A patent review of histone deacetylase 6 inhibitors in neurodegenerative diseases (2014–2019)

Abstract

Introduction

Histone deacetylase 6 (HDAC6) is unique in comparison with other zinc-dependent HDAC family members. An increasing amount of evidence from clinical and preclinical research demonstrates the potential of HDAC6 inhibition as an effective therapeutic approach for the treatment of cancer, autoimmune diseases, as well as neurological disorders. The recently disclosed crystal structures of HDAC6-ligand complexes offer further means for achieving pharmacophore refinement, thus further accelerating the pace of HDAC6 inhibitor discovery in the last few years.

Area covered

This review summarizes the latest clinical status of HDAC6 inhibitors, discusses pharmacological applications of selective HDAC6 inhibitors in neurodegenerative diseases, and describes the patent applications dealing with HDAC6 inhibitors from 2014–2019 that have not been reported in research articles.

Expert opinion

Phenylhydroxamate has proven a very useful scaffold in the discovery of potent and selective HDAC6 inhibitors. However, weaknesses of the hydroxamate function such as metabolic instability and mutagenic potential limit its application in the neurological field, where long-term administration is required. The recent invention of oxadiazole-based ligands by pharmaceutical companies may provide a new opportunity to optimize the druglike properties of HDAC6 inhibitors for the treatment of neurodegenerative diseases.

1. Introduction

Histone deacetylase (HDAC) are intriguing enzymes that regulate the reversible acetylation of histones and play a pivotal role in transcriptional regulation and histone metabolism [1–4]. FDA approval of several broad-spectrum HDAC inhibitors (Vorinostat, Panobinostat, Belinostat, and Romidepsin) for the treatment of several types of cancer has inspired the discovery of drug candidates targeting these proteins that regulate epigenetics for other disorders, especially neurodegenerative diseases [5–7]. However, concerns are rising that their broad HDAC inhibition might result in unwanted off-target effects and related toxicities that could eventually limit their clinical profile for indications beyond cancer [8]. Therefore, the search for subtype-selective (isoform-selective) HDAC inhibitors (HDACis) has become an important endeavor in developing new HDACis with improved biochemical and ADMET properties [9, 10]. Notably, according to a patent survey based on the recent progress of novel HDACi discovery from 2013 to 2017, most patent applications have focused on the discovery of new HDAC6 inhibitors (HDAC6is) [11].

Among the eleven zinc-dependent HDACs (comprising Classes I, IIa, IIb, and IV), HDAC6 is a member of Class IIb that is localized largely to the cytoplasm. It exhibits some unique characteristics, for it is the only isoform that contains tandem catalytic domains, namely CD1 and CD2, with the capacity to specifically deacetylate a variety of non-histone proteins, including α-tubulin, HSP90, peroxiredoxin, cortactin, survivin, and β-catenin (Figure 1a–b) [7]. α-Tubulin was the first HDAC6 substrate identified. Higher level of acetylated α-tubulin (Ac-α-tubulin) at lysine 40 enhances tubulin binding to the motor protein kinesin 1 and facilitates the transport of cargo proteins along with microtubule [12]. Moreover, Ac-α-tubulin also improves the ability of damaged organelles or misfolded proteins to leave synaptic regions [13]. Heat shock protein 90 (HSP90) was identified as the second substrate for HDAC6. It has been reported that HDAC6 inhibition or knockdown elevates the level of acetylated HSP90, inhibits HSP90-HDAC6 binding, and prompts cellular chaperone expression (e.g., HSP70), thereby triggering tau degradation [14, 15]. Peroxiredoxin 1 (Prx1) and 2 (Prx2), other two HDAC6 substrates, are proteins involved in the redox regulation of cell survival, which are increased in many neurodegenerative diseases [16, 17]. Thus, the accumulation of acetylated Prx1 and Prx2 upon HDAC6 inhibition modulates intracellular redox status and further reduces reactive oxygen species (ROS) production [18]. HDAC6 also has a hydrolase-like zinc finger domain, named ZnF-UBP, which is involved in the binding and transport of polyubiquitinated protein aggregates [19]. Additionally, HDAC6 regulates several proteins, such as tau, IIp45, and EGFR, through protein-protein interaction [20–22]. Growing evidence shows that pharmacological inhibition of HDAC6 effectively restores impaired α-tubulin acetylation, rescues mitochondrial transport deficits, degrades protein aggregates, or prevents neuronal oxidative stress to ameliorate the phenotypes in various models of neurodegenerative diseases, such as Alzheimer’s Disease (AD), Charcot−Marie−Tooth disease (CMT), and Amyotrophic Lateral Sclerosis (ALS), thus suggesting selective HDAC6 inhibition may become a therapeutic intervention in neurodegenerative diseases for which there is at present no effective treatments [5–7, 23].

Figure 1.

drHDAC6 protein structure and general pharmacophore of HDAC6 inhibitors. (a) Domain organization in drHDAC6: CD1 = catalytic domain 1, CD2 = catalytic domain 2, ZnF-UBP = zinc finger ubiquitin binding domain; (b) Crystal structure of drHDAC6 CD1 (left) and CD2 (right) (PDB code: 5G0I); (c) Typical pharmacophore of HDAC6is, illustrated for the case of Nexturastat A (1); (d) The active site of drHDAC6-CD2 as seen in the crystal structure of the enzyme complexed with Nexturastat A (PDB code: 5G0I).

2. General pharmacophore of HDAC6 inhibitors

Most HDAC inhibitors (HDACis) share a common pharmacophore that consists of a zinc-binding group (ZBG) that coordinates with the zinc ion (Zn²⁺) located in the bottom of the catalytic cavity, a linker that occupies the hydrophobic tunnel, and a capping group (cap) that interacts with the rim of the pocket (Figure 1c, illustrated for the HDAC6i of Nexturastat A (1) [24]). The phenylhydroxamate structural feature has been demonstrated to act as a useful ZBG as well as a hydrophobic linker in the discovery of selective and potent HDAC6is [9, 25, 26]. The recently published structures of the co-crystals of Danio rerio HDAC6 (drHDAC6) complexed with several ligands reveal that the majority of phenylhydroxamate-based selective HDAC6is engage with Zn²⁺ through an unusual monodentate coordination geometry [27, 28], while a canonical bidentate Zn²⁺ coordination geometry has been observed for HDACis possessing either flexible alkylhydroxamate ZBG-linker units or aromatic linkers lacking a bulky cap [29–31]. Moreover, the phenylhydroxamate moiety nestles in the hydrophobic channel formed by residues F583 and F643 (Figure 1d), engaging in a double π-stacking interaction. Most selective HDAC6is consist of a large and rigid cap that occupies the wide surface area of HDAC6 [27]. Monocyclic to tricyclic aromatic rings are often considered as effective capping groups to establish strong hydrophobic interactions with the selectivity-determining area, “L1 loop pocket”, defined by key residues H463, P464, F583, and L712 (Figure 1d) [30–32]. Moreover, the crystal structures also reveal that certain capping or linker groups are able to engage in additional hydrogen bonds with key residues, such as S531, that play important roles in HDAC6 substrate recognition [27, 29].

