The Trifluoromethyl Group as a Bioisosteric Replacement of the

May 3, 2019 - Kosterlitz Centre for Therapeutics, University of Aberdeen ... Department of Pharmacology & Toxicology, University of Toronto, Toronto M...
0 downloads 0 Views 485KB Size
Subscriber access provided by Bibliothèque de l'Université Paris-Sud

Article

The Trifluoromethyl Group as a Bioisosteric Replacement of the Aliphatic Nitro Group in CB1 Receptor Positive Allosteric Modulators (PAMs) Chih-Chung Tseng, Gemma Baillie, Giulia Donvito, Mohammed Mustafa, Sophie Juola, Chiara Zanato, Chiara Massarenti, Sergio Dall'Angelo, William T A Harrison, Aron H. Lichtman, Ruth Ross, Matteo Zanda, and Iain R. Greig J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00252 • Publication Date (Web): 03 May 2019 Downloaded from http://pubs.acs.org on May 3, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

The Trifluoromethyl Group as a Bioisosteric Replacement of the Aliphatic Nitro Group in CB1 Receptor Positive Allosteric Modulators (PAMs)

Chih-Chung Tseng,a Gemma Baillie,b Giulia Donvito,c Mohammed A. Mustafa,c Sophie E. Juola,c Chiara Zanato,a Chiara Massarenti,a Sergio Dall’Angelo,a William T. A. Harrison,e Aron H. Lichtman,c,f Ruth A. Ross,b Matteo Zanda,a,d,§,* Iain R. Greiga,*

a University

of Aberdeen, Kosterlitz Centre for Therapeutics, Foresterhill, Aberdeen, AB25 2ZD, Scotland, UK

b University

of Toronto, Department of Pharmacology & Toxicology, Toronto M5S 1A8, Canada

c

Department of Pharmacology and Toxicology, Virginia Commonwealth University, Richmond, Virginia,

23298

USA d

C.N.R. – I.C.R.M., via Mancinelli 7, 20131 Milan, Italy

e

Department of Chemistry, University of Aberdeen, AB24 3UE, Scotland, UK

f Department §

of Medicinal Chemistry, Virginia Commonwealth University, Richmond, Virginia, 23298, USA

Current address: Loughborough University, Centre for Sensing and Imaging Science, Sir David Davies building,

LE11 3TU Loughborough (UK).

ABSTRACT The first generation of CB1 positive allosteric modulators (PAMs; e.g., ZCZ011) featured a 3-nitroalkyl2-phenyl-indole structure. Although a small number of drugs include the nitro group, it is generally not regarded as being “drug-like”, and this is particularly true for aliphatic nitro groups. There are very few case studies where an appropriate bioisostere replaced a nitro group that had a direct role in binding. This may be indicative of the difficulty of replicating its binding interactions. Herein we report the design and synthesis of ligands targeting the allosteric binding site on the CB1 cannabinoid receptor, in which a CF3 group successfully replaced the aliphatic NO2. In general, the CF3-bearing compounds were more potent than their NO2 equivalents and also showed improved in vitro metabolic stability. The CF3-analogue (1) with the best balance of properties was selected for further

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

pharmacological evaluation. Pilot in vivo studies showed that (±)-1 has similar activity to (±)-ZCZ011, with both showing promising efficacy in a mouse model of neuropathic pain.

INTRODUCTION The endocannabinoid system plays a major role in modulating neurotransmitter release in the central nervous system and thus its modulation elicits a broad range of physiological effects. Compounds activating cannabinoid receptors represent a promising therapeutic opportunity for the treatment of many disorders, including depression, anxiety and pain. However, global activation of the cannabinoid CB1 receptor by direct agonism leads to psychoactive side effects, physical dependence and abuse liability. There is both an opportunity and a need to provide safer and more effective ways of harnessing the therapeutic benefits of CB1 receptor activation. Recent reviews have highlighted the therapeutic potential of targeting the allosteric binding site on the CB1 receptor using positive allosteric modulators (PAMs),1,2 which have the advantage of enhancing the actions of endocannabinoids directly at the orthosteric receptors expressed within spinal and brain pain processing centres. Thus, PAMs are predicted to be free of the psychoactive side-effects associated with direct CB1 agonists that globally activate these receptors. The impressive efficacy of small molecule tool CB1 PAMs in models of neuropathic pain and post-traumatic stress disorder, without any evidence of addiction or psychotropic side-effects, support this hypothesis.

