Novel 1-amidino-4-phenylpiperazines as potent agonists at human taar1 receptor: Rational design, synthesis, biological evaluation and molecular docking studies

pharmaceuticals

Article

Novel 1-Amidino-4-Phenylpiperazines as Potent

Agonists at Human TAAR1 Receptor: Rational

Design, Synthesis, Biological Evaluation and

Molecular Docking Studies

Valeria Francesconi ¹, Elena Cichero ^1,*, Evgeny V. Kanov ², Erik Laurini ³, Sabrina Pricl ^3,4

,

Raul R. Gainetdinov ^2,5and Michele Tonelli ^1,

*

1

Dipartimento di Farmacia, Università degli Studi di Genova, Viale Benedetto XV, 3, 16132 Genova, Italy;

francesconi.phd@difar.unige.it

Institute of Translational Biomedicine, St. Petersburg State University, 199034 St. Petersburg, Russia;

vonkanoﬀ@gmail.com (E.V.K.); gainetdinov.raul@gmail.com (R.R.G.)

Molecular Biology and Nanotechnology Laboratory (MolBNL@UniTS), Department of Engineering and

Architecture, University of Trieste, Piazzale Europa 1, 34127 Trieste, Italy; erik.laurini@dia.units.it (E.L.);

sabrina.pricl@dia.units.it (S.P.)

Department of General Biophysics, Faculty of Biology and Environmental Protection, University of Lodz,

90-236 Lodz, Poland

2

3

4

5

St. Petersburg University Hospital, St. Petersburg State University, 199034 St. Petersburg, Russia

Correspondence: cichero@difar.unige.it (E.C.); tonelli@difar.unige.it (M.T.)

*

ꢀꢁꢂꢀꢃꢄꢅꢆꢇ

Received: 17 October 2020; Accepted: 12 November 2020; Published: 14 November 2020

ꢀꢁꢂꢃꢄꢅꢆ

Abstract: Targeting trace amine-associated receptor 1 (TAAR1) receptor continues to offer an intriguing

opportunity to develop innovative therapies in different pharmacological settings. Pursuing our

endeavors in the search for effective and safe human TAAR1 (hTAAR1) ligands, we synthesized a

new series of 1-amidino-4-phenylpiperazine derivatives (1–16) based on the application of a combined

pharmacophore model/scaffold simplification strategy for an in-house series of biguanide-based TAAR1

agonists. Most of the novel compounds proved to be more effective than their prototypes, showing

nanomolar EC₅₀values in functional activity at hTAAR1 and low general cytotoxicity (CC₅₀> 80 µM)

when tested on the Vero-76 cell line. In this new series, the main determinant for TAAR1 agonism

ability appears to result from the appropriate combination between the steric size and position of

the substituents on the phenyl ring rather than from their different electronic nature, since both

electron-withdrawing and electron donor groups are permitted. In particular, the ortho-substitution

seems to impose a more appropriate spatial geometry to the molecule that entails an enhanced TAAR1

potency profile, as experienced, in the following order, by compounds 15 (2,3-diCl, EC₅₀= 20 nM),

2

(2-CH_3,EC₅₀= 30 nM), (2-OCH_3,EC₅₀= 93 nM) and (2-Cl, EC₅₀= 160 nM). Apart from the interest

6

3

in them as valuable leads for the development of promising hTAAR1 agonists, these simple small

molecules have further allowed us to identify the minimal structural requirements for producing an

efficient hTAAR1 targeting ability.

Keywords:traceamine-associatedreceptor1(TAAR1);humanTAAR1agonists;1-amidino-4-phenylpiperazines;

pharmacophore model; docking studies

1. Introduction

Trace amine-associated receptor 1 (TAAR1) is increasingly being recognized as a druggable target

in the treatment of several diseases, particularly for Central Nervous System (CNS) disorders [1,2].

Pharmaceuticals 2020, 13, 391; doi:10.3390 /ph13110391

ww w .mdpi.com /journal /pharmaceuticals

Pharmaceuticals 2020, 13, 391

2 of 26

TAAR1 is expressed in diﬀerent regions of the brain and several peripheral tissues (e.g., pancreas,

stomach, intestines, testes, and leukocytes) [3].

TAAR1 was identified as responsive to a class of endogenous ligands, called trace amines (TAs) [

3 ].

In particular, TAs such as tyramine (TYR), -phenylethylamine ( -PEA) and 3-iodothyronamine (T₁AM)

β

were identified as the most likely physiological, high-affinity TAAR1 agonists with potency in the

nanomolar range; however, both ligands and receptor-specific functions still deserve further investigations

to be clarified [4 ,5 ]. TAAR1 can also recognize classical monoamine neurotransmitters as endogenous

ligands, which are endowed with a micromolar degree of efficacy in stimulating TAAR1 [3 ,5

]. TAs and

classical biogenic amine pathways significantly overlap, even if TAs are expressed at markedly lower

levels in vivo.

Dysregulation of TAs is also considered as one of the contributing factors to the etiology of

many diseases, including not only schizophrenia, depression, attention deﬁcit hyperactivity disorder

(ADHD), but also sleep disorders and metabolic syndrome [2,6,7]. This scenario is currently stimulating

an increasing interest in speciﬁcally investigating the TAAR1 role in CNS, where receptor activation

is associated with precise control of monoaminergic circuits. TAAR1 is able to regulate dopamine

(DA) [8,9], 5-hydroxytriptamine (5-HT or serotonin) [10], and glutamate [11] neurotransmission,

decreasing the basal ﬁring rates and negatively modulating receptor sensitivity. In this context,

the initial studies revolved around drug repurposing strategies, drawing inspiration from the T₁AM

scaﬀold and known dopaminergic, adrenergic and serotonergic ligands. Diﬀerent structurally related

synthetic compounds and psychostimulant drugs have been screened and developed as potent TAAR1

agonists [12]. However, due to the structural similarity with their prototypes, most of these small

molecules continued to be aﬀected by an impaired selectivity proﬁle towards the TAAR1 receptor with

respect to other, highly related G-protein-coupled receptors (GPCRs) (monoaminergic systems) [13].

By chemical manipulation of T₁AM, Chiellini et al. developed a class of halogen-free diphenylmethane

derivatives, which displayed both in vitro and in vivo eﬃcacy [14,15]. Later analogues showed a

potency similar or even superior to that of their prototypes as TAAR1 agonists; however, these

molecules shared with T₁AM some non-TAAR1-mediated functional eﬀects (e.g., stimulation of hepatic

gluconeogenesis), which require further investigation for an adequate evaluation of the selectivity

issue [16,17].

Most eﬀorts in the TAAR1 ﬁeld were undertaken by the Hoﬀmann-La Roche pharmaceutical

company, whose research focused on an iterative series of structural modiﬁcations of adrenergic

medications; this led to the discovery of promising compounds (the so-called “RO family”), including

the amino-oxazoline α_2A-adrenergic receptor agonist S18616 [18]. A screening procedure involving

over 100 target proteins allowed the assessment of high TAAR1 selectivity for these RO compounds,

so they were widely used in functional assays to study the eﬀects of TAAR1 stimulation. As a result,

several potent full (e.g., RO5256390) [19] and partial TAAR1 agonists (e.g., RO5203648 and RO5263397,

Figure 1) were obtained [20,21], which also showed an improved eﬃcacy at hTAAR1.

The hTAAR1 agonists RO5256390 and RO5263397 also exhibited an addictive and synergistic effect

in combination therapy with the atypical antipsychotic drug olanzapine (Zyprexa) [22 ], suggesting,

for TAAR1 agonists, a potential antipsychotic-like activity along with a reduced risk of metabolic syndrome

induced by the marketed drug. Hofmann-La Roche recently patented a 5-ethyl-4-methylpyrazole-

3-carboxamide derivative based on the (2S)-(4-aminophenyl)morpholino scaﬀold as a ﬁrst-in-class

hTAAR1 partial agonist (EC₅₀= 58.5 nM, hTAAR1 eﬃcacy 42%, Figure 1) for the treatment or prevention

of psychiatric disorders [23 ]. Moreover, the selective TAAR1 agonists RO5166017 and RO5256390 were

also proposed for the treatment of diabetes and obesity [24 –26 ], being able to promote glucose-dependent

insulin secretion in -cells lines and human islets. In 2019, major progress in the quest for innovative

β

antipsychotic medications was achieved by Sunovion Pharmaceuticals with the identiﬁcation of the

molecule SEP-363856 as a mixed TAAR1 and 5-HT1A agonist [27] (Figure 1); this compound is currently

showing encouraging positive results in phase 2 and phase 3 clinical trials for the assessment of

Pharmaceuticals 2020, 13, 391

3 of 26

its eﬃcacy and safety in patients with schizophrenia (ClinicalTrials.gov Identiﬁer: NCT02969382,

NCT04109950, NCT04325737) [28].

Figure 1. Chemical structure and functional activity of the most promising human trace amine-associated

receptor 1 (hTAAR1) agonists.

From the analysis of the diﬀerent chemotypes thus far reported as TAAR1 agonists [1,3], a basic

core and an aromatic/heteroaromatic moiety were identiﬁed as mandatory pharmacophore features of

a TAAR1 ligand. Moreover, for a successful TAAR1 activity, these two required units need to be linked

through a spacer of variable nature and length; this, in turn, modulates the overall molecular ﬂexibility

and, ultimately, the related TAAR1 binding ability of the relevant compounds.

In this scenario, we started the study of new TAAR1 ligands, exploring phenyl or benzyl biguanides

(SET1, Figure 2) as novel chemotypes [29] that share a selective murine and human TAAR1 agonism

behavior with respect to murine TAAR5. In particular, the SET1 derivatives were endowed with higher

speciﬁcity towards the mouse receptor (nanomolar range) with respect to the human orthologue (low

micromolar range). Molecular docking studies at the m/hTAAR1 receptors enlightened the relevance of

the ligand’s basic core in forming a key salt bridge with a conserved m/hTAAR1 D3.32 aspartic acid,

and of the aromatic moiety in engaging π–π stacking and van der Waals contacts with a number of

recurrent aromatic residues characterizing the receptor cavity [1,29,30].