3. HDAC6 inhibitors in clinical trials

At present, five HDAC6is are being investigated in clinical trials for different types of cancer, autoimmune diseases, and peripheral pain, but none of them has been advanced into the clinic for the treatment of a neurodegenerative disease. Two partially selective HDAC6is, ACY-1215 (2, Ricolinostat) and ACY-241 (3, Citarinostat) developed by Acetylon Pharmaceuticals (Acetylon), are in extensive clinical trials for different types of cancer, such as multiple myeloma, lymphoma, breast cancer, and melanoma, through monotherapy or a combination approach (Table 1). The structures shown in Figure 2 indicate that both ACY-1215 and ACY-241 share a Vorinostat-like aliphatic long-chain hydroxamate scaffold bearing a large and rigid N,N-diphenyl 2-aminopyrimidine moiety as cap. These compounds exhibit nanomolar potency against HDAC6 (ACY-1215: IC₅₀ = 4.7 nM; ACY-241: IC₅₀ = 2.6 nM) and 12- to 13-fold selectivity over HDAC1 [33, 34]. Notably, it has been reported that ACY-1215 reverses nerve damage and attenuates pain, numbness, and muscle weakness resulting from chemotherapy and CMT [35–37], and a new Phase II trial of ACY-1215 for the treatment of diabetic neuropathic pain was initiated on October 31, 2019 (). Another HDAC6i, KA2507 (4, structure not disclosed), developed by Karus Therapeutics (Karus) is under investigation in a Phase I trial in patients with solid tumors. Chong Kun Dang Pharmaceutical Corp. (CKD) in Korea has advanced an HDAC6i, named CKD-504 (5, structure not disclosed), into a Phase I trial () to evaluate pharmacokinetics (PK)/ pharmacodynamics (PD) as well as toxicity profiles in healthy Korean and Caucasian adults, with the aim of developing potential therapeutics for Huntington disease. Moreover, a highly selective HDAC6i, CKD-506 (6, structure not disclosed), exhibits excellent HDAC6 potency (IC₅₀ = ~5 nM) and at least 400-fold selectivity over other HDAC isoforms [38]. In a mouse model of systemic lupus erythematosus (SLE), a chronic multisystemic autoimmune disease, CKD-506 suppressed lupus nephritis without significant adverse effects by repressing the expression of lupus disease-specific cytokines and chemokines in serum and kidneys [38]. Moreover, its safety, tolerability, and PK profile have been characterized in a Phase I study in Europe. In the meanwhile, a Phase 2a clinical trial is ongoing to evaluate its effects in arthritis patients [39].

Table 1.

Summary of HDAC6 inhibitors in clinical trials

Compound ID	Condition or disease	Phase	Clinical trial identifier
ACY-1215 (2)	Multiple myelomas	I/II (completed and active)	,
	Relapsed/refractory multiple myeloma	I/II (active)	,
	Relapsed/refractory lymphoid malignancies	I/II (recruiting)
	Diabetic neuropathic pain	II (not yet recruiting)
	Gynecological cancer	I (active)
	Metastatic breast cancer	I (active)
	Recurrent chronic lymphoid leukemia	I (active)
	Cholangiocarcinoma	I (withdrawn)
ACY-241 (3)	Malignant melanoma	I (completed)
	Multiple myelomas	I (active)
	Advanced solid tumors	I (active)
	Unresectable non-small cell lung cancer	I (recruiting)
KA2507 (4)	Solid tumor	I (recruiting)
CKD-504 (5)	Huntington disease	I (recruiting)
CKD-506 (6)	Autoimmune diseases	I (completed)	EudraCT201600281642

Figure 2.

Selective HDAC6 inhibitors in clinical and preclinical stages of development

4. HDAC6is and neurodegenerative diseases

Although the discovery of novel HDAC6is has become an attractive area of research in the last few years leading to the generation of diverse HDAC6is bearing novel caps, linkers, as well as ZBGs, only a limited number of candidates have been investigated that show effective results in models of neurodegenerative diseases. These efforts have originated predominately from one biotech company and two academic research groups (Table 2). First, besides the clinical-stage candidates ACY-1215 and ACY-241, Acetylon also invented two more brain-penetrable HDAC6is, ACY-738 (7) and ACY-775 (8) (Figure 2), with improved isoform selectivity [40]. Secondly, the Kozikowski group at the University of Illinois at Chicago (United States) discovered a phenylhydroxamate-based ligand containing a tetrahydro-γ-carboline ring system as cap, named Tubastatin A (TubA, 9). This was the first highly selective HDAC6i that was easy to synthesize, and that was shown to have some positive effects in animal models of neurodegenerative diseases [25, 41]. Inspired by the invention of TubA, a benzimidazole-capped selective Class II HDACi, 23d (10), containing a 3-fluorophenylhydroxamate moiety, was developed by the Kozikowski group that was found to exhibit improved biochemical properties and bioavailability [41]. The recently published tetrahydroquinoline-capped HDAC6i, SW-100 (11), exhibits a thousand-fold selectivity over other HDAC isoforms and shows enhanced brain penetration relative to TubA [42]. Moreover, the 8-aminoquinoline-capped compound W-2 (11) bearing a mercaptoacetamide moiety as an alternative ZBG is the only non-hydroxamate analog that exhibited efficacy in an animal model of Alzheimer’s Disease [43–46]. Thirdly, the Liou group at the Taipei Medical University (Taiwan) discovered two brain penetrable phenylhydroxamate-based compounds, the 5-Aroylindole 6 (12) and MPT0G211 (13) (Figure 2), which exhibit excellent HDAC6 potency and isoform selectivity, and which are able to rescue Alzheimer’s disease phenotypes in murine models [47–49]. The area of therapeutic applications of these HDAC6is are summarized in Table 2.

Table 2.

