Controlling Plasma Stability of Hydroxamic Acids: A MedChem

Controlling Plasma Stability of Hydroxamic Acids: A MedChem Toolbox
Paul Hermant, Damien Bosc, Catherine Piveteau, Ronan Gealageas, Baovy Lam, Cyril Ronco, Matthieu Roignant, Hasina Tolojanahary, Ludovic Jean,
Pierre-yves Renard, et al.
To cite this version:
Paul Hermant, Damien Bosc, Catherine Piveteau, Ronan Gealageas, Baovy Lam, et al.. Controlling Plasma Stability of Hydroxamic Acids: A MedChem Toolbox. Journal of Medicinal Chemistry, American Chemical Society, 2017, 60 (21), pp.9067-9089. ￿10.1021/acs.jmedchem.7b01444￿. ￿hal-03054238￿

HAL Id: hal-03054238 https://hal.archives-ouvertes.fr/hal-03054238
Submitted on 11 Dec 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Controlling Plasma Stability of Hydroxamic Acids: A MedChem Toolbox.
Paul Hermant,#1 Damien Bosc,#1 Catherine Piveteau,#1 Ronan Gealageas,1 BaoVy Lam,1 Cyril Ronco,1 Matthieu Roignant,1 Hasina Tolojanahary,1 Ludovic Jean,2 Pierre-
Yves Renard,2 Mohamed Lemdani,3 Marilyne Bourotte,1 Adrien Herledan,1 Corentin Bedart,1 Alexandre Biela,1 Florence Leroux,1 Benoit Deprez,1 Rebecca Deprez-Poulain *,1,4
1. Univ. Lille, Inserm, Institut Pasteur de Lille, U1177 - Drugs and Molecules for Living Systems, F-59000 Lille, France; 2. Normandie Université, COBRA, UMR 6014 &
FR 3038, Université de Rouen, INSA Rouen, CNRS, F-76821 Mont-Saint-Aignan Cedex, France.; 3 Univ. Lille. EA 2694 - Santé publique : épidémiologie et qualité
des soins. F-59000 Lille. France. Institut 4. Institut Universitaire de France, F- 75231, Paris, France.
Abstract Hydroxamic acids are outstanding zinc chelating groups that can be used to design potent and selective metalloenzyme inhibitors in various therapeutic areas. Some hydroxamic acids display a high plasma clearance resulting in poor in vivo activity, though they may be very potent compounds in vitro. We designed a 57-member library of hydroxamic acids to explore the structure-plasma stability relationships in these series and identify both which enzyme(s) and which pharmacophores are critical for plasma stability. Arylesterases and carboxylesterases were identified as the main metabolic enzymes for hydroxamic acids. Finally, we suggest structural features to be introduced or removed to improve stability. This work provides thus the first medicinal chemistry toolbox (experimental procedures and structural guidance) to assess and control the plasma stability of hydroxamic acids and realize their full potential as in vivo pharmacological probes and therapeutic agents. This study is particularly relevant to preclinical development as it allows to obtain compounds equally stable in human and rodent models.

