New 5-Aryl-Substituted 2-Aminobenzamide-Type HDAC Inhibitors with a Diketopiperazine Group and Their Ameliorating Effects on Ischemia-Induced Neuronal Cell Death

We previously synthesized new 5-thienyl-substituted 2-aminobenzamide-type HDAC1, 2 inhibitors with the (4-ethyl-2,3-dioxopiperazine-1-carboxamido) methyl group. K-560 (1a) protected against neuronal cell death in a Parkinson’s disease model by up-regulating the expression of XIAP. This finding prompted us to design new K-560-related compounds. We examined the structure activity relationship (SAR) for the neuronal protective effects of newly synthesized and known K-560 derivatives after cerebral ischemia. Among them, K-856 (8), containing the (4-methyl-2,5-dioxopiperazin-1-yl) methyl group, exhibited a promising neuronal survival activity. The SAR study strongly suggested that the attachment of a monocyclic 2,3- or 2,5-diketopiperazine group to the 2-amino-5-aryl (but not 2-nitro-5-aryl) scaffold is necessary for K-560-related compounds to exert a potent neuroprotective effect.


Ameliorating effects of HDAC inhibitors on cell damage after oxygen-glucose deprivation.
Primary rat cortical neurons prepared as indicated in the section of Materials and Methods were pre-incubated with a tested compound for 1 h and then subjected to oxygen-glucose deprivation (OGD) injury for 3 h. Neuronal cell death was assessed by performing a lactate dehydrogenase (LDH) assay 24 or 48 h after ischemia. Figure 3A shows the percentage of neuronal cell death resulting from incubation with K-560 (1a), K-562 (3), K-563 (4),    or K-564 (5) at 1 μM, and Fig. 3B depicts the percentage of cell death resulting from incubation with K-852 (6), K-854 (7), K-856 (8), or K-560 (1a) at 1 μM. The results obtained revealed that 3, 5, 6, and 8, as well as 1a, exerted protective effects (LDH assay in 24 h) against OGD-induced damage. Since K-856 (8) seemingly exerted the most promising protective effects (though with no significant difference between them) (Fig. 3B), its protective activity was compared with that of 1a by performing the LDH assay 48 h after ischemia (Fig. 3C). It is noteworthy that both K-560 (1a) and K-856 (8) maintained a lower percentage of neuronal cell death than the control (0.1%DMSO) even in 48 h. Figure 3D shows the percentage death of neuronal cells which were incubated with the 2-nitro form 4 (1b) of K-560 (1a), K-561 (2a), or the 2-nitro form 4 (2b) of 2a. None of these compounds exerted protective effects and, instead, they decreased cell viability compared with the control (by the LDH assay 24 h after ischemia). In order to monitor the changes in the protective effects of 1a and 8 with the concentrations, cortical neurons were incubated with them at ranging from 0.1 to 10 μM before OGD-induced injury, respectively, (Fig. 4A). It was substantiated that, although dose-dependency was not clearly observed, 1a and 8 reduced the percentages of cell death at 10-0.3 μM and at 3-0.1 μM (with significant differences vs. control), respectively, and that 8 was more effective than 1a at the lowest concentration (0.1 μM). Furthermore, it is suggested from the comparative experiment (Fig. 4B) with the clinically-used HDAC inhibitors, FK228 (HDAC1,2 selective inhibitor), MS-275 (HDAC Class I inhibitor), as well as 1a and 8, that 1a and 8 are more potent than FK228, since FK228 was more toxic even at 0.1 μM than the control, whereas they have nearly the same level of protective activity as MS-275. Figure 4C indicates that none of the 4-methyl-substituted 3-oxopiperazine form: OP-857 (9), the 4-unsubstituted 3-oxopiperazine-form: OP-858 (10) and the 4-ethylpiperazine-form: OP-859 (11) was effective with significant difference from the control, though compound 10 looked more effective than 9 and 11.