S

NO2 N H

NO2 N H

ZCZ-011 (5)

AZ-4 / GAT211

Figure 1

The activity of the first CB1 PAM - AZ-4,3 also known as F-0874 and now commonly referred to as GAT2115 (Figure 1), was reported in a conference poster by researchers at AstraZeneca Montreal,3 having been first synthesised by Noland and Lange in 1959.6 The structure has since been modified to

ACS Paragon Plus Environment

Page 2 of 37

Page 3 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

improve potency (ZCZ011,7 5 in Table 1) and the enantiomers of AZ-4 separated (GAT228/229)5, but no substantial development or comprehensive structure activity relationships have been reported. The primary problematic issue with these compounds is the presence of an aliphatic nitro group. Despite its presence in a small number of major drugs, including nifedipine, chloramphenicol, ranitidine and metronidazole, most academic and industrial medicinal chemists do not regard a nitro group as being “drug-like”. However, surprisingly little published evidence exists to support this blanket exclusion of the nitro-group from drug-like space. Justification for this exclusion may be based upon accepted wisdom, or derived from personal experience (see e.g. Muegge and co-workers;8 Beck and co-workers;9 and discussions on internet forums, including the well-regarded “In the Pipeline”10,11); or based on a well-established (if only potential) toxic liability associated solely with the aromatic nitro and its propensity for formation of an aryl nitrenium ion, which can bind to DNA.12 In additional examples of this general dismissal of the nitro group from drug-like space, the nitro was one of the groups eliminated prior to the studies leading to Baell’s identification of the range of panassay interference compounds13; likewise, any structures containing a nitro are explicitly excluded by Rankovic and Morphy14 as being acceptable for use in fragment-based lead discovery.

While little information exists on the toxicological issues associated with aliphatic nitro groups, the widely-held opinions described above usually require its removal or replacement early in the development of a bioactive compound. However, there are only a few case studies in which a nitro group has been successfully replaced by an appropriate bioisostere,9 which may be indicative of the difficulty of replicating its binding interactions. Nitro groups may act as hydrogen bond acceptors and may also stack with other groups, including ketones and alcohols.15 Additionally, Alston and coworkers16 reported that the nitroalkyl group can serve as a one- and two-electron reductant, oxidant, ambident nucleophile, electrophile, ligand, and leaving group in reactions with enzymes and other biomolecules and suggested that “the richness of its reactivity may have hindered the application of the nitroalkyl group in medicinal chemistry.” Much more commonly, nitro groups are initially present

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

in bioactive compounds as aromatic substituents, where their electron-withdrawing properties may polarise the ring, giving optimal interaction with electron-rich moieties within the biological target. In such cases, simple replacement with other strongly-electron-withdrawing substituents, e.g. halogens or haloalkyl groups, may be possible.

Because of this strong electron-withdrawing nature, the presence of a nitro group also facilitates many chemical reactions; thus, it is often found in screening collections (including the collection from which AZ-4 was originally identified). The reaction of an indole with nitrostyrene has been widely reported in the literature and there are a considerable range of methods available for this simple 1,4-conjugate addition reaction.17–24 There is literature precedent for the nitro group to be replaced with a carboxylic acid in this series of compounds, which is achieved by initial reaction of an indole with the highlyelectron deficient diethylbenzylidene malonate.21,22 The limited similarity of nitro group binding to carboxylic acid binding, in spite of the near perfect isosterism, has been discussed by Kelly and Kim.25 A small range of derivatives can be accessed from the carboxylic acid (e.g. CH2OH, CO2H, CO2Et, CONHNH2, CONH2 and CONHMe, Figure 2 - see Supporting Information); however, these were inactive as CB1 PAMs.

NR1R2 N H

COR1 N H

Figure 2. Likewise, simple derivatives based on reduction of the nitro to an amine (Figure 2), with or without subsequent substitution, gave inactive compounds (see Supporting Information).