In this context, we recently reported the development of two quantitative structure–activity

relationship (QSAR) models exploring the agonism ability oﬀered by diﬀerent chemotypes towards

murine and human TAAR1, including the aforementioned SET1 derivatives and the most potent

agonists disclosed by Roche [31] (Supplementary data S1). The results allowed us to outline some species

speciﬁcity preferences and to derive useful information for the synthesis of novel biguanide-based

compounds (SET2, Figure 2) exhibiting selective agonism towards the two orthologues.

In particular, in order to evaluate the proper distance between the two moieties, we designed

these new arylbiguanide analogues based on the piperazine ring as a bifunctional and rigid spacer

tethering the aromatic core with the biguanide moiety. These SET2 compounds also showed agonistic

activity at TAAR1, mostly with a mouse TAAR1 (mTAAR1) species-selective proﬁle, with the exception

of 8a, which bound only the human orthologue, but with poor aﬃnity (Figure 2 ) [31]. From a careful

analysis of the two SET1 and SET2 biguanide-based series, the best substitution pattern consisted of

lipophilic groups on the aromatic ring (Cl, F, Br, CF₃, CH₃), with the electron-withdrawing chlorine

atom representing the best performing substituent; on the contrary, the presence of polar groups

(OCH₃) on the phenyl ring resulted in negative eﬀects, sometimes abolishing the activity. Interestingly,

Pharmaceuticals 2020, 13, 391

4 of 26

a comparable trend was also reported for the most notable TAAR1 targeting chemotypes thus far

developed [1].

Figure 2. Chemical structure of previously studied biguanide-based derivatives as m/hTAAR1 agonists

.

2. Results and Discussion

2.1. Design of hTAAR1 Agonists

In this work, we proceeded with the rational design of a new series of novel 1-amidino-4-

phenylpiperazine derivatives (1–16) (Figure 3 ) based on information derived from the development of a

preliminary pharmacophore model (PM), built by taking into account the most potent oxazolines thus

far described (pEC₅₀values > 7.00 M) among those reported by Roche (see Table S1). This approach

was combined with a scaffold simplification strategy for an in-house series of biguanide-based TAAR1

agonists (SET2) to probe the hTAAR1 targeting ability of this new series of amidine-containing derivatives.

We simpliﬁed the biguanide moiety of arylpiperazino-derived biguanides (SET2), which was

replaced in favor of the simpler and ﬂexible amidino group, with a view to improving the TAAR1

agonism ability. The aromatic ring of the novel 1-amidino-4-arylpiperazino scaﬀold was properly

functionalized at diﬀerent positions (

electron-withdrawing substituents, but also including the polar and electron donor OCH₃group

), in order to conﬁrm the structure–activity relationship of the previously built SET1 and SET2

1–16, Figure 3) introducing the more eﬃcient lipophilic and

(6–8

of biguanide-based TAAR1 agonists. To investigate the more suitable contribution for targeting the

TAAR1 receptor, compounds 10 and 11 were decorated with electron-deﬁcient rings (pyridine-2-yl

and pyrimidin-2-yl, respectively) in place of their phenyl ring. These moieties were also included in

diﬀerent TAAR1 targeting scaﬀolds, being characterized by at least one electron-rich nitrogen, involved

in a potential H-bond, or salt bridge interactions after protonation [1].

The bulk of the evidence in the literature [3,5,7] reports a fold decrease in aﬃnity values and

functional activity by using mouse and rat TAAR1 (m/rTAAR1) receptors as surrogate models of human

orthologues (hTAAR1). In this regard, we limited the functional evaluation of the novel compounds only

at hTAAR1, in order to obtain informative data; meanwhile, cytotoxicity assays were performed against

Vero cells to explore the safety proﬁle of these novel hTAAR1-targeting compounds. The molecular

docking study explained the SAR of these compounds in relation to their binding mode to hTAAR1,

revealing the key interaction for further improvements.

Pharmaceuticals 2020, 13, 391

5 of 26

Figure 3. Chemical structures of the investigated hTAAR1 agonists.

These combined studies have led to the identiﬁcation of promising molecules worthy of further

studies to assess their TAAR1 selectivity proﬁle over monoaminergic GPCRs and to develop optimized

TAAR1 ligands.

2.2. Pharmacophore Modeling

Inourpreviouswork, wereportedthedevelopmentoftwoquantitativestructure–activityrelationship

(QSAR) models, exploring the agonism ability featured by different chemotypes toward murine

and human TAAR1, including the in-house biguanide-based analogues (SET1) as well as the potent

imidazolines and oxazolines published by Roche [31 ]. The relevant results allowed us to identify some

species specificity preferences and to derive useful information for the synthesis of the aforementioned

biguanides (SET2), which exhibited selective agonism towards the two orthologues.

Brieﬂy, we focused on those compound chemical descriptors more speciﬁcally in relation to the

human TAAR1 (model A), deciphering a limited number of key descriptors involved in efficient hTAAR1

targeting. Then, we assessed the role played by the same set of chosen descriptors in influencing agonistic

activity at the murine receptor. In this way, a further QSAR model (model B) related to the murine

orthologue agonism ability was built.

Interestingly, model A unveils that seven out of the eight selected descriptors belong to the 3D class,

mainly referring to area or volume measurements. These ﬁndings underscore the importance of the

compound spatial conformation in inﬂuencing hTAAR1 activation, featuring a balanced hydrophobicity

Pharmaceuticals 2020, 13, 391

6 of 26

and hydrophilicity proﬁle in tandem with a folded polar surface. Indeed, the unsubstituted 1-phenyl-4-

biguanylpiperazine 1a (hTAAR1 EC₅₀= 1380 nM, pEC₅₀= 5.86 M, Figure 2) was more potent

than 1-benzylbiguanide (BIG10) (hTAAR1 EC₅₀> 10,000 nM). Smaller and conformationally locked

compounds such as 1-phelylbiguanide (BIG2), BIG4 and BIG8 of the SET1 family proved to be all

inactive as hTAAR1-targeting ligands, except for the 2-Cl substituted derivative BIG2 (Figure 2).

Conversely, the 3- and/or 4-substitution in this class of compounds led to more interesting derivatives

when combined with the benzyl-based biguanides of SET1. On the whole, in SET1, these two structural

architectures were mandatory to guarantee a proper conformer positioning within the receptor crevice.

Herein, we proceeded with the development of a preliminary pharmacophore model considering

the most potent oxazolines reported by Roche (pEC₅₀values > 7.00 M, Table S1), with the aim of gaining

useful information about the most effective key features and reciprocal distances required in tailoring

promising hTAAR1 ligands. This study was also conceived to provide an interesting computational tool

to evaluate and guide the effectiveness of the simplification strategy, prior to the chemical synthesis and

biological evaluation of new TAAR1 ligands.

The derived pharmacophore model was generated using the pharmacophore consensus module

integrated into the MOE software. The program is based on the identiﬁcation and classiﬁcation of

the most commonly shared recurrent pharmacophore features within the proposed set of molecules.

Any pharmacophore group is classiﬁed by an identiﬁcation code associated with the program (ID),

the percentage by which this feature appears among the molecules considered (SCORE), by a radius

that exempliﬁes the maximum space within which this functional group can be placed within the

ligand (RADIUS), and by a symbol that represents its role in terms of interaction with the receptor

(EXPRESSION). Among these, aromatic rings and hydrophobic substituents at the ligands were

identiﬁed as Aro and Hyd features, while H-bonding groups were described as Don or Acc when

referring to donor or acceptor moieties included in the agonists.

The alignment of the oxazolines 4b, 5b, 9b–11b, 16b, 18b, 20b–23b, 25b, 29b, 33b–37b (Table S1,

pEC₅₀values > 7.00 M) is shown in Figure 4, taking the most potent TAAR1 agonist 20b (pEC₅₀= 8.05)

as the reference compound.

Figure 4. Alignment of the oxazolines included in the pharmacophore model generation (compounds

4b, 5b, 9b–11b, 16b, 18b, 20b–23b, 25b, 29b, 33b–37b, Table S1), based on the chemical structure,

as represented by the previous conformer analysis [31]. All oxazolines are shown as thin sticks, with the

following atom color code: C, gray; halogen, green; N, blue; O, red). The reference compound 20b is

depicted as bold sticks with the same atom color code except for C, here in yellow for better visualization.

Pharmaceuticals 2020, 13, 391

7 of 26

Based on the data obtained, the most important pharmacophore requirements (represented by at

least 80% of the molecules under examination) to design a hTAAR1 agonist include seven characteristic

groups, especially H-bonding features properly tethered to at least two (hetero)aromatic or hydrophobic

groups, as shown in Table 1.

Table 1. List of pharmacophore features shared by 80% of the oxazolines acting as hTAAR1 agonists,

along with their ID, score, radius and expression parameters. Pharmacophore features: Aro = aromatic;

Hyd = hydrophobic; Don = hydrogen bond donor; Acc = hydrogen bond acceptor.

ID

SCORE RADIUS

EXPRESSION

F1

F2

F3

F4

F5

F6

F7

89%

83%

94%

89%

100%

1.96

1.46

1.36

1.08

1.26

0.85

1.23

Aro|Hyd

Hyd

Don2

Acc2

Don2

Acc

Don

In particular, Figure 4 shows a bulky aromatic ring (namely F1 Aro|Hyd) properly connected

to a further hydrophobic core (F2:Hyd) positioning H-bonding features (such as F3:Don2, F4:Acc2,

F5:Don2) in proximity of the corresponding principal F6:Acc and F7:Don (shared by all the most potent

oxazolines). Interestingly, this information agrees with our previous QSARs, supporting a limited

number of positively charged atoms in TAAR1 ligands engaged in a salt bridge with a conserved

aspartic acid (D103) of the receptor.

