, that the indole-capping group, despite showing strong fluctuations, is well accommodated at the entrance of the lysing binding pocket of HDAC10 where it undergoes hydrophobic interactions with surrounding residues.

Fig. 4.

(a) RMSD plots of the compound 2a heavy atoms in drHDAC10 (PDB ID 6UHU)-compound 2a docked complexes for 3 repeated MD runs each for 100 ns. (b) RMSF plots of compound 2a heavy atoms in drHDAC10 (PDB ID 6UHU)-compound 2a docked complexes for three repeated MD runs each for 100 ns s (c). The measured distance between both oxygen atoms of the hydroxamate zinc binding moiety and catalytic zinc ion for three repeated MD runs. (d) Representative poses of the obtained three clusters with an occupancy > 10%, teal sticks for the first cluster (occupancy 55%), magenta sticks for the second cluster (occupancy 15%), and yellow sticks for the third cluster (occupancy 17%). Side chains of the relevant binding site residues are shown as white sticks and the catalytic zinc ion as orange spheres.

Conclusion

In this study, we synthesized novel bicyclic hydroxamic acid derivatives, characterized their inhibitory activity against class I, and class IIb HDACs. Our findings highlight compound 2e as a potent and selective inhibitor of HDAC10, while compound 2a exhibited the best inhibitory activity against HDAC10. Compound 2f displayed a preference towards HDAC6, and compounds 2c and 2d demonstrated high selectivity towards class IIb HDACs over class I HDACs.

Structural insights obtained from HDAC10–inhibitor complexes, along with the SAR study, provide a detailed understanding of the HDAC10 active site and highlight the critical structural features necessary for achieving potent and selective HDAC10 inhibition. Docking models with HDAC6 further elucidate the structural basis for HDAC10 selectivity over HDAC6, which is largely driven by the hydrophilic properties of the amino group integrated into the SAHA linker. Compounds lacking a strongly basic amino group exhibited a notable reduction in HDAC10 inhibitory activity. This decline can be explained by the loss of key interactions, such as salt bridges with D94 and E274, as well as cation-π interactions with W205 in HDAC10, which were consistently observed with other compounds in this study. As anticipated, only hydrogen bond interactions between the NH-group and D94 were detected (Figure S1b and Figure S1c). These results underscore the importance of the protonated nitrogen and its precise spatial orientation in enhancing binding interactions with HDAC10, providing a strong rationale for the design of further selective HDAC10 inhibitors.

On the other hand, molecular dynamics (MD) simulations reveal that the indole capping group can occupy one of two distinct hydrophobic regions at the rim of the binding pocket. In two clusters (with occupancies of 55% and 15%, respectively), the capping group engages in hydrophobic interactions with I27, A28, and P29. In a third cluster (17% occupancy), the capping group is positioned between F204 and W205 (Fig. 4C). The calculated data demonstrate that the indole capping group, despite exhibiting significant fluctuations, is well accommodated at the entrance of the lysine-binding pocket of HDAC10, where it participates in hydrophobic interactions with surrounding residues. This further supports the strategic design of inhibitors targeting HDAC10.

Experimental

Chemistry

General

Reagents and hydrous or anhydrous organic solvents were purchased from Sigma-Aldrich (Darmstadt, Germany) and Alfa Aesar Chemicals (Tewksbury, MA, USA) and were used without further purification. Reaction progress was monitored by thin-layer chromatography (TLC) on pre-coated 0.20 mm silica gel GF Uniplates from Macherey-Nagel (Düren, Germany). The TLC plates were visualized using a 254 nm UV lamp and chemical indicators such as ninhydrin, potassium permanganate (KMnO₄), dinitrophenylhydrazine (DNP), green girasol, and anisaldehyde. Column chromatography was performed with 63–200 µM silica gel with a 70–230 mesh size. Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker BB 400 MHz spectrometer. Chemical shifts (δ) are reported in ppm relative to tetramethyl silane (TMS) as an internal standard and coupling constants (J) are reported in Hz. The following abbreviations were used to denote multiplicity: s = singlet d = doublet t = triplet q = quartet dd = doublet of doublets dq = doublet of quartets m = multiplet.

General synthesis methods

Preparation of 1-tosyl-1H-indole-3-carbaldehyde (2)

This compound was prepared from 1H-indole-3-carbaldehyde and p-toluenesulfonyl chloride, according to the previously reported^37–39.

1H-indole-3-carbaldehyde (1) (1equiv., 3 g, 20.7 mmol) was dissolved in anhydrous N, N’-Dimethylformamide (DMF, 30 mL). The solution was stirred under nitrogen for 15 min in an ice bath. Sodium hydride (NaH) (1.1 equiv., 0.55 g, 22.9 mmol) was then added, followed by the slow addition of 4-methylbenzenesulfonyl chloride (1.4 equiv., 5.52g, 28.8 mmol) after 10 min. The reaction mixture was stirred overnight at room temperature. The reaction was quenched with distilled water (10 mL), and then a saturated sodium bicarbonate (NaHCO₃) solution (10 mL). The product was extracted three time with dichloromethane (CH₂Cl₂, 25mL each). The solvent was evaporated under reduced pressure, and the product was recrystallized using CH₂Cl₂ (15mL) and hexane (10mL). The resulting light yellow crystals were washed with cold methanol. The yield was 75% (4.61 g, 15.4 mmol), m.p 148–149 °C and R_f 0.88 (Hexane: EtOAc, 1:1). ¹H-NMR (400 MHz, CDCl₃-d, TMS): δ 2.36(s, 3 H, Ar-CH₃), 7.29(d, J = 8.4 Hz, 2 H, H-12), 7.36(ddd, J = 7.9, 7.3, 2.0 Hz, 1H, H-6), 7.42(ddd, J = 7.3, 8.2, 1.6 Hz, 1H, H-7), 7.87(d, J = 8.4 Hz, 2 H, H-11), 7.97(dt, J = 8.2, 2.0 Hz, 1H, H-8), 8.26(s, 1H, H-2), 8.27(dt, J = 7.9, 1.6 Hz, 1H, H-5), 10.11(s, 1H, CHO).¹³C-NMR (100 MHz, CDCl₃-d, 36 °C, TMS) δ 21.7(Ar-CH₃), 113.3(C-8), 122.4(C-3), 122.6(C-5), 125.1(C-6), 126.30(C-4), 126.33(C-7), 127.3(C-11), 130.4(C-12), 134.3(C-9), 135.2(C-10), 136.3(C-2), 146.2 (C-13), 185.4(-CHO). HRMS m/z [M + H]⁺ calcd. for C₁₆H₁₄NO₃S: 300.0694, found 300.0700.