Modelling studies on GAT22826 indicate that the nitro group may form polar interactions with a serine, a methionine and a cysteine residue in the absence of the orthosteric agonist (CP55,940). However, this case appears to apply to an orthosteric agonist. Although not stated as a conclusion by the

ACS Paragon Plus Environment

Page 4 of 37

Page 5 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

authors, the modelling suggests that, in the presence of an agonist, GAT228 moves deeper into the receptor; accordingly, the nitro group may be too far from these binding groups to retain a strong interaction. Thus, it remains uncertain that specific binding interactions provided by the nitro group are required for its activity as a PAM. However, the lack of activity shown by any of the compounds in Figure 2 suggests that such interactions will be necessary for PAM activity.

Herein we report that the CF3 offers a viable bioisosteric replacement for an aliphatic nitro group, demonstrating the removal of the non-drug-like nitro group from the most-commonly-studied and utilised CB1 PAMs, to achieve compounds with improved potency and metabolic stability.

RESULTS AND DISCUSSION Chemistry. A library of indole derivatives 1-57 (Table 1) was synthesised using a variety of methods. Nitro-indole derivatives 2-25 were obtained by InBr3-catalyzed Michael addition of 2-(2-nitrovinyl)-aryl/heteroaryl derivatives to the corresponding indoles, upon microwave heating (see Supporting Information Scheme SI-1 for details). Dicyano derivatives 29 and 32-34 having R2 = Ph and R4 = CH(CN)2 were obtained by addition of lithium indolyl-cuprate(I) to benzylidene-malononitriles (Supporting Information, Scheme SI-4). Compounds having R2 = Ph and R4 = CH2CN were obtained either via Fischer indole syntheses (28, 30), or via InBr3-promoted addition of phenyl-oxirane to an indole, followed by one-carbon homologation (31). A similar strategy, followed by one-carbon homologation using TMSCF3 and Barton-McCombie dehydroxylation was used to access the CF3-indoles 37 and 39 (Supporting Information, Scheme SI-5). The cyclopropyl CF3-indole 40 was also prepared using a multi-step process involving TMS-CF3 homologation and Barton-McCombie dehydroxylation (Supporting Information, Scheme SI-7). However, the most versatile method for efficiently preparing a wide range of CF3-indoles is shown in Scheme 1. The starting indole was iodinated at the 3-position with NIS to be then subjected to Sonogashira coupling to give 79. Terminal alkyne trifluoromethylation with in-situ generated "Cu-

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 37

CF3" under aerobic conditions gave CF3-alkynes 80,27 which were transformed into iodo-alkenes 81 with lithium iodide under mild acidic aonditions,28 which were then submitted to Stille coupling with the corresponding aryl stannanes leading to N-MOM-protected indoles 82. The latter were first deprotected with HCl and then reduced with Et3SiH in the presence of TFA to afford the target CF3-indoles 1, 41-50 and 57. CF3 R3

R1

a)-c)

N MOM

d)

R3 N 79 MOM

R1

R3 N 80 MOM

R1

e) R2

CF3

R2 g)

R1

I

CF3 f)

R3

R1

N H 1, 41-50, 57

82

R3 N MOM

R1 81

CF3 R3 N MOM

Scheme 1. Synthesis of CF3-PAMs via Stille coupling/ionic hydride reduction strategy.

Reagents and conditions: a) NIS, 0 °C, CH2Cl2; b) TMSCCH, PdCl2(PPh3)2, CuI, ET3N, 23 °C, DMF; c) TBAF, 0 °C, THF; d) TMSCF3, CuI, TMEDA, K2CO3, air, 23 °C, DMF; e) LiI, HOAc, CH2Cl2, 0 – 23 °C; f) R2SnBu3, PdCl2(MeCN)2, P(oTol)3, CuI, 55 °C, DMF; g) 6 M HCl, 60 °C, THF then Et3SiH, TFA, 0 – 23 °C, CH2Cl2.

Further late-stage functional groups manipulations, shown in Scheme 2, expanded the library of CF3indole derivatives to include compounds 52-56 and 83-89.