Selective HDAC6 inhibitors with efficacy in models of neurodegenerative disorders

Compound ID	Condition or disease	Reference
ACY-1215 (1)	Charcot-Marie-Tooth disease 2F	[36]
	Alzheimer’s Disease	[57]
	Rett Syndrome	[83]
ACY-738 (6)	Alzheimer’s Disease	[56, 58]
	Charcot-Marie-Tooth disease 2F	[36]
	Amyotrophic lateral sclerosis	[71, 72]
ACY-775 (7)	Charcot-Marie-Tooth disease 2F	[36]
Tubastatin A (8)	Alzheimer’s Disease	[52, 53, 55, 57]
	Amyotrophic lateral sclerosis	[71]
	Charcot-Marie-Tooth disease 2F	[62]
	Charcot-Marie-Tooth disease 2D	[63, 66]
	Parkinson’s disease	[18, 74, 75]
	Rett Syndrome	[81, 82]
	Huntington’s disease	[88]
Benzimidazole 23d (9)	Charcot-Marie-Tooth disease 2F	[41]
SW-100 (10)	Fragile X Syndrome	[42]
W-2 (11)	Alzheimer’s Disease	[45, 46]
5-Aroylindole 6 (12)	Alzheimer’s Disease	[47]
MPT0G211 (13)	Alzheimer’s Disease	[49]

4.1. HDAC6 inhibitors and Alzheimer’s Disease

Alzheimer’s disease (AD) is one of the most common neurodegenerative diseases associated with progressive cognitive decline, which is characterized by two key pathological features that include the extracellular deposition of the amyloid-β (Aβ) peptide resulting in senile plaques and the intracellular accumulation of hyperphosphorylated tau (p-tau) protein forming neurofibrillary tangles (NFTs). An increasing number of studies demonstrate that HDAC6 regulates acetylated α-tubulin/HSP90 levels and forms complexes with HSP90, ubiquitin, and tau, thereby playing a crucial regulatory role in mitochondrial axonal transport, protein aggregation/degradation, as well as tau phosphorylation. Based on these findings, HDAC6 has emerged as an attractive therapeutic target for AD [7, 50].

Impaired axonal transport and mitochondrial dysfunction are considered early pathological features of AD [51]. In 2012, Mook-Jung and co-workers found that a decreased level of Ac-α-tubulin could be detected in the frontal cortex isolated from 5XFAD mice expressing five AD-linked mutations driving Aβ overproduction [52]. Treatment with TubA (5 μM) rescued disrupted mitochondrial transport and restored impaired Ac-α-tubulin in rat hippocampal neurons in the presence of Aβ (2 μM). In 2014, Richter-Landsberg and co-workers detected HDAC6 in rat brain oligodendrocytes and found that TubA (2 μM) treatment led to morphological alterations, microtubule bundling, increases in the level of Ac-α-tubulin, and changes in tau-isoform expression and phosphorylation [53].

Following the encouraging results with TubA in neurons, several animal studies using TubA, ACY-1215, as well as ACY-738 were reported in 2014–2015. rTg4510 mice is a transgenic mouse model of human tau pathology, expressing the mutant htau^P301L, which phenocopies age-dependent neurofibrillary tangle formation, a prominent loss of neurons, and memory deficits [54]. Morgan and co-workers treated rTg4510 mice with TubA (25 mg/kg, IP) for two months and found amelioration of memory impairment and reversal of a hyperactivity phenotype [55]. They also found that TubA treatment promoted the decrease of total tau level which is positively correlated with memory improvement in these transgenic mice, while there was no impact on the phosphorylation status of tau, such as S202/T205, S396, and S199/202. Petrucelli and co-workers reported that HDAC6 inhibition regulated the tau acetylation on KXGS motifs (consisting of KIGS and KCGS motifs), which are hypoacetylated and hyperphosphorylated in frontal cortex from patients with AD and in the rTg4510 mouse model [56]. A two-day acute animal study of ACY-738 (0.5 mg/kg, SC) in FVB non-transgenic mice showed that selective HDAC6 inhibition simultaneously decreased the level of p-tau (12E8) and increased the level of acetylated tau (Ac-KIGS) at these crucial KXGS motifs in mice. Amyloid precursor protein/presenilin 1 (APP/PS1) double-transgenic mice, expressing AβPP^swe and PS1^L166P mutations, exhibit Aβ deposition in the cortex beginning at six weeks of age and cognitive deficits in learning and memory at seven months of age. Qin and co-workers conducted a 20-day study of the treatment of APP/PS1 mice with TubA (25 mg/kg, IP) or ACY-1215 (25 mg/kg, IP) and demonstrated that both compounds ameliorated memory function deficits without significant effects on anxiety levels, and increased the levels of Ac-α-tubulin in the hippocampus and cortex [57]. Moreover, both pharmacological inhibitors reduced Aβ deposition in the hippocampus and cortex, which was achieved by inhibiting β/γ-secretase cleavage of AβPP and promoting autophagy. Notably, HDAC6i treatment also led to decreased levels of various types of p-tau (S202/T205, T231, T181, and S396/404) via inhibiting the Akt/GSK3β signaling pathway, thus restoring microtubule stability, which was not observed in the rTg4510 mouse study described previously [55]. Pautler and co-workers studied the treatment of APP/PS1 mice (6–7-month old) with ACY-738 (100 mg/kg, chow-based formulation) for 21 days or 90 days, respectively [58]. They found that treatment with ACY-738 rescued axonal transport deficits, elevated Ac-α-tubulin levels, and reduced the p-tau level (S262) in both short and long treatment groups. Furthermore, ACY-738 also improved hyperactivity and fear-associated contextual learning and memory, while significantly decreasing the levels of insoluble Aβ_1–42 and the Aβ₄₂/Aβ₄₀ ratio.

In 2018, Liou and co-workers more extensively evaluated the effects of two selective HDAC6is, 5-Aroylindole 6 (13) and MPT0G211 (14), along with ACY-1215 in both in vitro and in vivo models of AD [47, 49]. They measured tau phosphorylation in SH-SY5Y and N2a cells transfected with plasmids encoding hAPP695 and hTau^P301L exhibiting hyperphosphorylated tau (S396). Treatment with 5-Aroylindole 6, MPT0G211, or ACY-1215 (1 μM) significantly reduced the level of p-tau (S396). Moreover, further evaluation on 5-Aroylindole 6 (1 μM) found that it remarkably attenuated tau phosphorylation at S396 and S404, slightly decreased APP levels in transfected cells, and diminished Aβ_1–40-induced tau phosphorylation. Inspired by these inhibitory effects on tau hyperphosphorylation, they further investigated if HDAC6is influence the polymerization of p-tau. The results in SH-SY5Y cells transfected with hAPP695 and htau^P301L revealed that treatment with 5-Aroylindole 6 (1 μM), MPT0G211(0.1 μM), or ACY-1215 (0.1 μM) significantly reduced levels of aggregated tau. On the other hand, clearance of tau aggregates may serve as a relevant therapeutic strategy, in which molecular chaperones, such as heat shock proteins (HSPs) and the ubiquitin-proteasome system (UPS) are highly involved in protein regulation and degradation [59]. The results in SH-SY5Y and N2a cells transfected with hAPP695 and htau^P301L further demonstrated that this HDAC6i treatment increased the levels of Ac-α-tubulin and acetylated HSP90 (K292/284), attenuated HDAC6-HSP90 binding, and promoted HSP70 expression. Notably, selective HDAC6 inhibition also produced an increase in the ubiquitin-p-tau (S396) complex leading to subsequent proteasomal degradation of polyubiquitinated aggregates. In the end, 5-Aroylindole 6 (50 and 100 mg/kg, IP) and MPT0G211 (50 mg/kg, IP) were investigated in a triple transgenic (3xTg-AD) mouse model expressing APP^swe and tau^P301L mutants that produced AD-like neuropathology (e.g., plaque and tangles). The animal behavior tests and immunohistochemical analysis showed that pharmacological HDAC6 inhibition ameliorated spatial memory impairment while enhancing Ac-α-tubulin, decreased the levels of Aβ, and down-regulated the levels of p-tau (S396 and S404) in the hippocampal CA1 region of the mouse brain without affecting the level of total tau.