General procedure for preparation of ((1-tosyl-1H-indol-3-yl) methyl) amine carboxylate, (quinolin-3-ylmethyl)amine carboxylate or (1H-indol-3-yl)methyl)amine carboxylate derivatives (1a-g)

To a solution of 1-tosyl-1H-indole-3-carbaldehyde (2) (1 equiv., 1.0–2.0 mmol), 3-quinolinecaboxaldehyde (3) (1 equiv., 1.5-2.0 mmol) or 1H-indole-3-carbaldehyde (1) (1.0 equiv, 1.5-3.0 mmol) in anhydrous dichloromethane (25 mL), the corresponding amine carboxylate (1.1-2.0 equiv., 1.1–4.5 mmol) was added. The reaction mixture was stirred under nitrogen for 2–4 h at room temperature. Sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.3-2.0 equiv., 1.3–4.5 mmol) was then added to the mixture in the ice bath, and the reaction was allowed to be stir overnight at room temperature. After completion, the mixture was quenched with distilled water (5 mL), followed by the addition of a saturated solution of sodium bicarbonate (5 mL). The mixture was extracted with CH₂Cl₂ (2 × 20 mL), dried over anhydrous sodium sulfate Na₂SO₄ and concentrated under reduced pressure. The crude products were purified by column chromatography using silica gel with the appropriate eluent to afford the final products (1a-g).

Preparation of ethyl 4-((((1H-indol-3-yl)methyl)amino)methyl)benzoate (1a)

This compound was synthesized from 1H-indole-3-carbaldehyde (1) (1.0 equiv., 0.43 g, 3.0 mmol), ethyl 4-(aminomethyl)benzoate (1.1 equiv., 0.59 g, 3.3 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.5 equiv., 0.95 g, 4.5 mmol) following the general procedure and reaction conditions described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: methanol: ethyl acetate (20:1:1) to yield white crystals. The product was obtained in a 42% yield (0.39 g, 1.26 mmol), m.p = 90–91 °C, R_f = 0.30 (CHCl₃:MeOH: EtOAc 20:1:1). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.44(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 3.96(s, 2 H, H-CH₂N-benzylic), 4.04(d, J = 0.8 Hz, 2 H, H-CH₂N-pyrrol), 4.42(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 7.13(m, 1H, H-2), 7.17(ddd, J = 7.8, 7.0, 1.1 Hz, 1H, H-6), 7.24(ddd, J = 8.1, 7.0, 1.1 Hz, 1H, H-7), 7.37(dt, J = 8.1, 1.1 Hz, 1H, H-8), 7.47(d, J = 8.3 Hz, 2 H, H-11), 7.69(dt, J = 7.8, 1.1, 1H, H-5), 8.06(d, J = 8.3 Hz, 2 H, H-12), 8.45(s, 1H, H1-NH). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.4(OCH₂CH₃), 44.2(C-CH₂N-pyrrol), 52.9(C-CH₂N-benzylic), 61.0(OCH₂CH₃), 111.3(C-8), 114.4(C-3), 118.9(C-5), 119.6(C-6), 122.2(C-7), 122.9(C-2), 127.1(C-4), 128.1(C-11), 129.2(C-13), 129.7(C-12), 136.5(C-9), 145.7(C-10), 166.7(C = O). HRMS m/z [M + Hac-H]⁺ calcd. for C₂₁H₂₃N₂O₄ 367.1663, found 367.2593. [M + H]⁺ calcd. for C₁₉H₂₁N₂O₂, 309.1597, found 309.1585.

Preparation of ethyl 4-(4-((1H-indol-3-yl)methyl)piperazin-1-yl)benzoate (1b)

This compound was synthesized from 1H-indole-3-carbaldehyde (1) (1.0 equiv., 0.43 g, 3.0 mmol), ethyl 4-(1-piperazinyl)benzoate (1.2 equiv., 0.84 g, 3.6 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.3 equiv., 0.83 g, 3.9 mmol) following the general procedure and reaction conditions described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: methanol (10:1) to yield white crystals. The product was obtained in a 73% yield (0.8 g, 2.19 mmol), m.p = 175–176 °C, R_f = 0.50 (CHCl₃: MeOH, 10:1). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.41(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 2.69(dd, J = 6.7, 3.1 Hz, 4 H, H-10), 3.35(dd, J = 6.7, 2.8 Hz, 4 H, H-11), 3.82(s, 2 H, H-CH₂N), 4.37(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 6.86(d, J = 9.1 Hz, 2 H, H-13), 7.17(m, 1H, H-2), 7.18(m, 1H, H-6), 7.25(m, 1H, H-7), 7.38(dd, J = 8.0, 1.2 Hz, 1H, H-8), 7.80(dd, J = 7.8, 1.2 Hz, 1H, H-5), 7.96(d, J = 9.1 Hz, 2 H, H-14), 8.47(s, 1H, -NH). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.5(OCH₂CH₃), 47.5(C-11), 52.7(C-10), 53.6(C-CH₂N), 60.4(OCH₂CH₃), 111.2(C-8), 112.0(C-3), 113.6(C-13), 119.5(C-5), 119.6(C-6), 119.9(C-15), 122.1(C-7), 124.0(C-2), 128.0(C-4), 131.2(C-14), 136.3(C-9), 154.2(C-12), 166.9(C = O). HRMS m/z [M + H]⁺ calcd. for C₂₂H₂₆N₃O₂, 364.2020, found 364.2050.