CF3

S

CF3

S a)

Ph

O O

50

Ph

HO

N H

52

N H

b)

S

S CF3 Ph

O O

83

N SEM

c)

O N

Ph O

S

S CF3

N 84 SEM

CF3

d)

Ph O

N 85 SEM

HO

S CF3

e) N

Ph N 86 SEM

f)

S CF3

O

Ph N H

N 87 SEM

i)

H 2N

Ph 54

N 53 H

Scheme 2. Late-stage functional group interconversions strategy.

ACS Paragon Plus Environment

S

O N H

Ph N 89 SEM i)

S

S

CF3

CF3

Ph N H

CF3 O

N 88 SEM

CF3

O

h)

i)

S CF3

O

Ph

i)

S

N H

S CF3

g)

Ph H 2N

N 55 H

O

S

O N H

Ph N 56 H

Page 7 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Reagents and conditions: a) Super-Hydride®, THF; b) SEMCl; c) HNMe(OMe)-HCl, iPrMgCl, THF, 0 oC; d) MeLi, THF, -78 oC; e) NH2OH-HCl; f) cyanuric chloride, DMF, 23 oC; g) N2H4 hydrate, EtOH, 100 oC; h) MeSO2Cl, Et3N, 23 oC; i) HF -TFA. aq

All the compounds above were obtained in racemic form, but pure enantiomers could be obtained for a number of derivatives, such as 1, 8, 26 and 50, using preparative HPLC separation on chiral column (see Supporting Information for details). Compound (+)-1 was crystallized from EtOAc/Hexane solution and its structure was determined by X-ray crystallography,29 which showed that the stereogenic centres C9 and C31 had (S)-configurations (Figure 3). Note, as a consequence of the change of priority of the substituents, (R)-()-1 and (S)-()GAT229 have different descriptors but share the same stereocentre’s topicity.

CF3

S

NC

N H (S)-(+)-1 (ABD1235) (S)-(+)-1 (ABD1235)

Figure 3. X-ray crystal structure of (S)-(+)-1.

In vitro pharmacology. Modification of the ZCZ011 structure indicated that either a 6-Me (5), 6-Cl (6), 6-CF3 (7) or 6-CN (8) substitution led to moderately potent and efficacious CB1 PAMs, as determined by enhancement of β-arrestin recruitment, induced by the CB1 agonist CP55,940 at its EC20 in the PathHunter® β-Arrestin assay (Eurofins DiscoverX), by a variation of the methods described previously.7 The results are shown in Table 1. Substitutions on the 5-position of the indole gave compounds with poor potency [5-Me (2), 5-CN (3), 5-NO2 (4)]. However, results from the same assay demonstrated that, alongside their PAM activity, most of these compounds were also allosteric agonists, giving receptor activation in the absence of the orthosteric ligand. This compares with previous reports5 that the enantiomers of AZ-4 (GAT228 and 229) show considerable differences in their properties, with the (+) enantiomer being an allosteric agonist and the () being a pure PAM.

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Separation of the enantiomers of (±)-8 (6-CN) confirmed that ()-8 was the more potent PAM but, contrary to GAT228/229,5 did not indicate that the (+)-8 was an agonist: both ()- and (+)-8 showed enhancement of CP55,940-induced β-arrestin recruitment (Emax = 117% and 51% for the () and (+) respectively, p10 µM) making them both effectively pure PAMs (Table 1). Replacement of the 2-phenyl group (R3) with different aryl or cycloalkyl groups as in 9-25 was not tolerated. Other researchers30 have noted that, in the absence of any (R1) substitutions on the indole, 4-fluoro, 4-chloro or 3-chloro-4fluoro substitution on the 2-phenyl (R2) appeared to be beneficial for potency in a cAMP assay; however, no examples featuring both an R1 and R2 substitution were shown.