4.2. HDAC6 inhibitors and Charcot-Marie-Tooth Disease

Charcot-Marie-Tooth (CMT) disease is the most common inherited disorder of the peripheral nervous system (PNS), caused by more than 80 different types of mutation, in which motor and sensory peripheral nerves are degenerated resulting in muscle weakness, motor problems, and sensory loss [60, 61]. In 2011, the Van Den Bosch group characterized a type of transgenic mice expressing two different mutations (S135F or P182L) in the heat-shock protein gene B1 (HSPB1/HSP27) that replicate the motor and sensory deficits seen in patients with CMT type 2F (CMT2F) or distal hereditary motor neuronopathies (HMN) type 2B, demonstrated by reduced compound muscle action potentials (CMAPs) and sensory nerve action potentials (SNAPs) [62]. Secondly, dorsal root ganglion (DRG) neurons isolated from HSPB1^S135F-expressing transgenic mice displayed a decrease in the number of total mitochondria and a defect in axonal transport of mitochondria compared to their wild-type littermates. Moreover, both HSPB1^S135F and HSPB1^P182L mice exhibit impaired acetylation of α-tubulin in the sciatic nerve. As HDAC6 recognizes α-tubulin as its preferred substrate and plays a critical role in controlling the axonal transport of mitochondria in cultured hippocampal neurons [12], a 3-week course of treatment with TubA (25 mg/kg, IP) was found to effectively restore impaired Ac-α-tubulin levels in the sciatic nerve and to rescue the CMT phenotype. Moreover, after treating DRG neurons from HSPB1^S135F mice with TubA (1 μM) for 12 h, the total number of mitochondria and the number of moving mitochondria were elevated, thus rescuing axonal transport defects [62]. In 2016, the Kozikowski and Van Den Bosch groups disclosed a second generation HDAC6 inhibitor, Benzimidazole 23d (10), which showed an improved ability to acetylate α-tubulin in N2a cells as well improved mitochondrial axonal transport in DRG neurons isolated from HSPB1^S135F mice compared to TubA [56]. In 2017, the Van Den Bosch group further conducted an evaluation of ACY-1215 (30 mg/kg, IP), ACY-738 (3 mg/kg, IP), and ACY-775 (3 mg/kg, IP) for three weeks in CMT2F mice expressing HSPB1^S135F. The compounds reversed the axonal deficits in motor and sensory nerves and induced reinnervation of the neuromuscular junction.

Aminoacyl transfer RNA (tRNA) synthetases are the largest gene/protein family implicated in CMT, in which glycyl-tRNA synthetase (GlyRS or GARS) was the first member identified among the dominant mutations associated with CMT type 2D (CMT2D). In 2018, the Van Den Bosch group investigated if HDAC6 could serve as a therapeutic target focusing on the mutant GlyRS-induced CMT2D.[63] Motor and sensory deficits reminiscent of CMT2D were observed in 1-year old transgenic mice expressing Gars^C201R/+ [64]. Secondly, the levels of Ac-α-tubulin in peripheral nerves and DRG neurons isolated from the Gars^C201R/+ mice are significantly impaired relative to samples from wild-type littermates, while normal mitochondrial trafficking is disrupted in the DRG from Gars^C201R/+ mice. A coimmunoprecipitation assay in N2a cells demonstrated that the interaction between GlyRS and HDAC6 was blocked after administration of TubA (1 μM), while Ac-α-tubulin was increased. A 40-day treatment with TubA (50 mg/kg, IP) ameliorated motor and sensory deficits and stimulated regeneration of motor and sensory nerves in 4-month-old Gars^C201R/+ mice. Moreover, this selective HDAC6 inhibition led to enhanced levels of Ac-α-tubulin in sciatic nerves and DRG homogenates from Gars^C201R/+ mice compared to the vehicle group. In the same year, the Yang group transfected murine motor neurons NSC-34 with V5-tagged GlyRS^CMT2D constructed of 11 different human mutations including C201R and P234KY, two mutations that have been identified in CMT2D mouse models [64, 65]. Significantly, all mutants demonstrated aberrant interaction with HDAC6 [66]. Moreover, GlyRS-HDAC6 interaction is observed in CMT2D mice expressing Gars^P234KY/+, but not in the wild-type littermates. In the sciatic nerves isolated from the Gars^P234KY/+ mice, the level of HDAC6 is significantly higher than that in the spinal cord or brain sample resulting in a decrease in the level of Ac-α-tubulin in the sciatic nerves. Moreover, axonal transport is significantly impaired in peripheral neurons prior to expressing CMT phenotypes in the Gars^P234KY/+ mice. It was found that DRG neurons from Gars^P234KY/+ mice, treated with TubA (2 μM), exhibited almost fully restored retrograde axonal transport. Subsequent treatment of the Gars^P234KY/+ mice for 14 days with TubA (50 mg/kg, IP) improved muscle strength and motor performance as the result of a significantly elevated level of Ac-α-tubulin in their sciatic nerves.

All these studies indicate that different genetic forms of CMT may originate through a similar pathogenic mechanism consisting of enhanced α-tubulin deacetylation and disrupted axonal transport. This insight broadens the therapeutic potential of selective HDAC6 inhibition to other types of CMT.

4.3. HDAC6 inhibitors and amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis (ALS) is caused by selective loss of motor neurons in the motor cortex, brainstem, and spinal cord leading to progressive muscle weakness, paralysis, and ultimately the death of the patient on an average of 2–5 years after the diagnosis. Mutations in the genes encoding superoxide dismutase 1 (SOD1), TAR DNA-binding protein 43 (TDP-43) and fused in sarcoma (FUS) are the most prevalent genetic causes of ALS. As axonal transport defects were observed in transgenic mice containing the G93A mutation in superoxide dismutase 1 (SOD1) [67], genetic deletion of HDAC6 significantly slowed down disease progression and prolonged survival of the mutant SOD1^G93A mouse model [68]. Moreover, it was reported that RNAi silencing of either TDP-43 or FUS reduced the mRNA expression of HDAC6 [69]. As most therapeutic strategies that have been found to be effective in the commonly employed mutant SOD1-ALS model, have failed to demonstrate clinical efficacy over the last two decades [70], better in vitro and in vivo models are required to assess the therapeutic effects of drugs including selective HDAC6is on ALS.