Introduction
The hydroxamic acid function is a strong zinc chelating group. It can be used to design potent and selective metalloenzymes inhibitors that can serve both as biological probes1 and leads. Hydroxamic acids can also be used as bioisosters of carboxylic acid thanks to their weak acid properties.2 In the field of infectious diseases, several hydroxamic acid series have been reported as antibacterials : inhibitors of peptide deformylase (PDF)3, inhibitors of the neurotoxin A metalloprotease of bacterium Clostridium botulinum (BoNTA)4, or inhibitors of UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase (LpxC) of gram-negative bacteria.5 Antivirals targeting HCV replication via a yet unknown mechanism,6 HIV integrase inhibitors,7 or inhibitors of metalloproteases of the Plasmodium falciparum parasite,8 have also been disclosed. Recently, inhibitors of the essential nematode-specific metalloprotease DPY31 were discovered via combined in silico and experimental methods.9 Several groups have explored the use of hydroxamic acids in other therapeutic areas inhibitors of glutamate carboxypeptidase II (GCPII) in neuropathic pain;10 inhibitors of Insulin-Degrading Enzyme in type-2 diabetes11, inhibitors of Matrix metalloprotease (MMP) in fibrinolysis control12. In osteoarthritis, matrix-metalloproteinase13 or aggrecanase inhibitors14 have been proposed, while inhibitors of Tumor Necrosis Factor Converting Enzyme (TACE, ADAM17) were designed for autoimmune diseases such as psoriasis or Crohn’s disease.15-16 In the area of cancer, several hydroxamic acids have been designed to inhibit various proteases such as the sheddase of HER-2.17 Importantly, with the growing interest in epigenetics, many teams explored the possibility to inhibit other zinc hydrolases, like histone deacetylases (HDACs). Several hydroxamic acid inhibitors of these targets have already been approved: vorinostat, and belinostat for T-cell lymphoma and recently panobinostat for multiple myeloma (Figure 1). Many other hydroxamic acids are currently following on these first clinical successes, not only in cancer, but also in several other therapeutic areas.18-19

vorinostat (SAHA)

belinostat

panobinostat

Figure 1: Marketed hydroxamic acid drugs.

The hydroxamic acid function undergoes phase I and phase II metabolism: it is mainly

hydrolysed to the corresponding carboxylic acid,20 but also it can be reduced to the

corresponding amide21, O-glucuronylated or O-sulfated22. The carboxylic acid is usually much

less active on the target and possesses different physical-chemical and permeability properties,

than its hydroxamic acid analogue.23 Depending on the fragility of the hydroxamic acid

function, and the substituents in α- or β-position,24 metabolism is driven towards the production

of carboxylic acid, or the production of O-glucuronides. For example, the main metabolite of

belinostat is the glucuronide form.25

We and other groups have tried to improve the pharmacokinetics properties of

hydroxamic acids thanks to chemical modulation.8,26 We have published a preliminary study of

in vitro structure-plasma stability relationships (SPSR) of such compounds.24 Here, we show

that susceptibility to the esterase activity in plasma is a critical driver of the stability of

hydroxamic acids.

Although often overlooked, esterases are worth considering in medicinal chemistry as

they are involved in the metabolism of 10% of drugs.27 Also, in specific situations, esterase

activity can be used to release a parent drug from the ester prodrug (e.g. angiotensin-converting

enzyme inhibitors, oseltamivir, tenofovir disoproxil)28 or to inactivate rapidly a compound like

in the soft-drug strategy29 (ROCK inhibitors30). Esterases can be sorted in two main types. Type

A (arylesterases) includes paraoxonase-1 (PON-1) and other aryl esterase activities. Type B

(serine esterases) includes acetyl- and butyl-cholinesterases (AchE and BuChE) as well as

carboxylesterases (CES1 and 2) (Table 1). Along with these main enzymes, other esterases like

arylacetamide deacetylase and acylpeptide hydrolase have been described.27 Last but not least,

Catalytic EC# Acro Name

albumin, which is the most abundant protein in plasma, possesses a pseudo-esterase activity

resulting in acylation of several of its nucleophilic residues such as Tyr411.31

mechanism

Table 1:Esterases classification.a

Type A esterases

Paraoxonase

Other aryl esterase activities

PON1

ArE

nym

3.1.8.1

3.1.1.2

Catalytic dyad His/His Ca2+ activated

undefined

Type B esterases

Cholinesterases

Carboxylesterases

AChE 3.1.1.7

BuChE 3.1.1.8

CES 3.1.1.1

Serine catalyzed Catalytic triad Ser, Glu, His

Substrates & Inhibitors

Others
Albumine
SA
6.1.1.16
Irreversible acetylation of Tyr411 and
other residues

RCOOAr R1O (R2O)-PO-OAr

RCOOPh

CH3COOR

BuCOOR, AlkCOOR

numerous prodrugs R1COOR2

CH3COOR

EDTA

DTNB

PMSF; tacrine; PMSF; tacrine;

huprine

profenamine

PMSF; BNPP

undefined

a EDTA: Ethylenediaminetetraacetic acid; DTNB: 5,5-dithio-bis-(2-nitrobenzoic acid); PMSF: Phenylmethylsulfonyl fluoride; BNPP: bis-para-nitrophenylphosphate.