Discussion
Three hydroxamate-type pan-HDAC inhibitors have been approved by the US Food and Drug Administration (FDA) to date for the treatment of the following cancers: SAHA for the advanced forms of cutaneous T-cell lymphoma (CTCL) 5,6,32 ; Belinostat (Beleodaq) for refractory peripheral T-cell lymphoma (PTCL) 33 ; Panobinostat (Farydak) for the combinatorial treatment of multiple myeloma 34 . Furthermore, the cyclic depsipeptide FK228 (Romidepsin) was licensed by the US FDA for the treatment of CTCL 5,6,32 , and the 2-aminobenzamide-type HDAC inhibitor MS-275 (Entinostat), for breakthrough therapy in the treatment of advanced breast cancers 35 .
On the other hand, HDAC inhibitors have emerged as an attractive therapeutic candidate for neurodegeneration in the last decade 36,37 because therapeutic options for neuroprotective therapies in subacute or chronic ischemic stroke currently remain very limited. It is anticipated that HDAC inhibitors may act against chronic neurodegeneration without tumorigenesis because they have been developed as treatment drugs for various cancers. Endovascular therapy after intravenous t-PA has been reported to lead to better functional outcomes [38][39][40][41] . Although pan-HDAC inhibitors such as SAHA and TSA were previously shown to prevent ischemic cell damage and improve functional recovery, they may cause side effects due to the non-selectivity of their HDAC isoforms. Since HDAC1, 2, and 3 are strongly expressed in the central nervous system, a promising strategy may be to develop selective inhibitors of these enzymes in order to assess their effectiveness as a new therapeutic agent for ischemic stroke. We were encouraged by the result showing that the HDAC1, 2-selective inhibitor K-560 (1a) exerted neuroprotective effects [23][24][25] . Therefore, we attempted to synthesize K-560-related compounds in order to assess their inhibitory activities against HDAC1, 2, 3, 8 (Class I) and 6 (Class II) and ameliorating effects on in vitro cerebral ischemia and also to examine their SAR. The results obtained revealed that all the compounds tested preferentially inhibited HDAC1 and 2 though with a 10-26-fold selectivity for HDAC1 over HDAC2. According to the SAR study between 5-thienyl-substituted 2-aminobenzamide HDAC1,2 inhibitors, even truncation (exemplified by 13) of the capping group and linker domain of compound 12 maintained a high selectivity for HDAC1 and HDAC2 versus HDAC3 and HDAC8, but reduced the preference for HDAC1 versus HDAC2 42 (Fig. 6). It is well known that the cap region of HDAC is predominantly responsible for selectivity 43 . Based on these findings, it is seems reasonable that our HDAC1,2 inhibitors, typified by 1a, having almost the same size of capping and linker domains as those of 12, showed the preferential inhibition for HDAC1 over HDAC2.
The OGD experiments indicated that the synthesized HDAC1, 2 inhibitors and K-560 (1a) exerted neuronal cell protection except for K-563 (4), K-854 (7), OP-857 (9), OP-858 (10) and OP-859 (11) (Figs 3A,B,C and 4C). It was also revealed that 8 and 1a had the promising neuroprotective activity, but the effect of 8 was higher than or at least equal to 1a. This finding was supported from the comparative OGD experiment with various concentrations of 8 and 1a, i.e., it showed that 8 was still effective at the lowest concentration (0.1 μM) where 1a was not effective (Fig. 4A). Another comparative experiment using the clinically used medicines FK228 (HDAC1, 2 inhibitor) and MS-275 (HDAC Class I inhibitor) suggested that both 8 and 1a have a higher neuroprotective activity than FK228, whereas they possess nearly the same activity as MS-275 (Fig. 4B). Since, however, MS-275 showed much higher percent (40.6%) of distribution of Sub-G 0 /G 1 phase cell (corresponding to apoptotic cell) than those SCientiFiC RepoRts | (2018) 8:1400 | DOI:10.1038/s41598-018-19664-9 (10.0-18.1%) for other compounds in the flow cytometric analyses of neuronal SH-SY5Y cells (Fig. 5), compound 8 is expected to be less toxic than MS-275 to neuronal cells.
Because the cyclo-L-prolylglycinylmethyl group reportedly exhibited both nootropic and anxiolytic activities 29 , it was unexpected that compound 7, which possesses this group, was not effective. We speculate that a bicyclic diketopiperazine ring such as that in 7 may be less suitable than a monocyclic diketopiperazine ring for interactions with biological molecules due to its bulkiness. Furthermore, since the HDAC1 and 2 inhibition of 1a and 3-11 was comparable (Table 1), it would be inappropriate that only their enzyme inhibition activities have responsibility for the neuroprotective effects. Instead, it is reasonable to assume that the diketopiperazine moieties (capping group) in these compounds are at least partially involved in their neuronal protection, in view of the reports that diketopiperazines themselves had neuroprotective activities [26][27][28] . This assumption was supported by   the finding that neither the monoketopiperazine-form 9 and 10 nor the piperazine-form 11 showed the neuroprotective effect as much as the diketopiperazine-form 8 and 1a (Fig. 4C).
The 2-nitro derivatives 1b and 2b as well as 2a, having the (4-methylpiperazine-1-carboxamido) methyl moiety, deteriorated cell viability from that with the control (Fig. 3D). The 2-amino group in the 2-amino-5-aryl-benzamide type HDAC inhibitor is known to play a role in chelation with zinc ion at the active site of the HDAC enzyme together with the 2-aminobenzamide-carbonyl, leading to its inhibition. Taken together, the presence of a monocyclic 2,3-or 2,5-diketopiperazine group as well as the HDAC1, 2 inhibition by the 2-aminobenzamide moiety appears to be crucial for the cell survival activity of K-560-related compounds against ischemic damage.
HDAC1 is a key molecular player between neuronal survival and death 44,45 . HDAC2 abolishes neurodegeneration-associated memory impairments via epigenetic blockade 46,47 , and mitigates remote fear memories 48 . In ischemic stroke, the inhibition of both or either HDAC1 and HDAC2 reduce ischemic damage. In neurological disorders, it is currently unclear which isozyme of HDAC enzymes needs to be inhibited because the type of HDAC being activated varies depending on the distinct pathophysiology, associated cell types, or the degree and severity of tissue damage. Further studies are underway to assess the validity of K-560-related compounds for CNS therapeutics against neurodegenerative diseases and stroke.