Preparation of ethyl 4-((((1-tosyl-1H-indol-3-yl)methyl)amino)methyl)benzoate (1c)

This compound was synthesized from 1-tosyl-1H-indole-3-carbaldehyde (2) (1.0 equiv., 0.45 g, 1.50 mmol), ethyl 4-(aminomethyl)benzoate (1.1 equiv., 0.30 g, 1.65 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.5 equiv, 0.48 g, 2.25 mmol) following the general procedure and reaction conditions described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: methanol (98:2), yielding white crystals with a 45% yield (0.31 g, 0.7 mmol), m.p = 110–112 °C, R_f = 0.54 (CHCl₃:MeOH, 98:2). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.42(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 2.33(s, 3 H, Ar-CH3), 3.89(s, 2 H, H-16), 3.92(s, 2 H, H-14), 4.40(q, J = 7.1 Hz, 2 H, -OCH2CH3), 7.21(d, J = 8.2 Hz, 2 H, H-12), 7.27(ddd, J = 8.2, 7.8, 1.0 Hz, 1H, H-7), 7.34(ddd, J = 8.2, 7.3, 1.3 Hz, 1H, H-6), 7.43(d, J = 8.2 Hz, 2 H, H-18), 7.53(s, 1H, H-2), 7.58(dt, J = 7.8, 1.3,, 1H, H-8), 7.78(d, J = 8.2 Hz, 2 H, H-11), 8.02(m,1H, H-5), 8.04(d, J = 8.2 Hz, 2 H, H-19).¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.4(OCH₂CH₃), 21.6( Ar-CH₃), 43.9(C-14), 53.0(C-16), 60.9(OCH₂CH₃), 113.8(C-5), 119.8 (C-8), 121.3(C-3), 123.2(C-7), 123.9(C-2), 124.9(C-6), 126.8(C-11), 128.0(C-18), 129.3(C-20), 129.8(C-19), 129.9(C-12), 130.3(C-4), 135.3(C-10), 135.5(C-9), 144.9(C-13), 145.3(C-17), 166.6(C = O). HRMS m/z [2 M + H]⁺ calcd. for C₅₂H₅₃N₄O₈S₂ 925.3299, found 925.3390. [M + Na]⁺ calcd for C₂₆H₂₆NaN₂O₄S 485.1506, found 485.1539.

Preparation of ethyl 4-(((1-tosyl-1H-indol-3-yl) methyl) amino) benzoate (1d)

This compound was synthesized from 1-tosyl-1H-indole-3-carbaldehyde (2) (1.0 equiv., 0.45 g, 1.50 mmol), Ethyl 4-aminobenzoate (1.2 equiv., 0.30 g, 1.8 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (2 equiv., 0.63 g, 2.0 mmol) following the general procedure and reaction conditions described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: ethyl acetate (4:1), yielding white crystals with an 85% yield (0.57 g, 1.3 mmol), 154–155 °C, R_f = 0.85 (CHCl₃: EtOAc, 4 :1). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.40(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 2.35(s, 3 H, Ar-CH₃), 4.35(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 4.47(d, J = 1.2 Hz, 2 H, -NCH₂), 6.60(d, J = 8.7 Hz, 2 H, H-15), 7.19(d, J = 8.4 Hz, 2 H, H-12), 7.26(ddd, J = 8.5, 7.7, 1.0 Hz, 1H, H-7), 7.36(ddd, J = 8.5, 7.2, 1.3 Hz, 1H, H-6), 7.50(s, 1H, H-2), 7.53(m, 1H, H-8), 7.70(d, J = 8.4 Hz, 2 H, H-11), 7.89(d, J = 8.7 Hz, 2 H, H-16), 8.04(m, 1H, H-5). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.5(OCH₂CH₃), 21.6(Ar-CH₃), 39.2(N-CH₂), 60.3(OCH₂CH₃), 111.8(C-15), 114.0(C-5), 119.2(C-17), 119.5(C-8), 119.7(C-3), 123.4(C-7), 124.3(C-2), 125.1(C-6), 126.8(C-11), 129.6(C-4), 129.9 (C-12), 131.5(C-16), 135.9(C-10), 135.6(C-9), 145.1(C-13), 151.5(C-14), 166.9(C = O). HRMS m/z [M + Cl]⁺ calcd for C25H24ClN2O4S 483.1146, found 483.1073.

Preparation of ethyl 4-(4-((1-tosyl-1H-indol-3-yl) methyl) piperazin-1-yl) benzoate (1e)

This compound was synthesized from 1-tosyl-1H-indole-3-carbaldehyde (2) (1.0 equiv, 0.45 g, 1.50 mmol), ethyl 4-(1-piperazinyl)benzoate (1.2 equiv., 0.42 g, 1.8 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.3 equiv., 0.41 g, 1.95 mmol) following the general procedure and reaction conditions described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: ethyl acetate (3:1), yielding white crystals with an 78% yield (0.6 g, 1.2 mmol), m.p = 166–167 °C, R_f = 0.62 (CHCl₃:EtOAc, 3:1). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.39(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 2.33(s, 3 H, Ar-CH₃), 2.60(m, 4 H, H-14), 3.31(m, 4 H, H-15), 3.69(s, 2 H, H-NCH₂), 4.36(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 6.86(d, J = 9.1 Hz, 2 H, H-17), 7.21(d, J = 8.4 Hz, 2 H, H-12), 7.25(m, 1H, H-7), 7.35(ddd, J = 8.4, 7.2, 1.3 Hz, 1H, H-6), 7.56(s, 1H, H-2), 7.74(dt, J = 7.8, 1.3 Hz, 1H, H-8), 7.80(d, J = 8.4 Hz, 2 H, H-11), 7.96(d, J = 9.1 Hz, 2 H, H-18), 8.04(dt, J = 8.4, 1.0 Hz, 1H, H-5). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.5(OCH₂CH₃), 21.6(Ar-CH₃), 47.5(C-15), 52.8(C-14), 53.4(C-NCH₂), 60.4(OCH₂CH₃), 113.6(C-17), 113.7(C-5), 119.2(C-19), 120.0(C-3), 120.6(C-8), 123.2(C-7), 124.9(C-2), 124.9(C-6), 126.8(C-11), 129.9(C-12), 130.8(C-4), 131.2(C-18), 135.3(C-10), 135.5(C-9), 145.0(C-13), 154.1(C-16), 166.7 (C = O). HRMS m/z [M + H]⁺ calcd. for C₂₉H₃₂N₃O₄S 518.2114, found 518.2221. [M + Na]⁺ calcd for C₂₉H₃₁NaN₃O₄S 540.1928, found 540.1935.