Selective Substrates Inhibitors

These enzymes display different substrate specificities (Table 1). Using a histidine dyad,32 PON-1 hydrolyzes both arylesters and phosphotriesters and is inhibited by chelating agents such as EDTA. On the opposite, other plasma arylesterase activity do not have the ability to cleave the phosphotriesters33 and are inhibited by 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB).34 BuChE displays a larger hydrophobic pocket that accounts for its less stringent substrate specificity over AChE. Both AChE and BuChE are inhibited by phenylmethylsulfonyl fluoride (PMSF), an irreversible inhibitor of serine proteases and tacrine, a prototypal cholinesterase inhibitor, or by specific non-covalent inhibitors (huprines35 and profenamine respectively). CES1 and 2 are both inhibited by PMSF and bis-(4-nitrophenyl)-phosphate (BNPP) but CES1 recognizes substrates with either small or large acyl groups.36

Importantly, tissue distribution of these enzymes differs among species,37 in particular in plasma.38-39 While CES are the major esterases in rodent plasma, they are absent in human plasma.40 On the contrary, AChE and BuChE are present at lower concentrations in rodent plasma. The major esterase in human plasma is BuChE while AChE is only present in trace amounts.41 Along with the plasma isoform of AChE, the E isoform of AChE (also called H for hydrophobic) is found in erythrocytes.42 Interestingly, PON-1 which is associated to HDL, displays a high polymorphism that impacts its expression and activity.32,43 Moreover, along with a presence in the serum, some esterases can be found in both liver and intestine. Importantly, CES1 is largely expressed in liver microsomes while CES2 is expressed in intestine microsomes,40 where they participate actively in drug metabolism.28 For these enzymes also, species differences in distribution have been reported.39,40
Compounds that are unstable in plasma tend to display a high clearance, resulting in poor oral bioavailability and thus poor or undetectable activity, though they may be very potent compounds in vitro. The utility of such compounds as tools to probe in vivo pharmacokinetic/pharmacodynamic relationships in animal disease models in early drugdiscovery is thus compromised.44
The goal of this study is to establish Structure Plasma Stability Relationships (SPSR), identify enzymes at play in hydrolysis, and provide structural guidance to chemists in the design of hydroxamic acids that show sufficient stability in rodents. Obtaining hydroxamates that are equally stable in rodent and human is key to engage the target in preclinical disease models at the lowest dose, generate animal proof-of-principle and make a first assessment of the safety window. This is of particular importance as rodent plasma is much more aggressive to hydroxamic acids than human plasma, a cause of poor systemic exposure in preclinical models. A common practice in drug discovery is to optimize compounds potency on both a human target and its rodent ortholog, as well as stability on human and rodent liver microsomes to allow for

a proof of concept in animal models. Optimization of rodent plasma stability is of similar importance for the particular case of hydroxamic acids, even if the lead compound is already stable in human plasma.
In order to provide medicinal chemists with tractable information for improving stability of hydroxamic acids, we report here a set of protocols and results (the toolbox) that comprises tools to assess plasma stability of hydroxamic acids, structure-plasma stability relationships and dashboards for the identification of involved esterases.
Figure 2: Hydroxamic acid library design.a
a n= 0-6; m =1-3; In gray highlight: diversity within each hydroxamic acid library subset; Cy = Bz or cHx or adamantyl; Ar = Bz, p-fluorobenzyl, 4-pyridinyl, biphenyl, heteroaryl.
Results
Library Design. To maximize the relevance of our study for medicinal chemists, we selected
substructures found in a wide range of biologically active hydroxamic acids. The resulting 57member library inspired from inhibitors of several classes of Zn metallohydrolases, allows in turn to explore different esterase pockets. We study the impact of chain length and volume (112, Figure 2A), nature of the linker (1,13-21, Figure 2B); α-substituents (13,20,22-51, Figure 2C); unsaturation (52-57, Figure 2D) and cyclic constraints (36,42,44,50-51, Figure 2E), and compare compounds.