The expected reciprocal distances between F1 and F7, shared by most of the oxazolines, reveal

useful information for the further development of novel ligands (Figure 5).

Figure 5. Pharmacophore features shared by most of the oxazolines acting as hTAAR1 agonists (Table S1).

The most potent derivative 20b is mapped onto the corresponding pharmacophore model (PM) features

and is depicted by sticks (colors as in Figure 4 ). Distances between the main groups are shown as green

lines and labelled (Å).

Indeed, for optimal PM mapping, the two hydrophobic rings should be folded towards each other,

as exempliﬁed by F1: Aro|Hyd and F2:Hyd, within a distance of 3.99 A. While F2:Hyd should be in the

proximity of the H-bonding groups (2.09 A and 3.42 A, respectively) from the main F6:Acc and F7:Don

moieties, the terminal aromatic ring F1:Aro|Hyd should be maintained at ideal distances of 4.40 A and

6.32 A from the aforementioned F6:Acc and F7: Don.

Pharmaceuticals 2020, 13, 391

8 of 26

On the other hand, the introduction of other H-bonding groups—as exempliﬁed by the F3:Don2

F4:Acc2, F5:Don2 features placed at 8.48 A, 5.35 A and 7.33 A from F1:Aro|Hyd—are well-tolerated,

representing the ﬁnal cut-oﬀ values for the rational design of new analogues.

Compound 20b (pEC₅₀= 8.05)—chosen as representative oxazoline—fulﬁlls these requirements,

through the primary amine group, the oxazoline ring and the terminal phenyl ring, which proved to

be properly folded thanks to the presence of the ethoxy chain (Figure 6).

Conversely, the less potent analogues 30b (pEC₅₀= 5.57) and 27b (pEC₅₀= 5.00), used as external

derivatives, were too bulky compared to 20b or were folded diﬀerently with respect to the previously

discussed pharmacophore model (Figure 6). As a consequence, they experienced lower potency values

as hTAAR1 agonists.

Figure 6. Pharmacophore features shared by most of the oxazolines acting as hTAAR1 agonists. The most

potent derivative, 20b, is depicted by bold sticks (colored as in Figure 4 ). The less potent analogues, 30b

(left) and 27b (right), are also shown for comparison (C atoms colored in white).

On this basis, we decided that it would be interesting to compare the previously identiﬁed

1-phenyl-4-biguanylpiperazine 1a of SET2 (Figure 2) and compound 1, as the prototype of the newly

developed ligands (1–16), with the structural information derived from the oxazoline analysis. As shown

in Figure 7 (left), the biguanide moiety of 1a was placed in the proximity of the amino-oxazoline

portion, showing additional H-bonding features to those described by the pharmacophore modeling,

while the phenylpiperazine group was too ﬂat compared with the phenoxy ethyl chain of 20b. This 1a

positioning quite closely resembles that previously described for oxazoline 30b. Accordingly, 1a

(pEC₅₀= 5.86) and 30b (pEC₅₀= 5.57) experienced comparable potency ability as hTAAR1 agonists.

Figure 7. Pharmacophore features shared by most of the oxazolines acting as hTAAR1 agonists (Table S1).

The most potent derivative 20b is depicted by bold sticks (colors as in Figure 4 ). The PM mapping of

the in-house developed hTAAR1 agonist 1a

(left) and the newly designed and synthesized prototype 1

(right) are also shown for comparison (C atoms colored in white and cyan, respectively).

Pharmaceuticals 2020, 13, 391

9 of 26

On the other hand, the amidine moiety of

1

highly mimics the polarity trend of the oxazoline ring

as well as of the primary amine group of 20b, while the piperazine spacer allowed the compound to be

folded in the proximity of the 20b phenoxy group (Figure 7, right).

Calculations of polar and hydrophobic properties at the molecular surface of

1 and 20b revealed

similar features in terms of H-bonding groups and lipophilic substituents, supporting the idea of

modifying the biguanide–piperazine substituent with the amidine-containing group (Figure 8).

Figure 8. Comparison of the Connolly molecular surface properties of 20b (left) and 1 (right) in terms

of hydrophobic, polar and H-bonding features. Agonists are shown as bold sticks (C atoms colored in

yellow and white, respectively).

2.3. Chemistry

Compounds

36 ], 36 ] and 13

1

[

32 ],

8

[

33 ], 14

[

34 ] were already described in the literature as free bases, while

2 [35 ],

4

[35

,

5

[

35 ] were reported as hemisulfates and 10

,

11 37 ] as benzoate salts. Since our

[

synthetic procedure allowed for the yield of the title compounds as pure monohydrochlorides, their

experimental properties have been reported herein.

An equimolar mixture of the proper 1-aryl(heteroaryl)piperazine monohydrochloride and

◦

cyanamide was fused at 220 C for 30 min, leading to 1-amidino-4-arylpiperazines of Scheme 1.

The reaction failed, starting from the 1-(pyridin-2-yl)piperazine; thus, compound 10 was prepared

by reﬂuxing an aqueous solution of the piperazine derivative with S-methylisothiourea for 2 h with

stirring (Scheme 2).

Scheme 1. Reagents and conditions: (a) fusion at 220 ^◦C, 30 min.

Scheme 2. Reagents and conditions: (a) H₂O, reﬂux 2h.

2.4. Biological Studies and SAR

Compounds 1–16 were evaluated for functional activity at hTAAR1 and for cytotoxicity against the

Vero 76 cell line (Table 2). The activity of the compounds was measured using human embryonic kidney

293 (HEK-293T) cells co-transfected with plasmids encoding hTAAR1 and a cAMP Bioluminescence

Pharmaceuticals 2020, 13, 391

10 of 26

Resonance Energy Transfer (BRET) biosensor, or empty vector as a control. TYR was used as a

positive control for agonism. Firstly, all the compounds were tested at 10 M either for agonistic or

antagonistic activity. Then, for the active compounds, a dose–response experiment was performed

using concentrations in the range from 10 nM to 10 M in order to calculate their corresponding

µ

EC₅₀values. The E_maxvalue for the functional activity data deﬁnes the degree of functional activity

compared to 100% for a full agonist TYR. The compounds sharing an E_max> 85% at hTAAR1 were

considered as full agonists.

Table 2. Functional activity at hTAAR1 receptor and cytotoxicity against Vero-76 of compounds 1–16.

hTAAR1

EC₅₀, nM

Vero-76 Cells

Compound

b

c

Activity ^a

CC₅₀, µM

1

2

3

4

5

6

7

8

9

71%

100%

90%

35%

inactive

64%

32%

inactive

80%

370

30

160

>2000

-

93

244

-

84 ± 5

80 ± 5

> 100

90 ± 3

95 ± 3

83 ± 5

87 ± 4

88 ± 2

>100

64

10

11

12

13

14

15

16

TYR

30%

>2000

74 ± 5

64 ± 2

78 ± 2

87 ± 5

94 ± 6

82 ± 5

89 ± 4

-

inactive

81%

-

20

71

66

49%

100%

a

b

As compared with 1 µM TYR (full agonism); Data are representative of four independent experiments and

c

are expressed as means (errors on EC₅₀are within 10%). Compound concentration (µM) required to reduce the

viability of Vero-76 cells by 50% as determined by the MTT method. The results are expressed as CC₅₀

three separate experiments performed in duplicate.

±

SEM of

The new truncated compounds were synthesized from the piperazino-based biguanide prototypes

of SET2 by replacing the more rigid biguanide chain with the amidino moiety. Most of them showed

a potent agonism activity at hTAAR1, reaching the nanomolar potency (EC₅₀= 20–370 nM) of the

endogenous ligands and of the above-reported RO compounds.

As observed for the previously developed biguanide series (SET1 and SET2), the lipophilic Cl and

CH₃groups on the aromatic ring were conﬁrmed as being valuable substitutions that were able to

enhance the activity at hTAAR1 (2–4, 15, 16); the OCH₃substituent also proved to be eﬀective (6, 7).

In this new series, the main determinant for the TAAR1 agonism ability seemed to be the result of an

appropriate combination between the steric hindrance and position of the substituent on the phenyl

ring rather than the diﬀerent nature of substituents, since both electron-withdrawing and electron

donor groups were tolerated. Regarding each series of isomers (

was more eﬃcient than the meta ones ( ), for which a lower level of activity was observed, while the

para-substitution provided a negative outcome ( ). Interestingly, the same comparable trend was

3–5 and 6–8), the ortho-position (3, 6)

4, 7

5, 8

shown by the disubstituted compounds, 15 and 16, where the 2,3-diCl substitution led to a 3,5-fold

increase in TAAR1 activity (EC₅₀= 20 nM) with respect to the 3,4-diCl one (EC₅₀= 71 nM). As a result,

the ortho-substitution seems to impose a more appropriate spatial geometry to the molecule that entails

the most relevant TAAR1 agonism proﬁle, as experienced, in the following order, by compounds 15

6 and 3.

, 2,

From the analysis of the steric size of the diﬀerent substituents on the phenyl ring, the presence

of the larger 2-CH₃and 2-OCH₃groups is seen to correlate with a larger gain in the potency proﬁle

(2 and 6, EC₅₀= 30 and 93 nM, respectively) than the corresponding smaller 2-Cl (3, EC₅₀= 160 nM)

Pharmaceuticals 2020, 13, 391

11 of 26

and 2-F (13, inactive) ones, with the only exception being 2-CN (12, inactive). The same trend was

also observed for the meta-substitution, as the activity decreased, passing from CF₃( , EC₅₀= 64 nM)

9

and OCH₃(7, EC₅₀= 244 nM) groups to the smaller Cl atom (4, EC₅₀> 2000 nM). Another permitted

substitution was represented by the 2-pyridine ring (10) in place of the phenyl ring, whose eﬃcacy

nevertheless decreased by some orders of magnitude in the low micromolar range (EC₅₀> 2000 nM).