In 2017, the Van Den Bosch group characterized induced pluripotent stem cells (iPSCs) from fibroblasts of ALS patients with different FUS mutations (R531H and P525L) [71]. Motor neurons cultured from these iPSCs exhibited typical cytoplasmic FUS pathology, hypoexcitability, and progressive axonal transport defects. Treatment with the HDAC6is TubA (1 μM) or ACY-738 (1 μM) restored impaired axonal transport and increased endoplasmic reticulum (ER)-mitochondrial overlay in these patient-derived motor neurons. In 2019, Van Den Bosch and co-workers further revealed their discovery of a transgenic mouse model of ALS overexpressing wild-type FUS (Tg FUS +/+ mouse) [72], which showed histone hypoacetylation in spinal cord and cortical tissue associated with its progressive neurodegeneration. ACY-738 (100 mg/kg, chow-based formulation) treatment significantly extended the survival rate of the Tg FUS +/+ mice compared to the vehicle group, and attenuated neuromuscular denervation and muscle atrophy, thus ameliorating ALS disease phenotypes. At the molecular level, ACY-738 restored the level of acetylated histone and decreased the cytoplasmic FUS level in the spinal cord of Tg FUS +/+ mice.

4.6. HDAC6 inhibitors and Parkinson’s disease

Parkinson’s disease (PD) is a common neurodegenerative movement disorder characterized by tremor, hypokinesia, rigid muscles, and impaired balance. The most prominent pathological features are progressive loss of dopaminergic neurons within the substantia nigra and loss of dopamine terminals in the striatum. Moreover, it has been reported that PD is associated with mitochondrial dysfunction and oxidative stress [73]. It is evident that HDAC6 inhibition upregulates mitochondrial axonal transport and prompts protein aggregate degradation in the models of AD, thus suggesting HDAC6 inhibition may also provide a therapeutic approach for PD. In 2014, Godena and co-workers found that leucine-rich repeat kinase 2 (LRRK2) mutations, the most common genetic cause of PD, lead to deacetylated microtubules and defective mitochondrial movement in the cortical neurons of rats expressing LRRK2 Roc-COR variants (R1441C or G2019S), associated with locomotor deficits in the corresponding transgenic Drosophila model [74]. Either treatment with the HDAC6i, TubA (2 μM), or overexpression of the α-tubulin acetyltransferase α-TAT1 specifically increased microtubule acetylation, thus significantly leading to the reversal of filamentous LRRK2 formation. Moreover, in a HDAC6 knockdown Drosophila model, the loss of HDAC6 effectively restored axonal transport deficits in LRRK2 Roc-COR mutant transgenic flies while rescuing climbing and flight deficits caused by mutant LRRK2. In 2016, Oliveira and co-workers investigated the effects of HDAC6 inhibition in a wild-type zebrafish (D. rerio) model using the dopaminergic toxin MPP+ to generate the associated behavioral and metabolic phenotypes [75]. Treatment of zebrafish with TubA (1 μM) increased the level of Ac-α-tubulin primarily distributed in the nervous system, and rescued decreases in diencephalic tyrosine hydroxylase and metabolic impairment induced by MPP⁺ without influencing the mitochondrial complex. Furthermore, TubA treatment modulated spontaneous zebrafish locomotion and attenuated the MPP⁺-induced head reflex impairment. In 2017, Zhang and co-workers treated 6-OHDA-induced PD mice with TubA (25 mg/kg, IP) for seven days to assess its effects on the nigrostriatal dopaminergic system [18]. They observed decreased expression of tyrosine hydroxylase in the substantia nigra and in the striatum together with an enhanced expression of HDAC6 in the nigrostriatal system subjected to the 6-OHDA injury compared to the sham group. In the above 6-OHDA PD mice, the acetylation status of Prx1 and Prx2, two important HDAC6 substrates, was impaired compared to the sham group but was restored after TubA treatment [18]. Moreover, TubA reduced reactive oxygen species (ROS) production and ameliorated dopaminergic neuronal damage in 6-OHDA lesioned mice.

4.4. HDAC6 inhibitors and Rett Syndrome

Rett Syndrome (RTT) is the most common neurodegenerative disorder in females associated with the impairment of intellectual function and motor abilities, which is predominately caused by mutations in the gene that encodes the X-linked methyl-CpG-binding protein 2 (MeCP2), an important reader of DNA methylation [76]. Expression of the gene encoding brain-derived neurotrophic factor (BDNF), a nerve growth factor essential for survival of neurons, was found to be regulated by MeCP2 which is relevant to the pathogenesis of RTT [77]. Reduced mRNA and protein levels, as well as defective trafficking of BDNF, have been detected in various brain regions of MeCP2-deficient mice and RTT individuals [78, 79]. Microtubule acetylation upregulated by HDAC6 inhibition increases both the anterograde and retrograde fast axonal transport of BDNF and mitochondria, thus reversing synaptic dysfunction [12, 80].

In 2014, Pozzo-Miller and co-workers reported that, in hippocampal neurons cultured from MeCP2 knockout mice, TubA (1 μM) treatment ameliorated the impaired BDNF trafficking tagged with yellow fluorescent protein (BDNF-YEP) in both the anterograde and retrograde directions by enhancing the level of Ac-α-tubulin, and also increased the level of activity-dependent release of BDNF back to a level comparable with that present in wild type (WT) neurons [81]. In 2015, Christodoulou and co-workers detected decreased levels of Ac-α-tubulin and increased levels of HDAC6 in the cortex of symptomatic 12-week-old Mecp2^T158A mice and in fibroblasts derived from RTT patients, and observed microtubule defects in patient cells expressing mutations in the MeCP2 gene [82]. In vitro TubA (0.5–20 μM) treatment enhanced the level of Ac-α-tubulin in both mouse and human neurons and also stabilized microtubules against nocodazole-induced depolymerization. In 2018, Campiani and co-workers characterized MeCP2-iPSC neurons with different mutations for recapitulating RTT features, which exhibited a significant decrease in Ac-α-tubulin, a reduced number of synapses, impaired spine density, as well as defective calcium signaling [83]. Furthermore, RNA-seq data analysis revealed an over-expression of the HDAC6 gene, while a reduction in Ac-α-tubulin in the MECP2-mutated neurons was detected compared to controls. The Ac-α-tubulin levels were elevated after the treatment with ACY-1215 (20 or 40 μM).

4.5. HDAC6 inhibitors and Huntington Disease

Huntington’s disease (HD) is a fatal, hereditary neurodegenerative disease caused by polyglutamine (polyQ) expansion mutations in the Huntingtin protein. Motor disorders are accompanied by weight loss, cognitive decline, and psychiatric disturbances. HDAC6 was proposed as a therapeutic target for HD, as it was evident that HDAC6 inhibition compensates intracellular transport deficits in HD models by increasing Ac-α-tubulin levels impaired in the brains of HD patients [84, 85]. In contrast, it was reported that genetic deletion of HDAC6 enhanced Ac-α-tubulin levels but did not improve BDNF transport, markedly reduced in HD, from the cortex to the striatum in the R6/2 transgene mice [86]. In another study, a HDAC6 knockout approach did increase levels of Ac-α-tubulin in both striatum and cortex as well as the BDNF protein level in the striatum of R6/1 mice, whereas it did not ameliorate the weight loss or the deficits in cognitive abilities, and even exacerbated social impairments and hypolocomotion in the R6/1 mouse model [87]. These findings weaken the validity of HDAC6 reduction as a possible therapeutic strategy for HD. Notably, in 2015, Oliveira and co-workers reported in vitro studies using primary cultures of striatal and cortical neurons, used as an experimental model to explore the differential vulnerability in HD neurons, to investigate the effects of pharmacological HDAC6 inhibition [88]. TubA (1 μM) treatment enhanced Ac-α-tubulin level, improved mitochondrial transport, and promoted mitochondrial fusion in striatal neurons. Furthermore, TubA increased autophagic flux in striatal neurons without blocking neuronal autophagosome–lysosome fusion or changing mitochondrial DNA levels, while diminished levels of diffuse mutant huntingtin in striatal neurons in a polyQ-dependent manner. These in vitro studies may suggest the potential of pharmacological HDAC6 inhibition to reduce striatal vulnerability to HD.