Preparation of ethyl 4-((quinolin-3-ylmethyl)amino)benzoate (1f)

This compound was synthesized from 3-quinolinecaboxaldehyde (3) (1 equiv., 0.314 g, 2.0 mmol), ethyl 4-aminobenzoate (1.1 equiv., 0.36 g, 2.2 mmol) and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.5 equiv., 0.64 g, 3.0 mmol) following the general procedure and reaction conditions described above^40–42. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: ethyl acetate (4:1), yielding white crystals with an 66% yield (0.4 g, 1.3 mmol), m.p = 135–136 °C, R_f = 0.38(CHCl₃:EtOAc, 4:1). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.35(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 4.31(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 4.58(d, J = 4.5 Hz, 2 H, -NCH₂), 4.92(s, br, -NH), 6.64(d, J = 8.8 Hz, 2 H, H-12), 7.55(ddd, J = 8.1, 6.9, 1.2 Hz, 1H, H-7), 7.71(ddd, J = 8.4, 6.9, 1.5 Hz, 1H, H-8), 7.76(dd, J = 8.1, 1.5 Hz, 1H, H-6), 7.89(d, J = 8.8 Hz, 2 H, H-13), 8.07(dd, J = 2.2, 1.0 Hz, 1H, H-4), 8.11(dt, J = 8.5, 1.0, 1.0 Hz, 1H, H-9), 8.90(d, J = 2.2 Hz, 1H, H-2). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.5(OCH₂CH₃), 45.4(-NCH₂), 60.3(OCH₂CH₃), 111.8(C-12), 119.5(C-14), 127.1(C-7), 127.7(C-6), 127.9(C-5), 129.1(C-9), 129.2(C-8), 131.3(C-3), 131.6(C-13), 134.1(C-4), 147.5(C-10), 150.3(C-2), 151.3(C-11), 166.8(C = O). HRMS m/z [M + H]⁺ calcd for C₁₉H₁₉N₂O₂ 307.1441, found 307.1473.

Preparation of ethyl 4-(4-(quinolin-3-ylmethyl)piperazin-1-yl)benzoate (1g)

This compound was synthesized from 3-quinolinecaboxaldehyde (3) (1 equiv., 0.236 g, 1.50 mmol), ethyl 4-(1-piperazinyl)benzoate (1.2 equiv., 0.42 g, 1.8 mmol), and sodium triacetoxyborohydride ((CH₃COO)₃BHNa) (1.3 equiv., 0.42 g, 1.95 mmol) following the general procedure and reaction conditions as described above. The crude product was purified by column chromatography using silica gel with an eluent system of chloroform: methanol (94:6), yielding white crystals with a 63% yield (0.35 g, 1 mmol), m.p = 153–155 °C, R_f = 0.68 (CHCl₃:MeOH, 94:6). ¹H-NMR (400 MHz, CDCl₃-d, TMS) δ 1.34(t, J = 7.1 Hz, 3 H, -OCH₂CH₃), 2.60(t, J = 5.0 Hz, 4 H, H-11), 3.29(t, J = 5.0 Hz, 4 H, H-12), 3.69(s, 2 H, CH₂N), 4.30(q, J = 7.1 Hz, 2 H, -OCH₂CH₃), 6.81(d, J = 8.7 Hz, 2 H, H-14), 7.52(t, J = 8.3 Hz, 1H, H-7), 7.68(ddd, J = 8.4, 6.6, 1.4 Hz, 1H, H-8), 7.79(dd, J = 8.3, 1.4 Hz, 1H, H-6), 7.91(d, J = 8.7 Hz, 2 H, H-15), 8.05(d, J = 2.1 Hz, 1H, H-4), 8.11(d, J = 8.4 Hz, 1H, H-9), 8.92(d, J = 2.1 Hz, 1H, H-2). ¹³C-NMR (100 MHz, CDCl₃-d, TMS) δ 14.5(OCH₂CH₃), 47.5(C-12), 52.8(C-11), 60.3(OCH₂CH₃), 60.4(NCH₂), 113.6(C-14), 120.1(C-16), 126.8(C-7), 127.6(C-6), 127.9(C-5), 129.2(C-8), 129.2(C-9), 130.7(C-3), 131.1(C-15), 135.7(C-4), 147.6(C-10), 152.0(C-2), 154.0(C-13), 166.6(C = O). HRMS m/z [M + H]⁺ calcd. for C₂₃H₂₆N₃O₂ 376.2020, found 376.2038.

General procedure for preparation of corresponding hydroxamic acid derivatives of (2a-g)

To a solution of benzoate derivatives (1a-g) (1 equiv., 0.27–0.45 mmol) in a mixture of anhydrous dichloromethane and methanol (1:2, 6 mL) at 0 °C, hydroxylamine (50 wt % in water, 7.99–13.37 mmol, 30 equiv.) was added, followed by the addition of sodium hydroxide (1.08–2.25 mmol, 4.0–5.0 eq). The reaction mixture was allowed to warm to room temperature and stirred for 24 h. Then the solvent was removed under reduced pressure, and the obtained solid was dissolved in water. The pH of the solution was adjusted to 7 using 1 N HCl. The resulting precipitate was filtered and dried under a high vacuum, and then the crude products were purified by column chromatography using silica gel with the appropriate eluent or soaking with dichloromethane to give the final products (2a-g).