Importantly, this class of compounds elicited additional interest for its general low toxicity

(CC₅₀> 70

as the ratio of CC₅₀to EC_50,Figure 9) ranged from 625 (

(15, 2, 9, 16 and 6), pointing to a very good safety proﬁle.

µ

M, Table 1) against Vero-76 cells; accordingly, the corresponding therapeutic index (deﬁned

3) to 4100 (15) for the most active compounds

Figure 9. Therapeutic index (CC₅₀/EC₅₀) values of the active 1-amidino-4-arylpiperazines.

On this basis, the molecular simpliﬁcation of the biguanide skeleton of piperazino-based biguanides

(SET2) has proven to be a valid strategy, which allowed for the identiﬁcation of the optimal features for

an eﬃcient TAAR1 agonism behavior with the amidino group as basic motif and a planar aromatic ring

as the lipophilic substituent, properly spaced by the bifunctional piperazine ring. Therefore, our results

provide the foundation to further investigate these 1-amidino-4-arylpiperazine derivatives in order

to assess their TAAR1 selectivity proﬁle and make them worthy of structural optimization towards

improved derivatives endowed with well-suited drug-like properties. Indeed, previous in vivo and

in vitro studies of amidino piperazines described the ability of some compounds (corresponding to

the present derivatives 1, 4 and 5) to produce a sympathomimetic eﬀect, causing a rise in the blood

pressure, potentiating the pressor eﬀect induced by tyramine and norepinephrine (NE). These eﬀects

were correlated to the inhibition of monoamine oxidase (MAO) activity and of the restoration of

NE in its storage site as a preliminary point, even if the authors did not exclude the possibility of

other unknown mechanisms [36]. Furthermore, the intraperitoneal administration of 100 mg/kg of

1-phenyl-4-guanylpiperazine (1) increased the concentration of NE by about 34.5% in the brain and

42.7% in the heart of rats [38]. These ﬁndings pointed out an appropriate bioavailability proﬁle for this

class of compounds, which were demonstrated to interfere with the adrenergic system, both in the

brain and in the periphery, a ﬁnding that was also observed for other classes of TAAR1 ligands thus

far reported.

2.5. Docking Studies

As the last step in this work, we performed molecular docking studies of the newly developed

amidine-containing derivatives exploiting our hTAAR1 model [29

most promising derivatives, and 15, and on -PEA as the reference compound. Additional

docking studies of the previously cited oxazolines 20b 27b and 30b allowed us to validate the whole

,31]. Speciﬁcally, we focused on the

1,

2,

6

β

,

Pharmaceuticals 2020, 13, 391

12 of 26

computational study and to better delineate the SAR proﬁle featured by the new series of TAAR1

agonists (see scoring functions in Table S2).

In particular, the docking mode of the endogenous TAAR1 agonist

β-PEA (pEC₅₀= 6.70 M)

revealed one ionic interaction between the protonated nitrogen atom and D103 while the aromatic ring

of the compound was stabilized at the receptor crevice by

with F185, F195, F267 and F268 (Figure 10).

π–π stacking and Van der Waals contacts

Figure 10. Docking mode of the potent oxazoline 20b (C atom; yellow) and of

β-phenylethylamine

(β-PEA) (C atom; green) at the hTAAR1 putative binding site (left). Ligands are shown as sticks. The most

relevant residues are labelled. A schematic perspective of the most relevant H-bonds experienced by the

two agonists is also shown (right).

Maintaining a ﬂexible chain connecting a basic core to a terminal aromatic ring, as experienced by

the potent agonist 20b (pEC₅₀= 8.05 M), allowed the compound to be properly folded within the GPCR

cavity, in order to display a further H-bond between the ethoxy alkyl chain and the F185 backbone

(Figure 10). On the other hand, the primary amine group linked to the oxazoline ring of 20b was

H-bonded to the D103 residue while the amino-oxazoline group and the phenyl ring were engaged

in polar contacts and cation–

revealed a high number of polar contacts within the protein binding site if compared to the endogenous

ligand -PEA, exhibiting higher potency values than this trace amine, featuring the necessary –cation

and H-bond interactions [29,31].

Contextually, the alkyl chain and the terminal phenyl ring are engaged in van der Waals contacts

stacking with the side chains of V184, I290 and F185, respectively. A diﬀerent point of view

π interaction with Y294 and H99, respectively. As a consequence, 20b

β

π

and π–π

for the discussed docking poses is reported in Figure S1.

On the whole, the related scoring value obtained for this protein–agonist complex is in agreement

with the potency observed for 20b (pEC₅₀= 8.05 M, S =

of this compound compared with the endogenous ligand

In fact, the docking mode of -PEA features the required key contact with D103 via the protonated

nitrogen atom, along with hydrophobic interactions and stacking between the ethyl chain and

−

99.1078) and supports the higher eﬀectiveness

β

-PEA (pEC₅₀= 6.70 M, S = 95.8775).

−

β

π–π

I290, and the phenyl ring with F195, F267 and F268, respectively. These ﬁndings are validated by the

docking positioning derived for the 20b rigid analogues 26b (pEC₅₀= 6.82 M), 27b (pEC₅₀= 5.00 M)

and 30b (pEC₅₀= 5.57 M), supporting a relevant role played by interacting with D103 and F185 to

stabilize the aﬀective agonist at the receptor binding site. Concerning the two enantiomers 26b (S)

Pharmaceuticals 2020, 13, 391

13 of 26

and 27b (R), only the ﬁrst one displayed promising potency values as a hTAAR1 agonist thanks to

polar contacts and H-bonds between the primary amine group and D103 and T100, S183, respectively

(Figure 11 ). This kind of positioning allowed the compound to exhibit

contacts with F185, F267 and V184, F85 and I290.

π–π stacking and Van der Waals

Figure 11. Docking positioning of the oxazolines 26b (C atom; green) and 27b (C atom; magenta) at

the hTAAR1 putative binding site (left). Ligands are shown as sticks. The most relevant residues are

labelled. A schematic perspective of the most relevant H-bonds experienced by the two agonists is also

shown (right).

The modest agonist 27b (R) experienced a shifted docking mode compared to 26b (S), moving

the oxazoline ring far from the D103 residue in order to detect only two H-bonds with T100 and F185

(Figure 11 ) and therefore motivating the lower potency trend of this derivative. However, 27b was

able to maintain a number of hydrophobic and π–π contacts with the surrounding residues F185, F267

and F268, supporting the idea of a main aromatic core tethered to proper H-bonding features.

We reasoned that that the limited ﬂexibility and dimensions of 27b, especially when combined

with a small hydrophobic substituent at the para position of the terminal phenyl ring, could impair the

capacity of the compound to properly occupy the receptor crevice with the expense of polar contacts

with D103, due to a consistent number of hydrophobic interactions with the GPCR cavity. Interestingly,

these results are also in accordance with the SAR developed within the amidine series, with 5, 8, 14 and

16 being inactive as hTAAR1 agonists. A bulkier substituent may be preferred to arrange and better

stabilize the ligand at the receptor binding site.

Accordingly, the oxazoline 30b (pEC₅₀= 5.57; S = 84.7643), bearing a biaryl substituent connected

−

to the oxazoline core, displayed several van der Waals contacts and aromatic stacking with I104, V184

and with F185, W264, F267, respectively, while the amino-oxazoline ring was H-bonded to T100, S183

and F185 (see Figure S2). In particular, this positioning allowed the compound to exhibit higher

potency than 27b, as conﬁrmed by the calculated scoring functions.

Concerning the newly developed amidine-containing derivatives, both the prototype 1 and the

most promising derivative 15 slightly shift their position within the protein cavity, bringing the amidine

moiety into the proximity of the receptor residues H99 and D103, thereby engaging their side chains in

H-bond contacts (Figures 12 and 13). Moreover, the folded piperazine, in tandem with the presence

of substituents at the phenyl ring, guides the aromatic core and the amidine moiety towards the

hydrophobic cavity delimited by I104, V184, F185, W264, I290.

Pharmaceuticals 2020, 13, 391

14 of 26

As shown in Figure 12, the unsubstituted compound

the aromatic core towards H99 and V184, thereby featuring van der Waals contacts and cation–

interactions, while the piperazine ring is involved in Van der Waals contacts with F267.

1

(pEC₅₀= 6.43 M, S =

−

85.7222) orients

π

Figure 12. Docking positioning of the amidine-containing derivatives 1 (C atom; yellow) and β-PEA (C

atom; green) as hTAAR1 agonists at the receptor crevice (left). Ligands are shown as sticks. The most

relevant residues are labelled. A schematic perspective of the most relevant H-bonds experienced by

the two agonists is also shown (right).

Moreover, the amidine moiety was H-bonded to D103 and I290, therefore being stabilized within

the receptor cavity. Notably, the discussed docking mode of compound was quite comparable with

that of -PEA, with the exception of stacking with F267 and F268, due to the rigid piperazine in

place of the ﬂexible ethyl chain (Figure 12). These ﬁndings are in agreement with the higher pEC₅₀

values of the endogenous trace amine -PEA (pEC₅₀= 6.70 M) in comparison with that of the amidine

1

β

π–π

β

derivative 1. A diﬀerent point of view for the discussed docking poses is reported in Figure S3.

Figure 13. Docking positioning of the amidine-containing derivatives 15 (C atom; magenta) and of the

oxazoline 30b (C atom; orange) as hTAAR1 agonists at the receptor crevice (left). Ligands are shown as

sticks. The most relevant residues are labelled. A schematic perspective of the most relevant H-bonds

experienced by the two agonists is also shown (right).

Pharmaceuticals 2020, 13, 391

15 of 26

Regarding the most potent analogue 15 (pEC₅₀= 7.70 M, S =

−

95.0332), the reported docking

mode revealed the eﬀective positioning of the agonist within the larger cavity of the hTAAR1 receptor,

including F185, S198, W264, F267 and F268, featuring stacking and Van der Waals contacts thanks

π–π

to the aryl moiety. This kind of docking mode allows the compound to better match the binding site

features, exhibiting a higher number of polar contacts with H99, D103 and S183 via its amidine moiety

(Figure 13).