4.7. HDAC6 inhibitors and Fragile X Syndrome

Fragile X Syndrome (FXS) is the most common inherited severe cognitive impairment in males, which causes developmental problems, such as learning disability, cognitive impairment, and behavioral characteristics. In most cases, FXS is caused by transcriptional silencing of the fragile X mental retardation 1 (FMR1) gene due to the methylation of a cytosine-guanine-guanine (CGG) trinucleotide repeat expansion (number of repeats in FXS > 200, normal < 50), leading to the loss of fragile X mental retardation protein (FMRP) which influences mRNA trafficking and dendritic translation. Significantly lower levels of mitochondria and defective mitochondrial movement were detected in hippocampal neurites of Fragile-X-associated tremor/ataxia syndrome (FXTAS, 50 < number of CGG repeats < 200) [89, 90]. As HDAC6 inhibition has been shown to improve the axonal transport of mitochondrial as well as vesicular trafficking of brain-derived neurotrophic factor (BDNF) [62, 81, 82], it would appear to offer a possible therapeutic intervention for FXS. In 2018, Kozikowski and co-workers demonstrated that the Ac-α-tubulin levels were 30% lower in the hippocampus of Fmr1^−/− mice, a mouse model of FXS, compared to that of WT mice, and that these could be restored after treatment with a new brain-penetrable HDAC6i, named SW-100 (11) [42]. Moreover, SW-100 rescued memory and learning impairments in Fmr1^−/− mice that model the intellectual deficiencies associated with FXS, thus suggesting an innovative approach for this neurodegenerative disorder.

5. HDAC6 inhibitor patent application summary (2014–2019)

The significant effects of selective HDAC6 inhibition on various models of neurodegenerative diseases are encouraging. While most patent applications typically disclose rather limited biochemical information, it is hypothesized that novel HDAC6i scaffolds, exhibiting more favorable drug-like properties, would lead to further improvements in efficacy and yield a higher potential to identify drug candidates. Herein, we have focused on HDAC6i patent applications disclosed from 2014 to 2019 the details of which have not been published in any research articles.

5.1. Alkylhydroxamate-based HDAC6 inhibitors

In 2014 and 2016, Karus Therapeutics (Karus) developed compounds containing an aliphatic long-chain hydroxamate [91, 92]. Unlike ACY-1215 and ACY-241, these compounds contain two monocyclic or bicyclic heteroaryl rings as caps directly linked to an amino group instead of bearing a pyrimidine core. As an example, compound 15 shown in Figure 3a exhibited an IC₅₀ value of 1.8 nM against HDAC6 and 126-fold selectivity over HDAC1 and is orally bioavailable (F% = 19) in a mouse PK study [91]. The plasma clearance rate was significantly improved in the case of compound 16 (mouse plasma clearance = 48.6 mL/min/kg) in comparison with compound 15 (mouse plasma clearance = 252.8 mL/min/kg) (Figure 3a) [92]. In 2016, CKD disclosed a Vorinostat-like series of compounds bearing substituted diphenylmethyl groups linked to a heterocyclic alkyl core [93]. Among them, compound 17 (Figure 3b) demonstrated low nanomolar potency against HDAC6 and 960-fold selectivity over HDAC1, indicating that its piperazine ring is responsible for its significantly improved HDAC6 selectivity compared to ACY-1215 bearing a pyrimidine core. Notably, treatment with 17 at three doses (1, 10, or 50 mg/kg, PO) alleviated symptoms in an adjuvant-induced arthritis rat model. In 2019, Shuttle Pharmaceuticals used biphenylyloxy groups connected to benzyl groups as capping groups to generate SP-2–225 (18, Figure 3c), which is about 65-fold and 22-fold more selective for HDAC6 compared to HDAC1 and HDAC3, respectively [94].

Figure 3.

Alkylhydroxamate-based HDAC6 inhibitors

5.2. Arylhydroxamate-based HDAC6 inhibitors

In 2015, Acetylon published the discovery of compound 19 that exhibited an IC₅₀ value of 3.7 nM at HDAC6 and hundred-fold selectivity over HDACs 1–3 (Figure 4a) [95]. Its structure is similar to that of ACY-775, while it contains difluoro substitution in the cyclohexane ring. Significantly, 19 showed a lower rat plasma clearance rate (3.7 L/h/kg), higher oral bioavailability (F% = 41.5), and improved mouse brain penetration (B/P ratio = 1.49) relative to ACY-738 and ACY-775 [95]. Compound 19, named ACY-1083 in the literature, at doses of 3 or 10 mg/kg (IP) reversed symptoms of peripheral neuropathy and restored mitochondrial transport in cisplatin-treated mice [35]. Recently, a crystal structure analysis of HDAC6 complexed with ACY-1083 (PDB code: 5WGM) revealed that its NH group attached to the pyrimidinyl group engages in a hydrogen bond directly to the hydroxyl group of S531 on the L2 loop that plays an important role in HDAC6 substrate recognition [28, 29]. In the same year, an Acetylon patent was published claiming analogs bearing substituted piperidine moieties in place of the cyclohexane ring. Compound 20 as a representative displayed IC₅₀ values of 1.9 nM, 38 nM, and 34 nM against HDAC6, HDAC1, and HDAC2, respectively [96]. The migration of neuroblastoma cells of the cell line SK-N-SK was reduced in the presence of compound 20. In 2017, it was reported that a partially selective pyrimidinylhydroxamate-based HDAC6i, compound 21 (30 mg/kg, PO), that does not contain a heterocyclic substituent at the benzylic position, evaluated alongside ACY-738 and ACY-1215, reversed tactile allodynia in STZ rats and restored the pain threshold in these rats to almost normal levels [97]. In 2018, Acetylon reported HDAC6is 22–24 containing phenylhydroxamate or N-hydroxypicolinamide as ZBG that exhibited improved DMPK profiles. Relative to compound 21, the replacement of its pyrimidinyl group with a phenyl group (22) enhanced HDAC6 selectivity and led to a more favorable rat plasma clearance rate (1–5 L/h/kg, 5 mg/kg, IV). N-Hydroxypicolinamide-based analogs 23 and 24 contain a diphenylmethyl group and diphenylyl group linked to a nitrogen atom, respectively. Mouse PK studies (5 mg/kg, IP) showed that brain to plasma ratios were greater than one (AUC_brain = 450–800 h*ng/mL vs AUC_plasma = 150–250 h*ng/mL), thereby indicating excellent brain permeability of 23 and 24 [98].