Preparation of 4-((((1H-indol-3-yl)methyl)amino)methyl)-N-hydroxybenzamide (2a)

Compound 2a was synthesized from 1a (0.10 g, 0.32 mmol, 1 equiv.), hydroxylamine (50 wt % in water, 0.64 mL, 0.32 g, 9.60 mmol, 30 equiv.) and sodium hydroxide (0.05 g, 1.28 mmol, 4.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by soaking with dichloromethane to yield compound 2a as white crystals. The yield was 52% (49 mg, 0.17 mmol), m.p = 158–160 °C. ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ 3.80(s, 2 H, H-CH₂N-benzylic), 3.86(d, J = 0.8 Hz, 2 H, H-CH₂N-pyrrol), 6.99(s, 1H, H-2), 7.08(m, 1H, H-6), 7.25(m, 1H, H-7), 7.36(m, 1H, H-8), 7.45(m, 2 H, H-11), 7.61(m, 1H, H-5), 7.74(m, 2 H, H-12), 10.91(s, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 44.2(C-CH₂N-pyrrol), 52.4(C-CH₂N-benzylic), 111.8(C-8), 113.9(C-3), 118.7(C-2), 119.3(C-5), 121.4(C-6), 124.0(C-7), 127.2(C-12), 127.5(C-4), 128.3(C-11), 131.5(C-13), 136.9(C-9), 144.9(C-10), 164.7(C = O). HRMS m/z 294.1250 [M-H]⁺ (calc for C₁₇H₁₆N₃O₂, 294.1242), 130.0656 [M-C₈H₉N₂O₂]⁺ (calc for C₉H₈N, 130.0657).

Preparation of 4-(4-((1H-indol-3-yl)methyl)piperazin-1-yl)-N-hydroxybenzamide (2b)

Compound 2b was synthesized from 1b (0.10 g, 0.27 mmol, 1 equiv.), hydroxylamine (50 wt % in water, 0.52 mL, 0.26 g, 8.25 mmol, 30 equiv.) and sodium hydroxide (0.04 g, 1.08 mmol, 4.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by soaking with dichloromethane to afford 2b as white crystals. The yield was 53% (50 mg, 0.14 mmol), m.p = 155–157 °C. ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ 2.53(m, 4 H, H-10), 3.20(m, 4 H, H-11), 3.72(s, 2 H, H-CH₂N), 6.87(d, J = 8.9 Hz, 2 H, H-13), 7.00(t, J = 7.4 Hz, 1H, H-6), 7.08(t, J = 7.5 Hz, 1H, H-7), 7.27(s, 1H, H-2), 7.37(d, J = 8.0 Hz, 1H, H-8), 7.57(d, J = 8.3 Hz, 2 H, H-14), 7.62(d, J = 7.9 Hz, 1H, H-5), 8.81(brs, 1H, -(NH)C = O), 10.93(s, 1H, NH(Ar)), 10.96(s, 1H, OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 47.7(C-11), 52.7(C-10), 53.6(C-CH₂N), 110.9(C-3), 111.8(C-8), 114.1(C-13), 118.9(C-6), 119.5(C-5), 121.4(C-7), 122.1(C-15), 125.2(C-2), 128.1(C-4), 128.5(C-14), 136.7(C-9), 153.2(C-12). HRMS m/z [M-H]⁺ calcd for C₂₀H₂₁N₄O₂ 349.1670, found 349.1678. [M + Cl³⁵]⁺ calcd for C₂₀H₂₂ClN₄O₂ 385.1437, found 385.1442.

Preparation of N-hydroxy-4-((((1-tosyl-1H-indol-3-yl)methyl)amino)methyl) benzamide (2c)

Compound 2c was synthesized from 1c (0.20 g, 0.43mmol, 1equiv.), hydroxylamine (50 wt % in water, 0.85mL, 0.42 g, 12.90mmol, 30equiv.) and sodium hydroxide (0.086 g, 2.15 mmol, 5.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by using silica gel with an eluent system of dichloromethane: methanol (98.7:1.3) to yield 2c as light-pink crystals. The yield was 37% (71 mg, 0.16 mmol), m.p = 151–153 °C, R_f = 0.49 (CH₂Cl₂:MeOH (87:13)). ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ 2.28(s, 3 H, Ar-CH₃), 3.75(s, 2 H, -NCH₂-16), 3.80(s, 2 H, -NCH₂-14), 7.25(t, J = 7.5 Hz, 7.5 Hz, 1H, H-7), 7.34(d, J = 7.9 Hz, 2 H, H-12), 7.35(t, J = 7.4 Hz, 1H, H-6), 7.41(d, J = 7.8 Hz, 2 H, H-18), 7.65(d, J = 7.5 Hz, 1H, H-8), 7.67(s, 1H, H-2), 7.75(d, J = 7.8 Hz, 2 H, H-19), 7.83(d, J = 7.9 Hz, 2 H, H-11), 7.93(d, J = 8.2 Hz, 1H, H-5), 9.04(brs, 1H, -((CO)NH), 11.17(brs, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 21.5( Ar-CH₃), 43.5(N-CH₂-14), 52.4(N-CH₂-16), 113.7(C-5), 120.8(C-8), 122.5(C-3), 123.7(C-7), 124.5(C-2), 125.2(C-6), 127.1(C-11), 127.3(C-19), 128.3(C-18), 130.6(C-12), 130.6(C-20), 131.6(C-4), 134.6(C-10), 135.2(C-9), 144.5(C-13), 145.8(C-17), 164.6(C = O). HRMS m/z [M-H]⁺ calcd. for C₂₄H₂₂N₃O₄S 448.1336, found 448.1342. [2 M + H]⁺ calcd for C₄₈H₄₅N₆O₈S₂ 897.2740, found 897.2751.

