Design, Synthesis, and Biological Evaluation of HDAC Degraders with CRBN E3 Ligase Ligands

Histone deacetylases (HDACs) play important roles in cell growth, cell differentiation, cell apoptosis, and many other cellular processes. The inhibition of different classes of HDACs has been shown to be closely related to the therapy of cancers and other diseases. In this study, a series of novel CRBN-recruiting HDAC PROTACs were designed and synthesized by linking hydroxamic acid and benzamide with lenalidomide, pomalidomide, and CC-220 through linkers of different lengths and types. One of these PROTACs, denoted 21a, with a new benzyl alcohol linker, exhibited comparably excellent HDAC inhibition activity on different HDAC classes, acceptable degradative activity, and even better in vitro anti-proliferative activities on the MM.1S cell line compared with SAHA. Moreover, we report for the first time the benzyl alcohol linker, which could also offer the potential to be used to develop more types of potent PROTACs for targeting more proteins of interest (POI).

In the past few decades, hundreds of HDAC inhibitors (HDACi) have been developed for the treatment of many malignant and non-malignant diseases [7,8]. Typical HDACis can be categorized into four subtypes, based on their chemical structure: (1) short-chain fatty acid; (2) hydroxamic acid; (3) benzamides; and (4) cyclic peptides [9][10][11]. Despite its structural distinctiveness, HDACi is generally considered to share a common key motif, a "war-head", such as hydroxamic acid, that interacts with the Zn + in the active site, which is crucial for histone deacetylation activity. To date, hundreds of hydroxamic acid-and benzamide-type HDACi have been submitted to clinical trials (Figure 1), such as vorinostat (SAHA), panobinostat, belinostat, and chidamide [12][13][14]. Unfortunately, these drugs exhibit limited selectivity, which has been attributed as a cause of their debilitating side effects in patients [12,15]. Hence, it is urgent to develop an approach to enhance selective HDAC binding. Recently, a new technology, the proteolysis-targeting chimera (PROTAC) strategy, has attracted a great deal of attention in drug discovery due to its unique mechanism, which was first proposed by Crews and Deshaies in 2001 [16][17][18][19][20]. PROTACs are heterobifunctional molecules consisting of three parts: an E3 ligase system recruiting ligand acting as an "anchor"; a protein of interest (POI) binding moiety acting as a "warhead"; and a chemical linker acting as a "bridge" [21]. Through binding to both proteins in cells, a POI is recruited by PROTACs to a ternary complex (TC) with the E3 ligase, which triggers the ubiquitin-proteasome system (UPS), causing the degradation of the targeted protein by 26S proteasome [22,23]. Recently, different completely new classes of small-molecule therapeutic agents PROTACs have been discovered and developed with cereblon (CRBN) and von Hippel-Lindau (VHL) as E3 ligases ( Figure 2). These have attracted much attention [24,25]. Thus, employing PROTAC technology in HDAC degradation is a promising technology in drug discovery.
C-220 (Figure 3), has been designed as the cereblon modulator, which is currently in clinical trials for the treatment of systematic lupus erythematosus (SLE) and relapsed and refractory multiple myeloma (MM) [26,27]. CC-220 has been demonstrated to bind to CRBN through the glutarimide moiety, same as lenalidomide and pomalidomide, and to feature a higher affinity for CRBN than these moieties [28]. Typically, most of the PROTACs with lenalidomide and pomalidomide as CRBN ligands were linked with the "warhead" in the four-amino position with an alkyl linker or a polyethylene glycol linker [29,30]. To the best of our knowledge, no degraders of HDACs with benzyl alcohol as a linker have been reported, and whether this type of PROTAC can promote degradation remains to be determined ( Figure 3).  Based on the above reasons, accordingly, we designed and synthesized a series of novel CRBN-recruiting HDAC PROTACs by linking hydroxamic acid and benzamide with lenalidomide and pomalidomide through linkers of different lengths and types ( Figure 3). Further, we report for the first time the benzyl alcohol linkage between the four-position of lenalidomide or pomalidomide and the "warhead", which can be used in other types of PROTACs ( Figure 3C). The resulting products were further evaluated in HDAC inhibition studies, in vitro anti-proliferative studies, and western blot assay; profiling in these cellbased assays revealed that our bifunctional small molecules selectively promoted the degradation of HDAC through the UPS.