This is in accordance with the higher eﬀectiveness of 15 (pEC₅₀= 7.70 M) compared with the

oxazoline 30b (pEC₅₀= 5.57), as shown in Figure 13. Indeed, the amidine moiety is a good bioisostere

of the amino-oxazoline ring while the 2,4-dichlorophenyl piperazine portion proved to be better

substituent than the biaryl moiety to eﬃciently bind the GPCR crevice. The positioning of compound

15 allowed the agonist to detect stronger polar and H-bond contacts with D103 and S183 than 30b.

A good similarity in terms of putative bioactive conformation can be noticed by comparing the

docking mode of compound 15 with that previously discussed for 20b (pEC₅₀= 8.05 M) and

(pEC₅₀= 6.70 M), as shown in Figure 14.

β-PEA

Figure 14. Docking positioning of potent oxazoline 20b (C atom; yellow), of the amidine-containing

derivative 15 (C atom; magenta) and of the endogenous agonist -PEA (C atom; green) as

hTAAR1-targeting compounds (left). Ligands are shown as sticks. The most relevant residues

β

are labelled. A schematic perspective of the most relevant H-bonds experienced by 15 and

also shown (right).

β-PEA is

While the 2,4-dichlorophenyl group of 15 highly mimicked the aromatic ring of

portion displayed additional H-bonds than the trace amine, supporting the higher potency of 15 in

comparison to -PEA. On the other hand, the amidine moiety guaranteed the required polar interactions

β-PEA, the amidine

β

with H99 and D103, as featured by 20b and additional contacts with S183, while the piperazine spacer

overlapped the ethoxyamine chain of the oxazoline, lacking H-bonds with F185. As a result, 15 experienced

quite comparable potency with respect to 20b.

Finally, the introduction of an electron donor substituent at the ortho position of the phenyl ring

led to the 2-methyl- or 2-methoxy-phenyl substituted analogues

agonist ability. Both compounds show a docking mode comparable to that of 15, with derivative

(pEC₅₀= 7.52 M, S = 94.6754) being more eﬀective than (pEC₅₀= 7.03 M, S = 94.5612). Indeed,

the 2-methylphenyl ring is better placed within the receptor cavity, mimicking the role played by the

2 and 6 exhibiting adequate hTAAR1

2

−

6

−

dichloro-substituted phenyl ring of 15. In any case, the amidine moiety of 2 and 6 properly displays

the key contacts with H99 and D103 (see Figure S4).

Pharmaceuticals 2020, 13, 391

16 of 26

2.6. Prediction of ADMET Properties

Nowadays, applying in silico methods for the computational prediction of descriptors related to the

pharmacokinetic (PK) and toxicity profile of novel molecules represents a useful tool that is accelerating the

lead compound discovery process [39 ]. A number of parameters involved in the absorption, distribution,

metabolism, excretion and toxicity properties (ADMET) of putative druglike derivatives can be managed

in silico, and we successfully applied this computational approach on our first series of TAAR1 agonists [29 ].

Thus, herein, we explored the putative ADMET proﬁle of the newly developed TAAR1 agonists

1–16 as well as of the endogenous ligand β-PEA and of the oxazoline derivative RO5263397, taken as

reference drug-like derivatives. In particular, the favorable pharmacokinetic proﬁle of RO5263397 has

been experimentally determined [18]. Brieﬂy, RO5263397 was characterized by an acceptable human

hepatocyte clearance and the inhibitory activity of the cytochrome isoforms, exhibiting in vitro safety

proﬁle performed by a (negative) AMES/MNT test and hERG IC₅₀evaluation.

In this work, for the aforementioned amidine-based derivatives, the logarithmic ratio of the

octanol–water partitioning coeﬃcient (cLogP), the ability to pass the blood–brain barrier (BBB) to the

extent of BBB permeation (LogBB), the rate of passive diﬀusion permeability (LogPS), human intestinal

absorption (HIA), the volume of distribution (Vd), the role played by plasmatic protein binding (%PPB),

and the ligand aﬃnity toward human serum albumin (LogKa ^HSA) were all taken into account in order

to determine the putative value of the oral bioavailability as a percentage (%F) (see Table 3).

Table 3. Calculated absorption, distribution, metabolism, excretion and toxicity properties (ADMET)

descriptors related to absorption and distribution properties. The most promising amidine-containing

derivative 15 has been highlighted in yellow, as well as the reference compounds

The other most interesting analogues, and 16, are shown in light cyan. The only one derivative

predicted as unable to pass the blood–brain barrier (BBB) is reported in gray.

β-PEA and RO5263397.

2, 6, 9

LogPS ^b

Vd (l/kg) ^d

LogKa ^HSA

Comp.

cLogP LogBB ^a

HIA (%) ^c

%PPB

%F (oral)

β-PEA

1.56

0.31

0.71

1.15

0.99

0.81

0.33

0.36

0.16

1.26

−0.31

−0.43

0.04

0.45

0.37

1.72

1.88

1.85

0.19

−0.08

−0.03

−0.04

−0.08

−0.13

−0.08

−0.09

−0.05

−0.13

−0.09

−0.04

−0.03

−0.01

0.13

−2.4

−3.6

−3.4

−3.2

−3.3

−3.7

−3.9

−3.1

−4.0

−4.1

−3.8

−3.5

−2.9

−2.8

−2.0

99

50

58

74

68

61

45

46

41

77

43

42

40

53

51

92

95

100

3.0

1.2

1.4

1.3

1.5

1.4

1.2

1.3

1.1

1.6

1.0

0.7

1.2

1.1

1.5

1.6

2.4

34.79

33.14

33.91

48.25

49.24

49.93

33.69

33.00

31.42

52.86

32.87

25.03

21.76

29.16

23.28

50.18

50.85

27.57

2.78

2.84

2.85

3.19

3.11

3.15

2.90

2.82

2.86

3.34

2.68

2.49

2.51

2.97

2.51

3.57

3.52

3.12

91.1

28.9

35.7

49.7

44.5

38.5

27.1

23.2

54.1

25.2

22.9

32.3

30.6

66.9

76.9

99.0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

0.19

0.35

RO5263397

a

Extent of brain penetration based on ratio of total drug concentrations in tissue and plasma at steady-state

b

conditions; Rate of brain penetration. The permeability–surface area product (PS) is derived from the kinetic

equation of capillary transport; ^chuman intestinal absorption (HIA) is expressed as percentage of the molecule able

to pass through the intestinal membrane; ^dprediction of volume of distribution (Vd) of the compound in the body.

The putative metabolism and toxicity proﬁles of any compound were determined on the basis of

a number of descriptors such as the ability to interact with the endocrine system and to inhibit the

hERG channel, as well as to act as cytochrome P450 3A4 and 2D6 inhibitors or substrates (see Table 4).

Pharmaceuticals 2020, 13, 391

17 of 26

Table 4. Calculated ADMET descriptors related to metabolism, excretion and toxicity properties. The most

promising amidine-containing derivative 15 has been highlighted in yellow, as well as the reference

compounds

light cyan.

β-PEA and RO5263397. The other most interesting analogues, 2, 6, 9 and 16, are shown in

hERG Inhibitor

Endocrine System Disruption ^a

CYP3A4 and CYP2D6

Comp.

Reliability

Index (R.I.)

LogRBA > −3

LogRBA > 0

(R.I. ≥ 0.3)

Inhibitor < 10 mM

(R.I. > 0.3)

Substrate

(R.I. ≥ 0.3)

Inhibitor

(R.I. ≥ 0.3)

β-PEA

0.06

0.07

0.08

0.11

0.15

0.10

0.09

0.15

0.43

0.07

0.05

0.07

0.12

0.14

0.21

0.27

0.06

0.40

0.34

0.35

0.29

0.22

0.56

0.35

0.33

0.30

0.25

0.27

0.24

0.30

0.36

No binder

0.02

0.01

0.02

0.01

0.02

0.01

>50%

33%

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

RO5263397

a

RBA: relative binding aﬃnity with respect to that of estradiol. Compounds showing LogRBA > 0 are classiﬁed as

3 are considered as non-binders. (RI: reliability index.

strong estrogen binders, while those showing LogRBA <

Borderline-allowed values for reliability parameter are ≥ 0.3; the most predictive fall in the range 0.50–1.0).

−

As shown in Table 3, all the newly synthesized compounds with the exception of compound 11

were predicted as being able to pass through the BBB thanks to organic cation transporters. Among them,

the promising compounds 15 and 16 were endowed with comparable lipophilicity values with respect

to the reference compounds β-PEA and RO5263397, turning in favorable values of human intestinal

absorption (HIA = 92–95%). While the reported prediction of volume of distribution (Vd) for 15

(Vd = 1.5 L/Kg; %PPB = 50.18%) and 16 (Vd = 1.6 L/Kg; %PPB = 50.85%) was noticeably lower than

that of the two references (Vd = 2.4–3.0 L/Kg; %PPB = 27.57–34.79%), the potential binding to the

plasmatic proteins fell in the allowed ranges (spanning from 40% to 80%), being higher than those of

β

-PEA and RO5263397. Accordingly, the predicted LogKa ^HASvalues were quite comparable (compare

15 and

16 (LogKa ^HSA= 3.52–3.57) to -PEA and RO5263397 (LogKa ^HSA= 2.78–3.12)). Regarding

, only the last one was endowed with a proper PK proﬁle based on the calculated cLogP ( cLogP =

1.26), HIA% (>70%), %PPB (>50%) and %F (>50%), as reported in Table 3.