Among the compounds selected, linear alkyl hydroxamic acid series (Figure 2A) with various chain length and terminal cycles or linkers (Figure 2B) have been partially inspired by HDAC inhibitors and BoNTA inhibitors.4 Diversely α-substituted hydroxamic acids (Figure 2C) were inspired by PfAM1 malonic inhibitors,8 LpxC inhibitors45 and hIDE inhibitors11. Other compounds were selected by homology to sulfonamides or sulfone-derived inhibitors of matrix-metalloproteases (MMP).46 Hydroxamic acids inspired by inhibitors of aggrecanases (AGG)47 or TACE48,16 were also designed and synthesized. Series of compounds with an unsaturation in α-position of the hydroxamic acid function (Figure 2D) have been inspired by cinnamic HDAC inhibitors49-50, PfAM1 inhibitors8, or BoNTA inhibitors4. Finally, a number of hydroxamic acids displaying cyclic constraints in α-position (Figure 2E) complete the library. They were inspired by MMP51-52, LpxC53 and TACE54 inhibitors.
Synthesis of the library. The studied 1-57 compounds were obtained either as previously described24 (37-38, 40-
42, 57) or via diverse synthetic approaches (1-36, 39, 43-56) summarized in schemes 1-7. Hydroxamic acids, scheme 1, were obtained either from the corresponding carboxylic acid by coupling with O-tritylhydroxylamine,55 or from the corresponding ester by aminolysis with either hydroxylamine hydrochloride and KOH56 or hydroxylamine and catalytic KCN57 (12, 1517, 19-20, 30-37, 39, 43-45, 50-51). Non-commercial precursors (22b-29b, 46b-49b, 52b, 54b56b) were synthesized as described in schemes 2-7.
Commercial 2‐(adamantan‐1‐yl)ethan‐1‐ol was tosylated and the resulting sulfonate was substituted using cyanide to afford 58. Then, the nitrile was hydrolyzed under basic conditions giving the corresponding carboxylic acid, which was converted into the methyl ester 12a (Scheme 2).

Scheme 1. From carboxylic acid or ester precursors to final hydroxamic acids.a
a or b

12a, 15a-20a, 22a-26a, 29a-36a,39a, 43a-51a, 55a
c or d

1-36, 39, 43-56 e

1b-11b, 13b-14b, 21b-29b, 46b-49b, 52b-56b
a Reagents and conditions: (a) KCN, aq. NH2OH 50 wt.%, MeOH, room temp., 2 h – 48 h, 4%-88% ; (b) NH2OH.HCl, KOH, MeOH, room temp., 16 h, 36% yield for 18; (c) (i) 2M NaOH, EtOH/THF, room temp., 16 h, (ii) 1N HCl, 52%-99%; (d) TFA, CH2Cl2, room temp., 2 h, quantitative yield; (e) (i) O-tritylhydroxylamine, EDCI, HOBt, N-methylmorpholine, DMF, room temp., 16 h; (ii) TFA, TIS, CH2Cl2, room temp., 15 min, 8%-99%.

Scheme 2. Synthesis of adamantane derivative 12a.a

a, b

c, d

58

12a

a Reagents and conditions: (a) 4-methylbenzenesulfonyl chloride, pyridine, room temp., 20 h, 75%; (b) KCN, DMF, 80 °C, 15 h, 84%; (c) KOH, EtOH, H2O, 90 °C, 20 h, 97%; (d) SOCl2, MeOH, 0 °C to room temp., 4 h, 21%.