Chemistry
The synthesis of the presented final compounds was outlined in Schemes 1-3. The synthetic routes towards 9a-9e from the commercially available compounds 1a-1d are outlined in Scheme 1. The synthesis of the intermediates 2a-2d and the linkers 4a-4d followed published procedures [31][32][33][34]. Next, with lenalidomide (5a) and pomalidomide (5b) as the CRBN ligands, the key intermediates (6a-6e) with varying linker lengths were obtained through an amide reaction followed by the deprotection of the benzyl group using palladium on carbon (Pd/C) in a hydrogen atmosphere with high yields. Subsequently, a condensation reaction was applied to couple 7a-7e with triphenylmethyl (Trt) hydroxylamine, using HATU as a condensation reagent, to obtain 8a-8e in yields between 62 and 79%. The subsequent Trt-deprotection of compounds 8a-8d and 8e created the corresponding lenalidomide-conjugated compounds 9a-9d and pomalidomide-conjugated compound 9e. Scheme  As shown in Scheme 2, the synthesis of compounds 15a-15d commenced with the conversion of lenalidomide (5a) and pomalidomide (5b) to carbamates 11a and 11b, respectively, using four-nitrophenyl chloroformate in THF. The reaction of 11a and 11b with amino acids (12a-12c) with varying lengths under mild conditions created the carboxylic acid 13a-13d, which were subsequently reacted with triphenylmethyl (Trt) hydroxylamine to create 14a-14d in acceptable yields. Similarly, 14a-14d also underwent mild acidic deprotection to create the corresponding lenalidomide-conjugated compounds 15a-15c and pomalidomide-conjugated compound 15d.
As shown in Scheme 3, 21a-21b and 23a-23b were synthesized according to the following procedure. Using lenalidomide-4-OH (16a) and pomalidomide-4-OH (16b) as starting materials and CRBN ligands, the compounds 19a-19b were obtained through a substitution reaction resulting in 18a-18b, followed by removal of the Boc-protecting group under acidic conditions. Employing compounds 19a-19b as intermediates, the final compounds 21a-21b were synthesized by condensation with o-phenylenediamine (20) and 23a-23b were synthesized by condensation with O-tritylhydroxylamine followed by the subsequent removal of the Trt group.

Discussion and Biological Evaluation
Because the length and type of the linker plays a key role in the potency of PROTAC degrader molecules, 9a-9e and 15a-15d, with varying linker lengths, were synthesized through amide linkage and urea linkage, respectively. The molecules 21a-21b and 23a-23b were also synthesized with benzyl alcohol linkage under moderate conditions. It should be noticed that inhibition does not always correlate with degradation [35]. The HDAC inhibitory ability of compounds is one of the most critical parameters in HDACi, and HDAC degraders. Thus, the HDAC inhibitory ability of these molecules was first evaluated on several recombinant HDAC enzymes, including HDAC1, HDAC3, and HDAC6, at a concentration of 20 µg/mL, with SAHA as the positive control. As shown in Table 1 and Figures S1-S3, 9b, 9d, 9e, 15a, 15b, 15c, and 15d exhibited high HDAC1 or HDAC3 inhibitory abilities with a nano-molar level IC 50 value. Unfortunately, compounds 9a and 9c displayed sharply decreased potencies in the inhibition of HDAC1 and HDAC3, suggesting that the inhibitory abilities are related to the linker lengths. The molecules 21a-21b and 23a-23b, with the benzyl alcohol linkage, also displayed similar or slightly weaker inhibitory abilities than SAHA, suggesting their potential use in the development of more types of potent PROTACs. Surprisingly, most of the molecules exhibited much weaker HDAC6 inhibitory abilities compared with the control SAHA, and the mechanism needs further elucidation.
Based on the high HDAC inhibition activities of the molecules, we next investigated the in vitro anti-proliferative activities of the newly synthesized compounds. It was reported previously that MM.1S cells were more sensitive to HDAC PROTAC molecules [1,12]. Consequently, in this study, in vitro anti-proliferative activities of the molecules chosen from the in vitro HDAC inhibition was performed against the MM.1S cell line. As shown in Table 2 and Figure S4, 21a exhibited the most potent in vitro anti-proliferative activities against the MM.1S cell line with an IC 50 value of 0.043 µM, which was approximately five times more potent than that of SAHA. Meanwhile, 21b also exhibited comparable anti-proliferative activity with SAHA, with an IC 50 value at a nano-molar level. However, the anti-proliferative activities of 9b, 9e, 15a, and 15d were significantly weaker than those of 21a, 21b, and 23b. A possible explanation for this phenomenon may be that degraders with a shorter linker could induce the degradation of HDAC and further increase the anti-proliferative activities, which deserves further immunoblot studies. Interestingly, the HDAC inhibitory abilities of 9b, 9e, 15a, and 15d were significantly stronger than those of 21a, 21b, and 23b. The discrepancy between IC50s in the inhibition of HDAC6 and the anti-proliferation for compound 21a strongly suggests that the anti-proliferation of 21a operates under a very different mechanism.
Subsequently, to evaluate the degradation potential of the molecules, we treated the MM.1S cells with 21a, 21b, and 23b with different concentration gradients based on the previous in vitro anti-proliferative activities for 3 h, and checked the HDAC1, HDAC3 and HDAC6 protein levels by immunoblotting. As illustrated in Figure 4A, the amount of HDAC proteins remained the same when the cells were treated with different concentrations of 21b and 23b, respectively, suggesting almost no degradation of HDAC1, HDAC3, or HDAC6. These results were significantly different from the anti-proliferative results, indicating that 21b and 23b function through an enzyme inhibition mechanism rather than through degradation. Among the three tested molecules, degrader 21a demonstrated superior HDAC6 degradation efficacies at a concentration of 0.12 µM to HDAC1 and HDAC3. The selective degradation of HDAC6 over HDAC1 and HDAC3 was unexpected, but not surprising, because it has been reported that selective degraders can be developed by tethering non-selective BRD ligands with an E3 ubiquitin ligase ligand. Compared with the HDAC inhibitory ability results, the HDAC6 degradation ability suggested that 21a functions through a degradation mechanism.  Comp.
0.20 ± 0.03 a Data are the mean ± SD value of three independent determinations. b Positive control.
To further investigate the mechanism of the action of degrader 21, the effective concentration range and dependence of 21a towards HDAC6 degradation was conducted in the MM.1S cell line, with SAHA as a control. As shown in Figure 4B, the degradation of HDAC6 by 21a was observed in a dose-dependent manner. However, in the present study, a notable increase in tubulin acetylation was not observed and the degradation potency of 21a still needs to be improved in further study. Overall, these results indicated that 21a offers the potential to selectively degrade HDAC6 and suggested the possibility of using benzyl alcohol linkage to develop more types of potent PROTACs, which deserves further evaluation.