,

β

2, 6

9

According to an overall analysis of the descriptors listed in Table 4, none of the compounds

proposed here should be involved in genotoxicity events such as binding with the endocrine system,

while only amidine 9 is predicted as a putative hERG inhibitor, being accompanied by a high reliability

index R.I = 0.56. Furthermore, none of them were proven to inhibit the cytochrome P450 3A4, as only

the reference agonist RO5263397 was the substrate for this enzyme. On the other hand, the endogenous

amine

β-PEA as well as the synthesized TAAR1 agonists 1–16 were predicted as substrates for CYP2D6

with high reliability indexes (RIs) (>50%).

Pharmaceuticals 2020, 13, 391

18 of 26

3. Materials and Methods

3.1. Chemistry

3.1.1. General Information

Chemicals, solvents and reagents (R-phenylpiperazines) used for the syntheses were purchased

from Sigma-Aldrich (Milan, Italy), and were used without any further puriﬁcation. Melting points

1

(uncorrected) were determined with a Büchi apparatus (Milan, Italy). H NMR spectra and ¹³C NMR

spectra were recorded on a Varian Gemini-200 instrument at 200 and 50 MHz, respectively; DMSO-d₆;

δ

in ppm rel. to Me₄Si as internal standard. J in Hz. Elemental analyses were performed on a Flash 2000

CHNS (Thermo Scientiﬁc, Milan Italy) instrument in the Microanalysis Laboratory of the Department

of Pharmacy, University of Genova. The results of elemental analyses indicated that the purity of all

compounds was ≥95%.

3.1.2. General Method for the Synthesis of 1-Amidino-4-Arylpiperazine Derivatives

A mixture of the proper phenylpiperazine hydrochloride (2.5 mmol) and cyanamide (2.5 mmol)

was fused at 220 ^◦C for 30 min. At r.t., the hard solid was rinsed with a mixture of EtOAc and Et₂O

and washed with the same solvents, aﬀording the ﬁnal products in the form of monohydrochlorides.

For some highly hygroscopic monohydrochloride salts, melting points have not been reported, since

they melted with decomposition in a wide range.

4-Phenylpiperazine-1-carboximidamide hydrochloride (

DMSO-d₆):

7.76 (s, NH₂⁺exchange with D₂O), 7.20 (pseudo s, 2ArH), 7.17–6.84 (m, 5H, 3ArH and

NH₂exchange with D₂O), 3.72–3.51 (m, 4H, 2CH₂piperazine), 3.21–3.08 (m, 4H, 2CH₂piperazine).

¹³C NMR (50 MHz, DMSO-d₆):

161.1, 149.7, 128.7 (2C), 122.6, 117.0 (2C), 54.3 (2C), 48.1 (2C). Anal.

Calcd. For C₁₁H₁₆N₄·HCl: % C, 54.88; H, 7.12; N, 23.27. Found: C, 54.63; H, 6.92; N, 23.68.

1

): Yield: 90%. M.p. 214–217 ^◦C. ¹H NMR (200 MHz,

δ

4-(o-Tolyl)piperazine-1-carboximidamide hydrochloride (

): Yield: 83%. M.p. 159–162 ^◦C. ¹H NMR (200 MHz,

DMSO-d₆):

7.80 (s, NH₂⁺exchange with D₂O), 7.43 (pseudo s, 1ArH), 7.23–6.86 (m, 5H, 3ArH and NH₂

exchange with D₂O), 3.62–3.56 (m, 4H, 2CH₂piperazine), 3.27–2.97 (m, 4H, 2CH₂piperazine), 2.41 (s,

CH₃). ¹³C NMR (50 MHz, DMSO-d₆):

162.3, 149.9, 131.1, 129.8, 126.7, 122.2, 117.9, 52.4 (2C), 47.1 (2C),

2

δ

19.6. Anal. Calcd for C₁₂H₁₈N₄·HCl: % C, 56.57; H, 7.52; N, 21.99. Found: C, 56.64; H, 7.80; N, 21.69.

1

4-(2-Chlorophenyl)piperazine-1-carboximidamide hydrochloride (

δ

3

): Yield: 67%. H NMR (200 MHz, DMSO-d₆):

+

7.79 (s, NH₂exchange with D₂O), 7.35–7.32 (m, 1ArH), 7.19–6.97 (m, 5H, 3ArH and NH₂exchange

with D₂O), 3.71–3.54 (m, 4H, 2CH₂piperazine), 3.17–2.91 (m, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz,

DMSO-d₆): 162.1, 151.1, 130.5, 127.3, 126.2, 123.0, 120.1, 51.7 (2C), 46.3 (2C). Anal. Calcd for C₁₁H₁₅ClN₄HCl:

% C, 48.01; H, 5.86; N, 20.36. Found: C, 48.11; H, 6.04; N, 20.29.

δ

·

1

4-(3-Chlorophenyl)piperazine-1-carboximidamide hydrochloride ( ): Yield: 45%. H NMR (200 MHz, DMSO-d₆):

7.61 (s, NH₂⁺exchange with D₂O), 7.42 (s, 1 ArH), 7.26–7.19 (m, 1ArH), 7.09–6.88 (m, 4H, 2ArH and

NH₂exchange with D₂O), 3.57 (pseudo s, 4H, 2CH₂piperazine), 3.21 (pseudo s, 4H, 2CH₂piperazine).

¹³C NMR (50 MHz, DMSO-d₆):

161.4, 151.9, 139.1, 130.7, 120.5, 117.8, 113.5, 51.2 (2C), 46.7 (2C). Anal.

4

δ

Calcd for C₁₁H₁₅ClN₄·HCl: % C, 48.01; H, 5.86; N, 20.36. Found: C, 48.14; H, 5.96; N, 20.04.

1

4-(4-Chlorophenyl)piperazine-1-carboximidamide hydrochloride (

δ

5

): Yield: 52%. H NMR (200 MHz, DMSO-d₆):

7.67 (s, NH₂⁺exchange with D₂O), 7.14–6.89 (m, 6H, 4ArH and NH₂exchange with D₂O), 3.58–3.49 (m,

4H, 2CH₂piperazine), 3.16–3.02 (m, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz, DMSO-d₆):

161.9, 150.6,

HCl: % C, 48.01; H, 5.86; N,

δ

129.1 (2C), 123.7, 116.5 (2C), 49.6 (2C), 45.7 (2C). Anal. Calcd for C₁₁H₁₅ClN₄

20.36. Found: C, 47.91 H, 5.96; N, 20.36.

·

1

4-(2-Methoxyphenyl)piperazine-1-carboximidamide hydrochloride (

δ

6

): Yield: 90%. H NMR (200 MHz, DMSO-d₆):

7.58 (s, NH₂⁺exchange with D₂O), 7.12–6.84 (m, 6H, 4ArH and NH₂exchange with D₂O), 3.72 (s, OCH₃),

3.47 (pseudo s, 4H, 2CH₂piperazine), 3.23 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz, DMSO-d₆):

Pharmaceuticals 2020, 13, 391

19 of 26

δ

162.7, 151.7, 142.5, 124.3, 122.6, 118.1, 113.7, 57.3, 50.7 (2C), 47.1 (2C). Anal. Calcd for C₁₂H₁₈N₄O

·HCl: %

C, 53.23; H, 7.07; N, 20.69. Found: C, 53.46; H, 7.02; N, 20.96.

1

4-(3-Methoxyphenyl)piperazine-1-carboximidamide hydrochloride (

7.64 (s, NH₂⁺exchange with D₂O), 7.18–7.11 (m, 1ArH), 6.89–6.41 (m, 5H, 3ArH and NH₂exchange with

D₂O), 3.74 (s, OCH₃), 3.53–3.49 (m, 4H, 2CH₂piperazine), 3.39–3.17 (m, 4H, 2CH₂piperazine). ¹³C NMR

(50 MHz, DMSO-d₆): 161.9, 159.8, 149.7, 131.2, 111.3, 107.6, 103.2, 56.9, 49.8 (2C), 47.3 (2C). Anal. Calcd for

7): Yield: 43%. H NMR (200 MHz, DMSO-d₆):

δ

C₁₂H₁₈N₄O·HCl: % C, 53.23; H, 7.07; N, 20.69. Found: C, 53.28; H, 7.22; N, 20.51.

1

4-(4-Methoxyphenyl)piperazine-1-carboximidamide hydrochloride (

δ

8

): Yield: 54%. H NMR (200 MHz, DMSO-d₆):

7.71 (s, NH₂⁺exchange with D₂O), 7.17–6.92 (m, 6H, 4ArH and NH₂exchange with D₂O), 3.78 (s, OCH₃),

3.61 (pseudo s, 4H, 2CH₂piperazine), 3.22 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz, DMSO-d₆):

162.4, 154.6, 143.8, 116.2 (2C), 114.8 (2C), 56.6, 50.1 (2C), 47.3 (2C). Anal. Calcd for C₁₂H₁₈N₄O HCl: % C,

δ

·

53.23; H, 7.07; N, 20.69. Found: C, 53.43; H, 7.16; N, 20.90.

4-[(3-Triﬂuoromethyl)phenyl]piperazine-1-carboximidamide hydrochloride (9

): Yield: 31%. M.p. 170–173 ^◦C.

¹H NMR (200 MHz, DMSO-d₆):

δ

7.76 (s, NH₂⁺exchange with D₂O), 7.58–7.01 (m, 6H, 4ArH and NH₂

exchange with D₂O), 3.58 (pseudo s, 4H, 2CH₂piperazine), 3.27 (pseudo s, 4H, 2CH₂piperazine). ¹³C

NMR (50 MHz, DMSO-d₆): 160.3, 157.6, 131.2 (q, J_C-F= 22.4 Hz), 128.3 (d, J_C-F= 22.35 Hz), 124.6 (d,

J_C-F= 271.25 Hz), 122.1 (d, J_C-F= 20.9 Hz), 121.3 (q, J_C-F= 37.5 Hz), 116.7 (q, J_C-F= 36.7 Hz), 50.1 (2C),

δ

44.6 (2C). Anal. Calcd for C₁₂H₁₅F₃N₄

N, 18.41.