Scheme 3. Synthesis of the secondary amines 17a, 27b-28b and cynnamyl precursors 52b, 54b, 55a and 56b.a

a

f

17a
b 52b

27b R = Me 28b R = Bn g,h
54b

c,d,e

i

55a

56b

a Reagents and conditions: (a) benzylamine, EtOH, room temp., 18 h, 84%; (b) malonic acid, pyridine, reflux, 1 h, quantitative yield; (c) MeOH, H2SO4, reflux, 16 h, 79%; (d) N,N-diethylethylenediamine, Et3N, dioxane, 90 °C, 16 h, 65%; (e) valeraldehyde, SnCl2.2H2O, AcOH, MeOH, 40 °C, 16 h, 48%; (f) (i) LiOH.H2O, MeOH, room temp., 20 min, (ii) benzaldehyde, room temp., 1 h, (iii) NaBH4, room temp., 30 min, 76%-100%; (g) chlorosulfonic acid, 10-12 °C, 16 h, 30%; (h) aniline, pyridine, CH2Cl2, room temp., 16 h, 66%; (i) iodobenzene, silver acetate, palladium acetate, acetic acid, 110 °C, 6 h, 61%.

Hydroxamic acid analogs bearing a secondary amine (17, 27 and 28) or a cinnamyl moiety (52 and 55) were synthesized as described in Scheme 3. The linear ester precursor 17a was obtained via an aza-Michael reaction whereas a reductive amination from benzaldehyde and the corresponding amino-acid afforded acidic precursors 27b and 28b.58 Benzenepropenoic acids 54b and 55a, analogues of HDAC inhibitors, were obtained according to the literature.4950 Compound 52b, the acid precursor of a botulinum neurotoxin A protease inhibitor59 was synthesized by a Döbner-Knöevenagel condensation from malonic acid and 2,4dichlorobenzaldehyde. At last, disubstituted alkene 56b was synthesized by palladiumcatalyzed Heck diarylation60 (Scheme 3).

Scheme 4. Synthesis of malonic precursors.a

b

c

59 R = H a 60 R = 2-naphtyl-CH2-

15b R = H, R1 = Ph 20b R = H, R1 = Bn 33b R = 2-naphtyl-CH2-, R1 = Bn

15a R = H, R1 = Ph 20a R = H, R1 = Bn 33a R = 2-naphtyl-CH2-, R1 = Bn

f

g

R = R' = H (commercial) 61 R = H, R' = 4-pyridinyl-CH - d
2

R = H, R' = Bn (commercial)

R = Me, R' = H (commercial)

62 R = Me, R' = Bn

e

R = R' = Me

R = R' = cyclopropyl

R = H 63 R' = Me 64 R' = Bn 65 R' = 4-pyridinyl-CH2-
66 R = Me, R' = Bn 67 R = R' = Me 68 R = R' = cyclopropyl

R1 = Bn 30a R = H, R' = Me 31a R = H, R' = Bn 32a R = H, R' = 4-pyridinyl-CH234a R = Me, R' = Bn 35a R = Me, R' = Me 36a R = R' = cyclopropyl
R = H, R' = Bn 39a R1 = pF-C6H4-CH243a R1 = 4-pyridinyl-CH2-

a Reagents and conditions: (a) (i) 2-naphthaldehyde, MeCN, room temp., 1.5 h, (ii) Hantzsch’s ester, L-proline, MeOH, room temp., 16 h, 68%; (b) aniline, MeCN, 80 °C, 2 h, 48% 15b, benzylamine, MeCN room temp., 3 h, 63% 20b, benzylamine, MeCN, MW 80 °C, 1.75 h, 95% 33b; (c) SOCl2, MeOH, 0 °C to room temp., 0.5-5.25 h, 43%-89%; (d) (i) NaH, THF, 0 °C to room temp., 30 min, (ii) 4-(chloromethyl)pyridine hydrochloride, 40 °C, 5.5 h, 23%; (e) (i) Sodium, abs. EthOH, 0 °C to 50°C, 1 h, (ii) bromomethylbenzene, 50 °C, 19 h, 37%; (f) KOH, EtOH, room temp., 24 h, quantitative yield; (g) amine, HBTU, DIPEA, DMF, 0 °C to room temp., 4h-18h, 41%75%.