Figure 4.

Arylhydroxamate-based HDAC6 inhibitors

In continuation of their efforts on alkylhydroxamates 15 and 16, Karus further introduced similar capping groups into the structures of arylhydroxamate-based HDAC6is (Figure 4b). Compounds 25-27 shown as examples contain two heteroaryl rings. These compounds afforded modest activity against HDAC1, while their potency against HDAC6 was better than 500 nM [99–101]. In a combination study with bortezomib, compound 27 (100 or 200 mg/kg, PO) effectively inhibited tumor growth (%T/C value < 40%) in the RPMI8226 xenograft model without significant weight change [101].

Unlike most researchers that focused on phenyl/heteroaryl-based hydroxamates, Forma Therapeutics (Forma) extended their interest into the discovery of spiro-fused tetraline/indane based hydroxamates containing aromatic caps (Figure 4c) (e.g., compounds 28–30) [102–104]. Biochemical assays indicated that IC₅₀ values against HDAC6 were lower than 100 nM, but no further biological information was reported in these patent applications. As 2-aminotetralin has been demonstrated as a useful linker to develop potent and selective HDAC6/8 dual inhibitors [105], compounds 28-30 might be expected to exhibit moderate HDAC8 activity. Subsequently, Forma also developed hydroxamate analogs bearing tetrahydrobenzo-1,4-oxazepine or tetrahydrobenz-1,5-oxazocine groups (e.g., 31–33) as linkers [106–108]. The exploration of the 1,4-oxazepine class was based on an investigation of different substituents on the nitrogen atom followed by another patent focusing on derivatives bearing various amide substituents. Most compounds contain an S-configured stereocenter bearing an aromatic ring at the C3 position and exhibited reasonably good HDAC6 inhibitory activity (IC₅₀ < 100 nM). However, their selectivity over other HDAC isoforms is unknown. The 1,5-oxazocine class was developed in a similar manner, and some of the compounds displayed nanomolar potency against HDAC6 (IC₅₀ < 100 nM).

In 2014–2015, CKD disclosed highly potent and selective HDAC6is bearing phenylhydroxamate moieties in four patents (Figure 4d). Compound 34 contains a urea core linked to a benzene ring and a heterocyclic ring [109], a structural pattern that may have been inspired by Nexturastat A (Figure 1). Compound 34 demonstrated low nanomolar potency against HDAC6 (IC₅₀ = 1 nM) and hundred-fold selectivity over other HDAC isoforms. Moreover, administration of 34 (10 or 30 mg/kg, IP) twice a day effectively ameliorated symptoms in an adjuvant-induced arthritis model and slowed down weight loss in a colitis mouse model, thus confirming the efficacy of this novel urea derivative. Compounds 35 and 36 have a 2-fluoro-2-methylpropyl fragment connected to bicyclic caps through piperidine and piperazine rings, respectively [110, 111]. Both compounds exhibited strong HDAC6 inhibition and weak activity against HDAC1. Moreover, a high concentration of 36 (50 mg/kg, PO) was detected in mouse brain (Conc_brain@1h = 3,000 ng/mL vs Conc_plasma@1h = 3,500 ng/mL), suggesting good metabolic stability and blood-brain-barrier (BBB) permeability [111]. The phenylhydroxamate-based compound 37 contains a unique (2S,6R)-2,6-dimethylpiperazine core linked to a 3-chlorobenzyl group and is believed to be the most selective HDAC6i (5,540-fold over HDAC1) in this class [112].

In 2018, Italfarmaco invented a series of phenylhydroxamate derivatives using triazole, thiadiazole, or oxadiazole as a core building block that was in turn linked to an aromatic ring. Compound 38, for example, displayed low nanomolar activity against HDAC6 (IC₅₀ = 8 nM) and 65-fold selectivity over HDAC3. The compound also demonstrated metabolic stability in rat/human plasma and in the liver S9 fraction. Mouse PK (1–5 mg/kg, IV and PO) and toxicity studies showed that 38 is orally bioavailable (F% = 25.8) and is tolerated at a dose of 50 mg/kg without causing significant weight loss [113]. In the same year, Medshine Discovery disclosed inhibitor 39 containing a pyridine core linked to a 2-phenyltetrahydrofuran moiety [114]. In a dog PK study, 39 displayed a low clearance rate (16 mL/min/kg) leading to a high oral bioavailability (%F = 53). In a multiple myeloma mouse model, the combination of 39 (30 mg/kg) and ixazomib (4 mg/kg) significantly inhibited tumor growth (%T/C value = 8.3; TGI% = 96.88)

5.3. Trifluoromethyloxadiazole derivatives as selective HDAC6is

Based on the successful story of TMP195 (developed by GSK), a selective Class IIa HDACi, the trifluoromethyloxadiazole has been shown to be a useful alternative ZBG to the hydroxamate group [115]. In recent patent applications, oxadiazole-type ZBGs have been widely applied to the discovery of novel selective HDAC6is.

In 2017, Merck disclosed compounds containing trifluoromethyl-1,2,4-oxadiazole as ZBG in combination with different heteroaryl rings as linker and cap (Figure 5a). As exemplified by compounds 40–42, these agents possess nanomolar activity against HDAC6 and a hundred-fold to a thousand-fold selectivity over HDAC1 [116–118]. In 2016–2017, Takeda developed several classes of trifluoromethyl-1,2,4-oxadiazoles that use benzene or isoindolin-1-one rings as linkers (Figure 5b), exemplified by compounds 43–46 These compound were shown to completely inhibit HDAC6 at a concentration of 1 μM [119–122]. Notably, in 2019, Takeda published a series of difluoromethyl-1,3,4-oxadiazoles bearing isoindolin-1-one as the linker (for example, compound 47), showing comparable HDAC6 inhibition with isoindolin-1-one based analogs 44 and 45 [123]. Inspired by the 1,2,4-oxadiazole derivatives developed by Merck and Takeda, Karus also developed trifluoromethyl-1,2,4-oxadiazole based inhibitors containing thiophene as the linker connected to the typical capping groups they had employed previously (Figure 5c). As an example, compound 48 exhibited excellent HDAC6 selectivity over other isoforms [124]. Moreover, Oryzon Genomics patented trifluoromethyl-1,2,4-oxadiazoles containing pyridine as linker and heteroaryl rings as caps (e.g., 49) (Figure 5d) [125].

Figure 5.