Materials and Methods
The experimental protocol was approved by the Committee of Osaka University Graduate School of Medicine, Kansai University, and Osaka University of Pharmaceutical Sciences. Cell culturing. Human neuroblastoma cell line SH-SY5Y (ATCC CRL-2266) was cultured in Dulbecco's Modified Eagle's Medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (Gibco) at 37 °C in a 95% air and 5% CO 2 humidified incubator. Cells were routinely subcultured when confluent.

Synthesis.
Animals. Wistar rats (Charles River) were used in this study. The experimental protocol was approved by the institutional animal care and use committee of Osaka University Graduate School of Medicine. Animals were kept four per cage under a 12 h light/dark cycle and standard housing conditions with ad libitum access to food and water before and after all procedures. Animal care was provided according to the Osaka University Medical School Guideline for the Care and Use of Laboratory Animals. Animal surgeries and experimental procedure were approved by the Osaka University Medical School Animal Care and Use Committee. All experiments were conducted according to the National Research Council's guidelines.
Primary cortical cultures. Primary cultures of rat cortical neurons were obtained as described previously 49,50 . Briefly, neuronal cultures were prepared from the cortex of embryonic day 16 (E16) rat embryos. Cells were dissociated with papain (papain dissociation system; Worthington) and plated onto 12-well plates, 4-well plates, and 60-mm dishes (Falcon, Becton Dickinson), or 4-chamber glass slides (Falcon) coated with polyethylenimine. Cells at a final concentration of 7.0 × 10 5 cells/mL were cultured in high-glucose DMEM (Sigma) containing 10% fetal calf serum (FCS; Invitrogen), 100 IU/mL penicillin, and 100 μg/mL streptomycin sulfate. Twenty-four hours after seeding, the medium was changed to Neurobasal medium (Invitrogen) supplemented with B-27 (Invitrogen). Cells were cultured at 37 °C in a humidified atmosphere of 95% air and 5% CO 2 and used after 10-11 days in vitro when most cells showed a neuronal phenotype.
Oxygen-glucose deprivation (OGD). OGD was performed by placing cultures in a 37 °C incubator housed in an anaerobic chamber as previously described 49,50 . Cultures were washed with phosphate-buffered saline and incubated with glucose-free Eagle's balanced salt solution (Biological Industries). Cultures were subjected to an anaerobic environment of 95% and N 2 /5% CO 2 , producing an O 2 partial pressure of 10-15 Torr, as measured with an oxygen microelectrode at 3 h. OGD was terminated by bringing the cultures back to the original medium and placing them in a normoxic chamber.
Analysis of DNA histogram by flow cytometry. SH-SY5Y cells were plated onto 60-mm diameter dishes (1.0 × 10 6 /dish). After incubation for 24 h, the cells were washed twice with the serum-free medium (1 mL) and suspended with the serum-free medium (5 mL) for 24 h. After removal of the medium, the cells were washed twice with the serum-free medium (1 mL) and incubated in the serum-free medium (5 mL) with a test compound (10 μM) for another 48 h. The adherent cells were treated with 0.25% trypsin (Invitrogen) and combined with the floating cells. All the cells were treated with a Cycle Test Plus DNA reagent Kit (Catalog No. 340242, Becton Dickinson). DNA content was measured with a FACSTMCant II. (Becton Dickinson).
Statistical analysis. Data are expressed as the mean ± standard error mean. Statistical analyses were performed using a one-way analysis of variance (SPSS). A p-value of less than 0.05 denotes a significant difference.