Conclusions
In this study, a series of HDACi-based PROTACs were designed and synthesized by linking hydroxamic acid and benzamide with lenalidomide, pomalidomide, and CC-220 through linkers of different lengths and types as a novel class of HDAC degraders. Overall, with the synthesis of these heterobifunctional molecules, we have demonstrated the feasibility of the targeted degradation of HDAC. It seems likely that the linker lengths and types of HDAC degraders will exert a profound effect on their HDAC and antiproliferative activity in cells. Among these degraders, 21a, with a new benzyl alcohol linker exhibited comparably excellent HDAC inhibition activity on different HDAC classes and degradative activity on the MM.1S cell line compared with SAHA. Furthermore, 21a also exhibited even better in vitro anti-proliferative activities on the MM.1S cell line, demonstrating great therapeutic potential for the treatment of certain disorders through HDAC degradation. Taken together, these results demonstrate important parameters for PROTAC design and reveal that further optimization is still needed to develop this novel type of HDAC technology. Moreover, we report for the first time the benzyl alcohol linkage, which could also offer the potential to be used to develop more types of potent PROTACs for targeting more POIs.

Materials
All the reagents were obtained from commercial sources and dried prior to use, unless otherwise stated. Anhydrous solutions of reaction mixtures were transferred via an ovendried syringe or cannula. All the reactions were monitored by thin-layer chromatography (TLC) on 25.4 mm × 76.2 mm silica 394 gel plates GF-254 and UPLC-Mass on Waters ACQUITY UPLC H-Class. The 1 H NMR and 13 C NMR spectra were recorded at 23 • C in CDCl 3 and DMSO-d 6 on a Bruker DRX-400 (400 MHz) using TMS as the internal standard. Chemical shifts were reported as δ (ppm) and signal splitting patterns were described as singlet (s), doublet (d), triplet (t), quartet (q), quintet (quint), or multiplet (m), with coupling constants (J) in hertz. High-resolution mass spectra (HRMS) were obtained on an electron spray injection (ESI) Thermo Fisher Scientific LTQ FTICR mass spectrometer. The purity of all the tested compounds was ≥ 95%, as estimated by HPLC analysis performed on Agilent Diamonsil C18 (250 mm × 4.6 mm).