·

HCl: % C, 46.68; H, 5.22; N, 18.15. Found: C, 46.79; H, 5.44;

1

4-(Pyrimidin-2-yl)piperazine-1-carboximidamide hydrochloride (11): Yield: 29%. H NMR (200 MHz, DMSO-d₆):

δ

7.97–7.36 (m, 4H, 2H pyrim. And NH₂⁺exchange with D₂O), 7.31–6.64 (m, 3H, 1H pyrim. And NH₂

exchange with D₂O), 3.71 (pseudo s, 4H, 2CH₂piperazine), 3.59 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR

(50 MHz, DMSO-d₆): 161.4, 159.1, 157.6 (2C), 112.1, 46.8 (2C), 44.3 (2C). Anal. Calcd for C₉H₁₄N₆HCl: %

C, 44.54; H, 6.23; N, 34.63. Found: C, 44.49; H, 6.34; N, 34.34.

δ

·

4-(2-Cyanophenyl)piperazine-1-carboximidamide hydrochloride (12): Yield: 61%. M.p. 236–238 ^◦C. ¹H NMR

(200 MHz, DMSO-d₆):

7.98–7.52 (m, 6H, 2ArH, NH₂and 7.77, s, NH₂⁺superimposed, exchange with

D₂O), 7.28–7.07 (m, 2ArH), 3.64 (pseudo s, 4H, 2CH₂piperazine), 3.23 (pseudo s, 4H, 2CH₂piperazine).

¹³C NMR (50 MHz, DMSO-d₆):

162.2, 154.5, 136.7, 130.6, 122.1, 118.6, 117.4, 106.2, 51.1 (2C), 46.3 (2C).

δ

Anal. Calcd for C₁₂H₁₅N₅·HCl: % C, 54.24; H, 6.07; N, 26.35. Found: C, 54.00; H, 6.24; N, 26.18.

4-(2-Fluorophenyl)piperazine-1-carboximidamide hydrochloride (13): Yield: 57%. M.p. 145–148 ^◦C. ¹H NMR

(200 MHz, DMSO-d₆):

with D₂O), 3.63 (pseudo s, 4H, 2CH₂piperazine), 3.07 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz,

DMSO-d₆): ¹³C NMR (50 MHz, DMSO-d₆):

161.9, 159.1 (d, J_C-F= 243.1 Hz), 141.9 (d, J_C-F= 9.7 Hz),

δ

7.80 (s, NH₂⁺exchange with D₂O), 7.43–6.81 (m, 6H, 4ArH and NH₂exchange

δ

128.1, 123.7 (d, J_C-F= 7.2 Hz), 120.0(d, J_C-F= 6.3 Hz), 116.4 (d, J_C-F= 22.1 Hz), 59.3 (2C), 47.8 (2C). Anal.

Calcd for C₁₁H₁₅FN₄·HCl: % C, 51.07; H, 6.23; N, 21.66. Found: C, 50.84; H, 6.35; N, 21.28.

4-(4-Fluorophenyl)piperazine-1-carboximidamide hydrochloride (14): Yield: 53%. ¹H NMR (200 MHz,

+

DMSO-d₆):

δ

7.76 (s, NH₂exchange with D₂O), 7.41–6.84 (m, 6H, 4ArH and NH₂exchange with

D₂O), 3.61 (pseudo s, 4H, 2CH₂piperazine), 3.16 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz,

DMSO-d₆): 162.4, 157.8 (d, J_C-F= 244.5 Hz), 148.1, 120.2 (d, J_C-F= 7.6 Hz; 2C), 115.6 (d, J_C-F= 23.4 Hz;

δ

2C), 52.3 (2C), 47.4 (2C). Anal calcd for C₁₁H₁₅FN₄

H, 6.38; N, 21.43.

·

HCl: % C, 51.07; H, 6.23; N, 21.66. Found: C, 50.90;

1

4-(2,3-Dichlorophenyl)piperazine-1-carboximidamide hydrochloride (15): Yield: 37%. H NMR (200 MHz,

+

DMSO-d₆):

δ

7.86 (s, NH₂exchange with D₂O), 7.69–6.98 (m, 5H, 3ArH and NH₂exchange with

163.1, 149.2, 135.1, 128.2,

HCl: % C, 42.67; H, 4.88; N, 18.10.

D₂O), 3.76–2.97 (m, 8H, 4CH₂piperazine). ¹³C NMR (50 MHz, DMSO-d₆):

126.3, 125.8, 115.5, 51.9 (2C), 46.2 (2C). Anal. Calcd for C₁₁H₁₄Cl₂N₄

Found: C, 42.27; H, 4.94; N, 17.94.

δ

·

Pharmaceuticals 2020, 13, 391

20 of 26

4-(3,4-Dichlorophenyl)piperazine-1-carboximidamide hydrochloride (16): Yield: 45%. M.p. 235–237^◦C. ¹H

NMR (200 MHz, DMSO-d₆):

7.53–7.38 (m, 3H, 1ArH and 7.45, s, NH₂⁺superimposed, exchange with

D₂O), 7.19–7.15 (m, 1ArH), 7.03–6.91 (m, 3H, 1ArH and NH₂exchange with D₂O), 3.61 (pseudo s, 4H,

2CH₂piperazine), 3.27 (pseudo s, 4H, 2CH₂piperazine). ¹³C NMR (50 MHz, DMSO-d₆):

162.6, 149.1,

HCl: % C, 42.67; H,

δ

131.9, 129.1, 120.2, 116.8, 114.7, 50.4 (2C), 46.7 (2C). Anal. Calcd for C₁₁H₁₄Cl₂N₄

4.88; N, 18.10. Found: C, 42.59; H, 4.64; N, 18.29.

·

3.1.3. Synthesis of 4-(Pyridin-2-yl)Piperazine-1-Carboximidamide Hydrochloride (10)

1-(Pyridin-2-yl)piperazine (2.65 mmol) was added to a solution of S-methylisothiourea hydrochloride

(3.95 mmol) in 10 mL of H₂O, and then refluxed for 2 h with stirring. At r.t., the reaction mixture was

alkalinized with 6N NaOH and extracted with CH₂Cl₂; the organic layer was washed with H₂O, then

dried with anhydrous Na₂SO₄and filtered. After evaporation, the oily residue was converted into the

corresponding monohydrochloride with 1N ethanolic solution of HCl. Yield: 59%. M.p. 205–208 ^◦C. ¹H

NMR (200 MHz, DMSO-d₆):

D₂O), 6.97–6.61 (m, 4H, 2H pyr. and NH₂exchange with D₂O), 3.61–3.54 (m, 8H, 4CH₂piperazine). ¹³C

NMR (50 MHz, DMSO-d₆): 162.3, 158.5, 149.6, 136.7, 115.6, 111.0, 46.7 (2C), 46.0 (2C). Anal. Calcd for

δ

8.14–7.48 (m, 4H, 2H pyr. and 7.78, s, NH₂⁺superimposed, exchange with

δ

C₁₀H₁₅N₅·HCl: % C, 49.69; H, 6.67; N, 28.97. Found: C, 49.82; H, 6.76; N, 28.60.

3.2. In Vitro Biological Tests

3.2.1. Screening of hTAAR1 Agonists by Means of Bioluminescence Resonance Energy Transfer

(BRET) Technology

BRET screening was described in detail elsewhere [40 ]. HEK-293T (ATCC) cells were transiently

с

o-transfected with plasmids encoding hTAAR1 and a cAMP BRET biosensor using Lipofectamine^®

(ThermoFisher) reagent, and then plated in 96-well plates (Corning) at 50

following day, culture medium was removed and 70 L of phosphate-buffered saline containing calcium

and magnesium was added to each well, followed by the addition of 10 L 200 mM 3-isobutyl-1-

methylxanthine solution (Sigma), and 10 L 50 M coelenterazine-h solution (Promega). All tested

compounds were dissolved in DMSO to yield 10 mM stock solutions. After 10 min incubation, either 10

of vehicle or 10 of the concentrated solution of compound to be tested (10 M final concentration) was

added. Readings were collected using a Mithras LB943 multimodal plate reader (Berthold Technologies).

The BRET signal is determined by calculating the ratio of the light emitted at 505 to 555 nm to the light

emitted at 465 to 505 nm.

×

10⁴cells per well. On the

µ

µL

×

µ

For active compounds, separate dose–response experiments were performed in order to calculate

the EC₅₀values. Curves were ﬁtted by applying non-linear regression models on GraphPad Prism

6 (GraphPad Software). Data are representative of 4 independent experiments and are expressed as

means (errors in EC₅₀are within 10%).

3.2.2. Cell Cytotoxicity Assay

Vero-76 cells (ATCC CRL 1587 Cercopithecus Aethiops) were seeded at an initial density of 4 × 10⁵cells/mL

in 24-well plates, in culture medium (Dulbecco’s Modiﬁed Eagle Medium (D-MEM) with L-glutamine,

supplemented with fetal bovine serum (FBS), 0.025 g/L kanamycin). Cell cultures were then incubated

◦

at 37 C in a humidiﬁed, 5% CO₂atmosphere in the absence or presence of serial dilutions of test

compounds. Cell viability was determined after 48–96 h at 37 ^◦C by the MTT staining method. The results

are expressed as CC₅₀, which is the concentration of compound necessary to inhibit cell growth by 50%.

Each CC₅₀value is the mean and standard deviation of at least three separate experiments performed

in duplicate.

Pharmaceuticals 2020, 13, 391

21 of 26

3.3. Molecular Modelling Studies

3.3.1. Ligand Preparation and Pharmacophore Analysis

All the compounds investigated herein through computational studies were built, parameterized

(AM1 partial charges as in the calculation method) and energy minimized within MOE [41 ], following the

same procedure we applied in our previous work [29 ,31 ]. Oxazolines were explored in terms of geometry

and conformation energy by means of the systematic Conformational Search tool implemented in MOE,

in order to be further exploited for pharmacophore analysis. Briefly, Systematic Conformational Search

generates molecular conformations by systematically rotating bonds in a molecule by discrete increments.