Preparation of N-hydroxy-4-(((1-tosyl-1H-indol-3-yl)methyl)amino)benzamide (2d)

Compound 2d was synthesized from 1d (0.20 g, 0.45 mmol, 1.0 equiv.), hydroxylamine (50 wt % in water, 0.88 mL, 0.44 g, 13.37 mmol, 30 equiv.) and sodium hydroxide (0.09 g, 2.25 mmol, 5.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by using silica gel with an eluent system of dichloromethane: methanol (94:6) to yield 2d as light-pink crystals (CH₂Cl₂:MeOH(94:6)). The yield was 34% (66 mg, 0.15 mmol), m.p = 183–184 °C, R_f = 0.32 (CH₂Cl₂:MeOH(94:6)). ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ 2.29(s, 3 H, Ar-CH₃), 4.43(d, J = 5.6 Hz, 2 H, -NCH₂), 6.65(d, J = 8.4 Hz, 2 H, H-15), 6.71(t, J = 6.2 Hz, 1H, -NH-CH₂), 7.26(t, J = 7.5 Hz, 1H, H-7), 7.33(d, J = 8.2 Hz, 2 H, H-12), 7.33(t, J = 7.2 Hz, 1H, H-6), 7.53(d, J = 8.3 Hz, 2 H, H-16), 7.72(d, J = 7.5 Hz, 1H, H-8), 7.75(d, J = 8.3 Hz, 2 H, H-11), 7.77(s, 1H, H-2), 7.91(d, J = 8.2 Hz, 1H, H-5), 8.72(brs, 1H, -((CO)NH), 10.81(brs, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 21.5( Ar-CH₃), 38.2(N-CH₂), 111.9(C-15), 113.8(C-5), 120.00(C-17), 120.9(C-8), 121.2(C-3), 123.8(C-7), 125.1(C-2), 125.4(C-6), 127.0(C-11), 128.6(C-16), 130.3(C-4), 130.60(C-12), 134.4 (C-10), 135.2(C-9), 145.8(C-13), 151.3(C-14). HRMS m/z [M + Na]⁺ calcd for C₂₃H₂₁N₃NaO₄S 458.1151, found 458.1156. [2 M + H]⁺ calcd. for C₄₆H₄₃N₆O₈S₂ 871.2584, found 871.2610.

Preparation of N-hydroxy-4-(4-((1-tosyl-1H-indol-3-yl)methyl)piperazin-1-yl)benzamide (2e)

Compound 2e was synthesized from 1e (0.15 g, 0.29 mmol, 1 equiv.), hydroxylamine (50 wt % in water, 0.57 mL, 0.287 g, 8.7 mmol, 30 equiv.) and sodium hydroxide (0.058 g, 1.45 mmol, 5.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by using silica gel with an eluent system of dichloromethane: methanol (90:10) to yield 2e as light-pink crystals. The yield was 53% (77 mg, 0.15 mmol), m.p = 180–182 °C, R_f = 0.43 (CH₂Cl₂:MeOH(90:10)). ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ 2.26(s, 3 H, Ar-CH₃), 2.49(t, J = 8.2 Hz, 4 H, H-14), 3.20(t, J = 4.8 Hz, 4 H, H-15), 3.65(s, 2 H, H-CH₂N), 6.91(d, J = 8.5 Hz, 2 H, H-17), 7.25(t, J = 7.4 Hz, 1H, H-7), 7.33(d, J = 7.9 Hz, 2 H, H-12), 7.33(t J = 7.6 Hz, 1H, H-6), 7.68(d, J = 8.3 Hz, 2 H, H-18), 7.72(s, 1H, H-2), 7.73(d, J = 8.9 Hz, 1H, H-8), 7.84(d, J = 7.8 Hz, 2 H, H-11), 7.94(d, J = 8.2 Hz, 1H, H-5), 8.86(brs, 1H, -NH), 10.97(brs, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 21.4( Ar-CH₃), 47.6(C-15), 52.7(C-CH₂N,14), 113.7(C-5), 114.2(C-17), 119.8(C-19), 121.2(C-8), 122.2(C-3), 123.8(C-7), 125.3(C-2), 125.7(C-6), 127.1(C-11), 128.6(C-18), 130.6(C-12), 131.1(C-4), 135.2(C-10), 135.6(C-9), 145.8(C-13), 153.2(C-16), 164.8(C = O). HRMS m/z. [M-H⁺]⁺ calcd for C₂₇H₂₇N₄O₄S⁻ 503.1758, found 503.1766.

Preparation of N-hydroxy-4-((quinolin-3-ylmethyl)amino)benzamide (2f)

Compound 2f was synthesized from 1f (0.10 g, 0.32 mmol, 1 equiv.), hydroxylamine (50 wt % in water, 0.63 mL, 0.32 g, 9.60 mmol, 30 equiv.) and sodium hydroxide (0.05 g, 1.28 mmol, 4.0 equiv.) following the general procedure and reaction conditions described above. The crude product was purified by soaking with dichloromethane to afford 2f as white crystals. The yield was 53% (49 mg, 0.17 mmol), m.p = 201–203 °C. ¹H- NMR (400 MHz, DMSO-d₆, TMS) δ 4.56(d, J = 6.0 Hz, 2 H, -NCH₂), 6.65(d, J = 8.4 Hz, 2 H, H-12), 6.94(t, J = 6.0 Hz, 1H, -NH-CH₂), 7.53(d, J = 8.4 Hz, 2 H, H-13), 7.59(t, J = 7.5 Hz, 1H, H-7), 7.73(t, J = 8.4 Hz, 1H, H-8), 7.94(d, J = 8.1 Hz, 1H, H-6), 8.01(d, J = 8.4 Hz, 1H, H-9), 8.25(d, J = 2.2 Hz, 1H, H-4), 8.69(brs, 1H, -NH), 8.93(d, J = 2.2 Hz, 1H, H-2), 10.77(s, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 44.4(-NCH₂), 111.9(C-12), 120.3(C-14), 127.3(C-7), 128.0(C-5), 128.3(C-6), 128.8(C-13), 129.2(C-9), 129.6(C-8), 133.1(C-3), 133.9(C-4), 147.3(C-10), 151.2(C-11), 151.4(C-2), 165.4(C = O). HRMS m/z [M-H]⁺ calcd. for C₁₇H₁₄N₃O₂ 292.1086, found 292.1084.