Experimental Procedures
Compounds 2a-2d were synthesized according to literature reports [31,32]. General procedures for the preparation of compounds 6a-6e. To a solution of 4a-4d (1.0 eq) in anhydrous DMF, HATU (2.0 eq) and DIEA (3.0 eq) were added. After stirring in an ice bath for 0.5 h, 5a or 5b (1.0 eq) was added. The mixture was stirred at room temperature overnight. After the reaction completed, the residue was diluted with H 2 O and extracted with EtOAc. The combined organic layers were then washed with HCl (0.1 M), saturated brine, dried over anhydrous Na 2 SO 4 , and concentrated in vacuo to afford the crude product, which was purified by column chromatography with DCM/ MeOH (30:1-20:1) to produce 6a-6e. General procedures for the preparation of compounds 7a-7e. To a solution of 6a-6e (1.0 eq) in MeOH, 10% Pd/C (10% eq) was added. The reaction mixture was purged with N 2 and then H 2 and stirred at room temperature for 2 h. After the reaction completed, the residue was filtered through celite and washed with DCM. The combined organic layers were then washed with saturated brine, dried over anhydrous Na 2 SO 4 , and concentrated in vacuo to produce 7a-7e. General procedures for the preparation of compounds 8a-8e. To a solution of 7a-7e (1.0 eq) in anhydrous DMF, HATU (2.0 eq) and DIEA (3.0 eq) were added. After stirring in an ice bath for 0.5 h, O-tritylhydroxylamine (1.0 eq) was added. The mixture was stirred at room temperature overnight. After the reaction completed, the residue was diluted with H 2 O and extracted with EtOAc. The combined organic layers were then washed saturated brine, dried over anhydrous Na 2 SO 4 , and concentrated in vacuo to produce the crude product, which was purified by column chromatography with DCM/ MeOH (30:1-20:1) to produce 8a-8e. General procedures for the preparation of compounds 9a-9e. To a solution of 8a-8e (200 mg) in TFA (0.3 mL) and DCM (6 mL), triethylsilane (0.15·mL) was added dropwise. The mixture was stirred at room temperature for 30 min. After the reaction completed, the mixture was concentrated in vacuo to create the crude product, which was purified by column chromatography with DCM/ MeOH (20:1-15:1) to produce 9a-9e.  13  General procedures for the preparation of compounds 11a-11b. To a solution of 5a or 5b (1.0 eq) in THF, 10 (1.5 eq) was added. The mixture was stirred at reflux for 4 h. After the reaction completed, the mixture was cooled to room temperature, filtered and washed with THF to produce 11a-11b. General procedures for the preparation of compounds 13a-13d. To a solution of 11a or 11b (1.0 eq) in DMF, 12a-12c (1.5 eq) was added. The mixture was stirred at room temperature overnight. After the reaction completed, was diluted with H 2 O and extracted with EtOAc. The combined organic layers were then washed saturated brine, dried over anhydrous Na 2 SO 4 , and concentrated in vacuo to create the crude product, which was purified by column chromatography with DCM/ MeOH (15:1-10:1) to produce 13a-13d.

In Vitro HDAC Inhibition Assay
All three full-length rhHDACs were expressed in insect High5 cells using a baculoviral expression system, and all the His6-tagged and GST-fusion proteins were purified using Ni-NTA (QIAGEN). The deacetylase activity of rhHDACs (recombinant human HDACs) 1 and 3 were assayed with an HDAC substrate (Ac-Lys-Tyr-Lys(ε-acetyl)-AMC), and the HDAC6 was assayed with another HDAC substrate (Boc-Lys(ε-acetyl)-AMC).The total HDAC assay volume was 25 µl and all the assay components were diluted in a Hepes buffer (25 mM Hepes, 137 mM NaCl, 2.7 mM KCl, and 4.9 mM MgCl2; pH 8.0).The reaction was carried out in black 384 well plates (OptiPlateTM-384F, PerkinElmer). In brief, the HDAC assay mixture contained an HDAC substrate (5-50 µM, 5 µl), rhHDAC isoforms (20-200 nM) and an inhibitor (1 µl). The positive controls contained all the above components except the inhibitor. The negative controls contained neither an enzyme nor an inhibitor. The HDAC6 assay components were incubated at room temperature for 3 h, and HDAC 1 and 3 were incubated for 24 h. The reaction was quenched with the addition of 25 µl trypsin (diluted to final concentration 0.3125%). The plates were incubated for 30 min at room temperature to allow the fluorescence signal to develop. The fluorescence generated was monitored at wavelengths 355 nm (excitation) and 460 nm (emission) using Envision (PerkinElmer).

Anti-Proliferative Activities
The cells were seeded in 96 well plates at a proper density in growth media. After 24 h, a range of concentrations of the test compounds was added and the plates were incubated at 37 • C for 72 h. The cell proliferation was determined by the MTS assay. After 3 h, the absorbance was collected using SpectraMAX 340. The IC 50 values were calculated by fitting a concentration-response curve using a SoftMax pro-based four-parameter method.

Western Blot Analysis
The 2 × 10 6 MM.1S cells (multiple myeloma) were cultured in a cell culture plate and incubated for 1 h. Next, the MM.1S cells were treated with compounds at different concentrations for 3 h. Subsequently, the cells were collected, washed, lysed in a Ripa buffer containing a sample reducing agent, and analyzed using SDS/PAGE. Finally, the gel was blotted and the total HDAC levels were detected through standard Western blot.
Supplementary Materials: Figure S1: IC50 curve of synthesized molecules and SAHA against HDAC1; Figure S2: IC50 curve of synthesized molecules and SAHA against HDAC3; Figure