In addition to the potential toxicological concerns, the methylene group adjacent to the nitro group was expected to be acidic and potentially a metabolic liability, contributing to the short half-life. ZCZ011 (5) was shown to be highly unstable in the rat and mouse liver microsomal assays (t½ < 10 min), with moderate stability in human liver microsomes (t½ = 41 min) (Cyprotex Ltd). No significant increase in stability was seen with the 6-Cl (6) and 6-CF3 (7) derivatives or with the placement of an electronwithdrawing 4-Cl on the 2-phenyl group (16) (see Table 1). We therefore concluded that further development of the nitro-bearing compounds was unlikely to lead to drug-like compounds (suitable either for preclinical development or as tool compounds for in vivo target validation studies), and set about developing the synthetic methods required to permit replacement of the nitro group with a range of bioisosteres. Modelling software (FieldStere and FieldAlign, products available from Cresset Biomolecular at the time) was used to suggest moieties that might mimic the electronic properties of the nitro group, with the primary feature modelled being that of a hydrogen bond acceptor. Many structures were identified, but synthetic methods were generally lacking. Various 5-membered heterocycles were

ACS Paragon Plus Environment

Page 8 of 37

Page 9 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

investigated (Fig. 4); however, none of these were active and the syntheses were not straightforward. Thus, this line of investigation was terminated because of unproductivity. N N N

N N N N

N N N N

N H

N H

N N

N H

N H

Figure 4. Inactive heterocyclic analogues of ZCZ011 investigated in this work (see Supporting Information: Het-1 – Het-4).

The cyano group is sometimes proposed as being a bioisostere for a non-aromatic nitro group;31 although perhaps the most widely known example of this is in fact the use of nitro as a bioisostere of a cyano group, as seen in the development of ranitidine from cimetidine.32 This bioisosterism was affirmed by the (albeit modest) potency of compounds 30 and 32, which featured replacement of the CH2NO2 with either a CH2CN or a CH(CN)2 moiety. Compound 31 showed trends for somewhat reduced potency and efficacy as compared to its nearest nitro analogue (not significant), and showed similarlypoor metabolic stability in rat liver microsomes, possibly again related to the acidity of the proton adjacent to the CN group(s). Studies aimed at replacement of the stereogenic carbon atom with an amine led to the preparation of a range of compounds bearing a -COCF3 substituent, some of which (such as compounds 58 and 59, Fig. 5) showed a modest degree of enhancement (~50% at 1 µM). O N

N H

O CF3

58

N

N H

CF3

59

Figure 5. Trifluoroacetamide derivatives tested in this work.

It was assumed that these compounds would be highly susceptible to hydrolysis and they were not regarded as development candidates or even appropriate for use as tool compounds. However, these

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

results suggested that the –CF3 group might be a suitable bioisostere. Consequently, we set about developing synthetic methodologies to allow access to target compounds.

ACS Paragon Plus Environment

Page 10 of 37

Page 11 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Journal of Medicinal Chemistry

R2

R4 R3

R1 N H aPAM

Comp

R1

R2

R3

R4

2 3 4 5 6 7 (±)-8 (+)-8 (-)-8 9 10 11 12 13 14 15 16 17 18 19 20 21

5-Me 5-CN 5-NO2 6-Me 6-Cl 6-CF3 6-CN 6-CN 6-CN 6-Cl 6-Cl 6-Cl 6-Cl 6-Cl 6-Cl 6-Cl 6-Me 6-Cl 6-Cl 6-Me 6-Cl 6-Cl

2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl

Ph Ph Ph Ph Ph Ph Ph Ph Ph 4-OMe-C6H44-NMe2-C6H44-CN-C6H44-CO2Me-C6H44-SO2Me-C6H44-COOH-C6H44-Ac-C6H44-Cl-C6H44-Cl-C6H44-F-C6H44-F-C6H43-OMe-C6H43-CN-C6H4-

CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 (+)CH2NO2 (-)CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2NO2

Activity: >50 % enhancement * at 1000 nM ** at 500 nM *** at 200 nM * NA NA ** ** ** * * * NA NA NA NA NA NA NA NA NA NA NA NA NA