Preparation of N-hydroxy-4-{4-[(quinolin-3-yl)methyl]piperazin-1-yl}benzamide (2g)

Compound 2g was synthesized from 1g (0.10 g, 0.27 mmol, 1 equiv.), hydroxylamine (50 wt % in water, 0.52 mL, 0.26 g, 7.99 mmol, 30 equiv.) and sodium hydroxide (0.04 g, 1.08 mmol, 4.0 equiv) following the general procedure and reaction conditions described above. The crude product was purified by soaking with dichloromethane to afford 2g as white crystals. The yield was 73% (71 mg, 0.20 mmol), m.p = 215–216 °C. ¹H-NMR (400 MHz, DMSO-d₆, TMS) δ = 2.55(s, 4 H, H-11), 3.26(s, 4 H, H-12), 3.74(s, 2 H, H-CH₂N), 6.93(d, J = 8.6 Hz, 2 H, H-14), 7.62(t, J = 7.3 Hz, 1H, H-7), 7.64(d, J = 8.3 Hz, 2 H, H-15), 7.76(t J = 6.6 Hz, 1H, H-8), 7.99(d, J = 8.3 Hz, 1H, H-6), 8.03(d, J = 8.3 Hz, 1H, H-9), 8.26(s, 1H, H-4), 8.84(brs, 1H, -NH), 8.90(s, 1H, H-2), 10.96(s, 1H, -OH). ¹³C-NMR (101 MHz, DMSO-d₆, TMS) δ 47.6(C-12), 52.8(C-11), 59.8(NCH₂), 114.2(C-14), 122.4(C-16), 127.2(C-7), 127.9(C-5), 128.4(C-6), 128.5(C-9), 129.2(C-8), 129.6(C-15), 131.4(C-3), 135.8(C-4), 147.4(C-10), 152.5(C-2), 153.2(C-13). HRMS m/z [M + 2 H]⁺ calcd. for C₂₁H₂₄N₄O₂: 364.1888, found: 364.1885.

Pharmacological/biological assays

The cytotoxic activity

The cytotoxic activity of all pure compounds was determined towards a panel of four human solid tumour cell lines: melanoma (SK-MEL), epidermal carcinoma (KB), breast carcinoma (BT-549), and ovarian carcinoma (SK-OV-3). Moreover, non-cancer kidney cell lines (LLC-PK1 and VERO) were also employed to determine if the anti-cell proliferative activity of these compounds was selective for the tested tumour cell lines. All cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The cells were seeded in 96-well plates (10,000 cells/well) and incubated for 24 h. All ligands were dissolved in DMSO, diluted in media, and added to the cells at concentrations of 80, 40, 20, and 10 µM. After incubating for 48 h, cell viability was determined using a tetrazolium dye WST-8, which is converted to a water-soluble formazan product in the presence of 1-methoxy PMS by the activity of cellular enzymes. The colour of the formazan product was measured at 450 nm on a plate reader. Doxorubicin was used as a positive control for the cytotoxicity assay, and DMSO (0.25%) was used as the vehicle control. The IC₅₀ values were obtained from concentration-response curves. The values are represented as mean ± standard deviation (n = 3).

Enzymatic in vitro HDAC inhibitory activity

Recombinant human HDAC1, HDAC2, HDAC3/NCOR1, and HDAC6 were purchased from ENZO Life Sciences AG (Lausen, CH). Barinka et al.³² produced recombinant drHDAC10 wild type. Recombinant human HDAC8 was produced by Romier et al.⁴³. In vitro testing of the inhibitors in an enzymatic assay was carried out as described in previous publications by us^32,43–46. Inhibitory activity for HDAC1, HDAC2, HDAC3, and HDAC6 was determined using a discontinuous assay with a substrate peptide derived from p53 (Ac-RHKK(Ac)-AMC) in a 384 well-plate. The enzyme at final concentrations (10 nM HDAC1, 3 nM HDAC2 and HDAC3, and 1 nM HDAC6,) and the inhibitors at various concentrations were incubated for 5 min in assay buffer (50 mM HEPES, 150 mM NaCl, 5 mM MgCl2, 1 mM TCEP and 0.2 mg/mL BSA, pH 7.4 adjusted with NaOH). The reaction started with the addition of substrate (20 µM HDAC1-3 and 5 µM HDAC6). Afterwards, the fluorescence was developed with a 0.5 mg/mL trypsin solution (final concentration), and the fluorescence readout was done with an Envision 2104 Multilabel Plate Reader (PerkinElmer, Waltham, USA) with λ_Ex = 380 ± 10 nm and λ_Em = 430 ± 8 nm.

For HDAC8 the fluorogenic peptide derivate Abz-SRGGK(thio-TFA)FFRR-NH2 was applied as described in⁴³. For HDAC10 a spermidine derivative Ac-spermidine-AMC was applied. The HDAC8 assay was performed in the same assay buffer as described above. For HDAC10 the assay buffer was 20 mM HEPES, pH 7.4, and 0.5 mg/mL BSA. The enzyme (1.5 nM HDAC8 or 5 nM HDAC10 final concentration) was incubated for 5 min with various concentrations of the inhibitor. The reaction was started with the addition of 50 µM substrate and the readout was done continuously with an Envision 2104 with λ_Ex = 330 ± 75 nm, and λ_Em = 430 ± 8 nm.

For HDAC1, 2, and 3 a fluorogenic peptide derived from p53 (Ac-RHKK(Acetyl)-AMC) was applied. The measurements were performed in assay buffer (50 mM HEPES, 150 mM NaCl, 5 mM MgCl2, 1 mM TCEP, and 0.2 mg/mL BSA, pH 7.4 adjusted with NaOH) at 37 °C. An Envision 2104 Multilabel Plate Reader (PerkinElmer, Waltham, MA), with an excitation wavelength of 380 ± 8 nm and an emission wavelength of 430 ± 8 nm was considered to measure the fluorescence intensity. For HDAC6 the substrate (Abz-SRGGK(thio-TFA)FFRR-NH2) was applied as described before⁴³. The HDAC10 inhibition assay was performed as described before²⁹ using Ac-spermidine-AMC as substrate. The idea of this discontinuous assay is the HDAC10 mediated generation of a primary amino function. The released primary amine is reacted wirh 2,3-Naphthalenedicarboxaldehyde (NDA) resulting in an NDA-spermidin-AMC derivative which differences in the AMC fluorescence intensities as compared to Ac-spermidin AMC. The assay was performed in black 96-well plates (PerkinElmer, OptiPlateTM-96 F). The compounds to be tested were incubated for 25 min at 25 °C. Before measuring fluorescence (POLARstar plate reader, λex = 330 nm, λem = 390 nm) each well was filled with 200 µL stop solution containing 16mM NDA). (for assay details please refer to²⁹. The enzyme inhibition of HDAC8 was determined by using a homogenous fluorescence assay⁴³. The enzyme was incubated for 90 min at 37 °C, with the fluorogenic substrate ZMAL (Z(Ac)Lys-AMC) in a concentration of 10.5 µM and increasing concentrations of inhibitors. Fluorescence intensity was measured at an excitation wavelength of 390 nm and an emission wavelength of 460 nm in a microtiter plate reader (BMG Polarstar).