Trifluoromethyloxadiazole derivatives as selective HDAC6 inhibitors

On the other hand, CKD has been further pursuing its discovery of 1,3,4-oxadiazole based selective HDAC6is (Figure 5e). Compounds 50–53 contain an N-(hetero)aryl bearing sulfonamide group as cap and a trifluoromethyl-1,3,4-oxadiazole or difluoromethyl-1,3,4-oxadiazole as ZBG [126, 127]. The majority of these compounds displayed potent activity against HDAC6 and a thousand-fold selectivity over HDAC1. Designed in analogy to their arylhydroxamate-based HDAC6is with a urea core (e.g., 34), 1,3,4-oxadiazoles (e.g., 54 and 55) demonstrated strong HDAC6 inhibition and modest HDAC1 activity, indicating improved selectivity [128]. Intriguingly, CKD also investigated the application of a 1,3,4-oxadiazole ZBG (56 and 57) in combination with the cap and linker contained in ACY-738, leading to selective HDAC6is as well [129]. Moreover, the effects of some of these compounds on mitochondrial axonal transport were investigated in hippocampal neurons from SD rat embryos, thus highlighting the therapeutic potential of novel 1,3,4-oxadiazole HDAC6is in neurodegenerative diseases.

6. Conclusion

In summary, we have updated the clinical developmental status of HDAC6is, showing that the majority of clinical trials still focus on various cancers. We have further summarized the pharmacological applications of selective HDAC6is in neurodegenerative diseases. Despite the fact that the discovery of novel HDAC6is has become an attractive area of research in the past few years, only a limited number of selective HDAC6is have been investigated in in vitro and in vivo models of neurodegenerative diseases. In addition, we have described the patent applications dealing with HDAC6is from 2014–2019 including the structures claimed and biological results, which show that arylhydroxamate-based HDAC6is still represent the predominate class, while the recent invention of oxadiazole-based HDAC6is may bring a new direction into this area of research.

7. Expert opinion

The recent disclosure of various crystal structures of HDAC6 complexed with different ligands facilitates the discovery of novel selective HDAC6 inhibitors. Moreover, mounting evidence suggests that selective HDAC6 inhibition is an effective therapeutic strategy to intervene in a variety of neurodegenerative diseases and neurological disorders.

Phenylhydroxamate, first reported as a ZBG/linker unit in Tubastatin A, has been considered the most useful linker and ZBG, leading to the identification of selective HDAC6is at the preclinical stage, thus overcoming the selectivity concern of broad spectrum HDACis. However, toxicity and poor stability remain hurdles in the advancement of hydroxamate HDAC6is to the clinical stage. Some hydroxamic acids are readily hydrolyzed to carboxylic acids in plasma due to the presence of esterases resulting in a high plasma clearance and poor DMPK profiles [130, 131]. Moreover, preclinical stage HDAC6is, such as TubA and SW-100, were found to exhibit short half lives in hepatocellular stability assays, thereby resulting in modest oral bioavailability [41, 42]. Additionally, the hydroxamate-based HDAC6is may possess mutagenic potential due to Lossen rearrangement leading to mutagenic isocyanates [8]. The hydroxylamine released upon enzymatic hydrolysis of the hydroxamate is a mutagen as well [132]. All hydroxamate HDACis on the market are positive in the Ames test and cause chromosomal aberrations in rodent cells. Although several HDAC6is have shown negative results in the Ames test [31, 42, 133], full genotoxicity evaluation using chromosomal aberration and micronucleus assays is needed prior to advancing them into clinical trials for application beyond cancer. To function as a therapeutic treatment for CNS diseases, drug candidates are required to efficiently cross the blood-brain-barrier (BBB). Due to the high polarity of the hydroxamate group, the discovery of brain-penetrant HDAC6is remains challenging. It has been reported that high pK_a values of capping groups of HDAC6is, which often contain open-chain or heterocyclic amine functionality, may enhance the compounds’ zwitterionic character, which would result in poor membrane permeability [42]. On the other hand, bulky, hydrophobic caps of HDAC6is are beneficial for enhancing their lipophilicity. However, such groups may result in higher plasma or brain homogenate binding thereby decreasing the concentration of the unbound drug [134].

To overcome such limitations, the mercaptoacetamide group has been investigated as an alternative ZBG that can lead to selective HDAC6is. The recently reported crystal structure of a HDAC6-mercaptoacetamide ligand complex revealed that the thiol group of the mercaptoacetamide is ionized to form a negatively charged thiolate that coordinates to the Zn²⁺ ion present in the active site of HDAC6 [135]. Furthermore, its NH group donates a hydrogen bond to H574 in the HDAC6 pocket, which is not formed with H142 in the active site of smHDAC8, thereby leading to a 200-fold HDAC6 selectivity over HDAC8. Additionally, the mercaptoacetamide-based HDAC6i W-2 exhibited promising results in AD models, while continued SAR studies have led to more potent and selective HDAC6is with improved BBB permeability [136, 137]. However, the mercaptoacetamide moiety exhibits several liabilities, namely it is able to undergo ready dimerization (to form a disulfide) and its poor metabolic stability in liver microsome studies, which thus limits its development as a drug candidate [137]. The recent discovery of oxadiazole-based HDAC6is led by pharmaceutical companies has provided a new direction for the HDAC field. Oxadiazole motifs have often been used in drug discovery projects, as they represent bioisosteric replacements for ester and amide functionalities [138]. The replacement of hydroxamate with oxadiazole-based ZBGs provides an opportunity to refine the ADMET profiles of HDAC6is, especially in regard to their possible genotoxicity. 1,3,4-Oxadiazole and 1,2,4-oxadiazole moieties have been successfully applied to the discovery of novel HDAC6is. Pharmacological HDAC6 inhibition experiments indicate that different types of HDAC6i share a similar capacity to restore impaired α-tubulin acetylation, rescue mitochondrial transport deficits, and promote protein aggregate formation/degradation, thereby showing an ability to rescue disease phenotypes of distinct neurodegenerative disorders. Notably, some oxadiazole-based HDAC6is developed by CKD have been investigated in axonal transport in vitro, which suggests that they may be effective in models of neurodegenerative diseases.

Article highlights.

• The discovery of novel HDAC6 inhibitors has become an attractive area of research in the last few years leading to the generation of diverse HDAC6 inhibitors, thus overcoming the selectivity concern of broad spectrum HDAC inhibitors.

• Five HDAC6 inhibitors are being investigated in clinical trials for different types of cancer, autoimmune diseases, and peripheral pain, but none of them has been advanced into the clinic for the treatment of a neurodegenerative disease.

• Only a limited number of candidates have been investigated that show effective results in models of neurodegenerative diseases.

• Arylhydroxamate has been considered the most useful linker and ZBG for discovering selective HDAC6 inhibitors, while toxicity and poor stability remain hurdles in the advancement of hydroxamate HDAC6 inhibitors to the clinical stage.

• Oxadiazole-type ZBGs have been widely applied to the discovery of novel selective HDAC6 inhibitors, which provides an opportunity to refine the ADMET profiles of HDAC6 inhibitors, especially in regard to their possible genotoxicity.