EC50

fT

Emax

1/2 (min)

bPAM

cAgo

dPAM

dAgo

HLM

RLM

MLM

283±140 462±363 284±211 863±211 664±215 -

777±363 1530±1128 974±780 >10000 >10000 -

106±18 154±10 163±17 51±18## 117±19## -

55±24 82±48 106±47 9±4# 21±6# -

49.4 41.2±3.1 32.7 30.7 24.2 -

8.2 6.4±1.2 13.8 15.6 9.8 -

-

-

10.4±2.1 -

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

22 23 24 25 (±)-26 (+)-26 (-)-26 27 28 29 30 31 32 33 34 35 36 37 38 39

6-Me 6-Me 6-Me 5-CN -

40

6-Me

41 42 43 44 45 46 47 48 49 (±)-50 (+)-50 (-)-50 51

6-Me 6-Me 6-Cl 6-Cl 6-Cl 6-CF3 6-CF3 5-Me 5-Me 6-CO2Me 6-CO2Me 6-CO2Me 6-CO2H

6-Me 6-Cl 6-Me 6-Me 6-Cl 6-Me 6-Me

2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl Phenyl 2-Thienyl 2-Thienyl Phenyl Phenyl Phenyl Phenyl Phenyl Cycloprop -yl methyl 2-Thienyl 2-Furyl 2-Furyl 2-Thienyl Phenyl Phenyl 2-Thienyl Phenyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl

Page 12 of 37

Cyclopropyl Cyclopentyl Cyclohexyl Cyclohexyl Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph

CH2NO2 CH2NO2 CH2NO2 CH2NO2 CH2CO2H CH2CO2H CH2CO2H CH2OH CH2CN CH(CN)2 CH2CN CH2CN CH(CN)2 CH(CN)2 CH(CN)2 COCF3 CHOHCF3 CH2CF3 COCF3 CH2CF3

* * NA NA NA NA NA NA NA NA * * * * NA NA NA * ** **

1113±401 -

2354±601 -

-

-

100±36 -

24±7 -

21.4 62.7 42.7 26.2 45.7 43.2 39.3 31.2 29.6 51.1 39.4 85.2 267 87.2

3.8 14.5 14.8 18.7 22.5 12.4 4.9 13.5 41.4 18.2 40.7 24.8 28.2

-

Ph

CH2CF3

*

-

-

-

-

74.4

19.8

-

Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph

CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3

** * ** ***

322±252 160±147‡‡ 113±62‡ -

801±554 682±211‡ 416±167‡ -

203±54 88±28## 156±41## -

95±15 26±5## 51±11## -

108 59.1 63.9 157 98.6 49.7 -

17.5 10.3 46.3 54.3 78.0 36.7 -

-

*** ** NA NA *** *** *** NA

ACS Paragon Plus Environment

Page 13 of 37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Journal of Medicinal Chemistry

52 53 (±)-1 (S)-(+)-1 (R)-(-)-1 54 55 56 57

6-CH2OH 6-NHAc 6-CN 6-CN 6-CN 6-Ac 6-NH2 6-NHSO2Me 6-CF3

2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl 2-Thienyl

Ph Ph Ph Ph Ph Ph Ph Ph Cyclopentyl

CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3 CH2CF3

NA NA *** ** ** *** NA NA **

177±66 342±142‡ 363±169‡‡ 158±60 301±66

647±270 1206±345‡ 828±310‡‡ 741±362 755±237

100±25 79±32# 125±29# 121±23 162±47

30±14 17±4 25±6 39±9 68±14

72.3±30.5 51.2 47.1

34.9±7.4 12.2 44.4

26.5±1.5 -

Table 1: Enhancement of β-arrestin recruitment induced by the CB1 agonist CP55,940 and metabolic stability data from human, rat and mouse liver microsomal stability studies (HLM, RLM and MLM). aPilot data were generated as described previously7 and represents data from a single experiment. bSelected compounds were evaluated by Eurofins DiscoverX in three independent experiments, using the PathHunter® β-Arrestin assay. Positive allosteric modulation was assessed using EC20 CP55,940. The EC50 value represents the compound concentration at which 50% of the maximal achievable EC20 CP55,940induced β-arrestin recruitment is seen (± standard deviation) for each specific compound. cAgonism was evaluated in the same assay and EC50 and Emax values were generated as described for positive allosteric modulation, using compound alone in the absence of agonist. dThe Emax for each compound is the maximum EC20 CP55,940-induced β-arrestin recruitment relative to the maximum achievable with CP55,940. eHuman, rat and mouse liver microsomal stability studies were conducted Cyprotex Ltd. Half-life values represent data from a single experiment (5 timepoints), except where ± stdev is given (three independent experiments). ## P