Molecular docking

Molecular docking

Available crystal structures of drHDAC10, hsHDAC1, hsHDAC6 and hsHDAC8 were downloaded (PDB ID: 6UHU, PDB ID: 5ICN, PDB ID: 5EDU, and PDB ID: 2V5X respectively) from the Protein Data Bank (PDB;www.rscb.org)⁴⁷. All ligands and all protein-ligand complexes were prepared using similar methods as published before^29,30. Validation of the molecular docking method was performed by re-docking the ligands co-crystallized in HDAC10 as reported in our previous publication³⁰. For the preparation of the proteins, the wizard implementation of the Schrödinger version 2019.1 was used with the following steps: hydrogen atoms addition, protonation states assignment, and finally, restrained energy minimization using the OPLS force field 2005. The ligand structure was generated using the 2D Sketcher of Schrödinger (version 2019.1). Afterward, the LigPrep tool (Schrödinger version 2019.1) was used for the preparation of the ligands with energy minimization using the OPLS2005 force field^48,49. 64 conformers for each ligand were generated with the ConfGen tool (Schrödinger version 2019.1). Preparation of the receptor grid for the docking procedure was performed by assigning the co-crystallized ligands as the centroid of the grid box in each PDB crystal structure using the receptor grid preparation module in Schrödinger (version 2019.1). Lastly, docking of all generated conformers was done in the Standard Precision mode using the Glide (Schrödinger-release 2019.1).

Molecular dynamics (MD) simulations

AMBER22 was used to perform GPU-based MD simulations. PDB4Amber command was used for the preparation of the protein structures for further usage within the tLEaP program. Topology and force field parameters of the ligands were assigned with Antechamber⁵⁰ package using the second generation general Amber force field (GAFF2) and the semi-empirical AM1-BCC (Austin Model1 with bond charge correction) as atomic charge method^51,52. Afterwards, the protein-ligand complexes were created with the AMBER22 tLEaP module. The second-generation general AMBER force field (GAFF2) was used as the ligand force field while force field 14 Stony Brook-ff14SB was used for the protein structures^53,54. For the catalytic Zn2+, the 12-6-4 LJ-type nonbonded ion model was applied⁵⁵. After combining the protein and the ligands, complexes were solvated by transferable intermolecular potential 3P-TIP3P water model as an octahedral box around the protein with a 10 Å margin. Then Na + and Cl- ions were added to neutralize the whole system. Parameter/topology files of the entire system were created with tLEaP and the files were used as a starting point for the MD simulations. The solvated systems were first subjected to two energy minimization steps involving 1000 cycles of steepest descent followed by 2000 cycles of conjugate gradient totaling 3000 cycles of minimization. In the first energy minimization step, only the solvent molecules and counter-ions (Na + and Cl-) were minimized while applying constraints with a force constant of 10 kcal*mol-1*Å-2 to the proteins, ligands, and zinc ion. In the second minimization step, the whole system was minimized without constraints. Subsequently, the systems were heated from 0 to 300 K over 100 ps while applying the same constraints on the solute as in minimization step1. Constant volume periodic boundary was set to equilibrate the temperature of the system by Langevin thermostat using a collision frequency of 2 ps-1. Subsequently, a pressure equilibration routine with a constant pressure of 1 bar and at 300 K was performed for 100 ps. Finally, 100 ns free molecular dynamic simulations with the time step of 2 fs were applied utilizing the Particle Mesh Ewald method⁵⁶. The system temperature was kept at 300 K with a Langevin thermostat using 2 ps-1 collision frequency and the pressure of the system was maintained at 1 bar with the usage of isotropic position scaling and a relaxation time of 2 ps. Implementation of the SHAKE algorithm was done to constrain all bonds containing hydrogens. A total of 1000 frames were written for 100 ns long MDs. Simulations of the prepared crystal structures of humanized drHDAC10 (PDB IDs; 7U6B, 7U69, 7U6A, 7U69) X-ray structures and the docking pose of ligand 1 in drHDAC10 (PDB ID:6UHU) were repeated three times from the minimization steps with non-identical random seeds. The analysis of the MD trajectories was performed using the CPPTRAJ module in AMBER 22. The RMSD of the protein was calculated for the backbone atoms using CPPTRAJ. RMSF of the ligand as well as distances between the hydroxamate group oxygens and the zinc ion were calculated to further examine the stability of the protein-ligand complexes using CPPTRAJ. Additionally, the obtained trajectories were clustered based on the ligand-heavy atoms using the K-means algorithm and pytraj implementation of AMBER.

PAINS filter

Inhibitors described herein were filtered for pan-assay interference compounds (PAINS). For this purpose, PAINS1, PAINS2, and PAINS3 filters, as implemented in Schroedinger’s Canvas program (Schrödinger version 2019.1), were employed. None of the compounds was flagged as PAINS.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, A.T., W.S., S.R. and S.A.; methodology, A.T., W.S. and S.R.; software, M.S.; validation, W.S., A.T., S.A. and S.R.,.; formal analysis, M.S., T.Y., M.Z., D.R. and C.B.; investigation, A.T, M.S., T.Y., M.Z., D.R. and C.B.; resources, S.R. and W.S.; data curation, A.T and W.S.; writing—original draft preparation, A.T.; writing—review and editing, S.A., M.S. and S.R.,; visualization, W.S and A.T.; supervision, A.T, W.S. and S.R.; project administration, A.T and W.S.; funding acquisition, S.R and W.S. All authors have read and agreed to the published version of the manuscript.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.