The purpose of Systematic Conformational Search is to generate a collection of reasonable molecular

conformations, which may or may not be at local minima. A generated conformation is rejected if it

contains two atoms whose mutual van der Waals energy exceeds a threshold (by default, 10 kcal/mol).

This ensures that the output conformations contain no conformations with heavily overlapped atoms.

Conformations generated by this method may be strained due to bonded interactions as well as some

non-bonded strain. Then, a preliminary pharmacophore model has been developed taking into account

the most potent oxazolines reported by Roche (pEC₅₀values > 7.00 M) including compounds 4b

, 5b,

9b 11b 16b 18b 20b 23b 25b 29b and 33b 37b from the whole set 1b 37b, while the remaining

–

,

–

,

–

modest analogues were considered as an external set. The final model has been calculated using the

pharmacophore search module implemented in the MOE software, starting from the alignment of

the aforementioned oxazolines onto the most potent 20b, taken as reference compound. The module

pharmacophore consensus of this software generates a set of suggested features based on the exploited

alignment of the selected oxazoline conformations. These chemical features are characterized by a position,

radius and a type expression. The relevance of any feature when based on equal scores is assessed by

secondary keys in the following order: radius, number of molecules, number of conformations, length

of the expression and alphabetical sequence. This kind of analysis is described in the literature as a

successful strategy for drug design and for virtual screening approaches. Accordingly, the computational

part of pharmacophore modeling has significantly improved in reent years thanks to the availability of

software packages, such as MOE software [42 –46 ].

3.3.2. Molecular Docking Studies

Docking calculations within the hTAAR1 receptor have been performed based on the in-house

homology modelled hTAAR1, which has already been extensively exploited by us to explore the putative

bioactive conformation and binding mode of different series of TAAR1-targeting compounds [29 ,31 ].

The reliability of our previous study, as well as of the binding site we discussed, was supported by

mutagenesis experiments performed by Reese [47 ]. Herein, we proceeded with molecular docking

studies at the same binding site, focusing on the newly synthesized compounds,

a higher potency trend within the amidine series. The reliability of the applied docking protocol was

assessed by performing preliminary docking calculations on the oxazolines 20b 26b 27b and 30b

as reference compounds, and on -phenylethylamine ( -PEA), as a pharmacological tool. On the

whole, docking studies were performed by means of the DOCK tool implemented in MOE, choosing

as a binding site the one identified in our previous studies [29 31 ]. The alpha triangle, as a placement

1, 2, 6 and 15, featuring

,

β

,

algorithm, was selected. In this case, poses are generated by the superposition of ligand atom triplets

and the triplets of receptor site points. The receptor site points are alpha sphere centers that represent

the locations of tight packing. At each iteration, a random conformation is selected. A random triplet of

ligand atoms and a random triplet of alpha sphere centers are used to determine the pose. The calculation

of the enthalpy-based Affinity ∆G scoring function allowed us to score the generated thirty poses, while

the induced fit method was applied to refine the previous poses and create the final ones. These were

rescored based on the alpha HB methodology, which is focused on H-bonding estimation.

Pharmaceuticals 2020, 13, 391

22 of 26

This Aﬃnity ∆G function estimates the enthalpic contribution to the free energy of binding using

a linear function:

∆G = C_hbf_hb+ C_ionf_ion+ C_mligf_mlig+ C_hhf_hh+ C_hpf_hp+ C_aaf_aa

where the f terms fractionally count atomic contacts of speciﬁc types and the C_iterms are coeﬃcients

that weight the term contributions to the aﬃnity estimate. The individual terms are shown below.

Subscript

Description

hb

Interactions between hydrogen bond donor–acceptor pairs. An optimistic view is taken;

for example, two hydroxyl groups are assumed to interact in the most favorable way

Ionic interactions. A Coulomb-like term is used to evaluate the interactions between charged

groups. This can contribute to or detract from binding aﬃnity

Metal ligation. Interactions between nitrogens/sulfurs and transition metals are assumed to be

metal ligation interactions

ion

mlig

hh

Hydrophobic interactions, for example, between alkane carbons. These interactions are

generally favorable

hp

aa

Interactions between hydrophobic and polar atoms. These interactions are generally unfavorable

An interaction between any two atoms. This interaction is weak and generally favorable

The Induced Fit approach allows us to maintain ﬂexible protein sidechains within the selected

binding site, which are to be included in the refinement stage. The derived docking poses were prioritized

by the score values of the lowest energy pose of the compounds docked to the protein structure,

as shown below.

S

The ﬁnal score, which is the score of the last stage of reﬁnement.

E_conf

The energy of the conformer. If there is a reﬁnement stage, this is the energy calculated at the

end of the reﬁnement

E_place

E_score1

E_score2

E_reﬁne

Score from the placement stage

Score from rescoring stages 1 and 2

Score from the reﬁnement stage, calculated to be the sum of the van der Waals electrostatics and

solvation energies, under the Generalized Born solvation model (GB/VI)

3.3.3. In Silico Evaluation of Pharmacokinetic and Toxicity Properties

The prediction of parameters related to ADMET properties was performed by means of the

Advanced Chemistry Development (ACD) Percepta platform 2015 (v14.0.0, ww w .acdlabs.com). Any

ADMET descriptor was evaluated by Percepta, relying on the training libraries implemented in the

software, which include several sets of molecules whose pharmacokinetic and toxicity behaviors have

been experimentally determined.

4. Conclusions

The present work reports on the discovery of 1-amidino-4-phenylpiperazines as very promising

hTAAR1 agonists, as a result of a combination strategy based on pharmacophore model studies and a

scaﬀold simpliﬁcation strategy for an in-house series of biguanide-based TAAR1 agonists. Interestingly,

most of them showed a potent agonism activity at hTAAR1, comparing favorably with the nanomolar

potency of the endogenous ligands and of notable compounds from La Roche. Our study has proven

to be a valid strategy, allowing for the identiﬁcation of the minimal structural requirements for an

eﬃcient hTAAR1 agonism behavior, with the amidino group as a basic moiety and a planar aromatic

ring as the hydrophobic substituent, adequately spaced through the bifunctional piperazine ring and

able to impose an adequate geometry to the molecule.

Pharmaceuticals 2020, 13, 391

23 of 26

Additionally, they exhibited low cytotoxicity values; thus, the corresponding therapeutic indices

were very high for the most eﬀective hTAAR1 agonists (15 16 and ). In other previous works,

,

2

,

9,

6

some of these compounds were proven to modulate the adrenergic system, both in the brain and in the

periphery, an activity that was observed for other classes of TAAR1 ligands thus far identiﬁed.

Therefore, the present results prompt us to explore these 1-amidino-4-arylpiperazine derivatives

at a later stage, with a view to assessing their TAAR1 selectivity proﬁle over monoaminergic GPCRs,

whose activation or downregulation might be correlated with the onset of unwanted side eﬀects.

Accordingly, this series deserves further structural improvements and more in-depth biological studies

of their mechanism of action towards the design of more eﬀective molecules endowed with well-suited

drug-like properties.

Supplementary Materials: The following are available online at http://ww w .mdpi.com /1424-8247 /13 /11 /391 /s1:

Chemical structure and biological activity of 2-aminooxazolines (1b –37b ) (Table S1); Table of scoring functions

for the selected docking poses of the discussed hTAAR1 agonists (Table S2); Docking positioning at the hTAAR1

putative binding site of the oxazoline derivatives 20b (Figure S1) and 30b (Figure S2), of the amidine derivatives

(Figure S3), and 15 (Figure S4); Analysis of the ChEMBL database depending on diﬀerent similarity thresholds

1

2

to the present 1-amidino-4-phenylpiperazines 1– 3, 6, 7, 9,

16 (Table S3); ¹H and ¹³C NMR spectra of compounds

12, 15 and 16.

Author Contributions: Conceptualization, M.T.; methodology and validation, M.T., E.C., S.P., R.R.G.; software,

E.C.; investigation, V.F., M.T., E.C., E.V.K., E.L.; resources, M.T., E.C., S.P., R.R.G.; writing—original draft

preparation, M.T., E.C.; writing—review and editing, M.T., S.P., R.R.G.; visualization, M.T.; supervision, M.T.,

E.C., S.P., R.R.G.; project administration, M.T. All authors have read and agreed to the published version of

the manuscript.

Funding: This work was ﬁnancially supported by the University of Genoa. E.V.K. and R.R.G. and BRET screening

for TAAR1 activity were supported by the Russian Science Foundation (grant number 19-75-30008).

Acknowledgments: M.T. thanks O. Gagliardo for performing the elemental analyses.

Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1.

Tonelli, M.; Cichero, E. Trace amine associated receptor 1 (TAAR1) modulators: A patent review (2010–present).

Expert Opin. Ther. Patents 2020, 30, 137–145. [CrossRef] [PubMed]

2.

Schwartz, M.D.; Canales, J.J.; Zucchi, R.; Espinoza, S.; Sukhanov, I.; Gainetdinov, R.R. Trace amine-associated

receptor 1: A multimodal therapeutic target for neuropsychiatric diseases. Expert Opin. Ther. Targets 2018, 22,

513–526. [CrossRef] [PubMed]

3.

4.

Rutigliano, G.; Accorroni, A.; Zucchi, R. The case for TAAR1 as a modulator of central nervous system

function. Front. Pharmacol. 2018, 8, 987. [CrossRef] [PubMed]

Scanlan, T.S.; Suchland, K.L.; Hart, M.E.; Chiellini, G.; Huang, Y.; Kruzich, P.J.; Frascarelli, S.; Crossley, D.A.;

Bunzow, J.R.; Ronca-Testoni, S.; et al. 3-Iodothyronamine is an endogenous and rapid-acting derivative of

thyroid hormone. Nat. Med. 2004, 10, 638–642. [CrossRef] [PubMed]

5.

6.

Gainetdinov, R.R.; Hoener, M.C.; Berry, M.D. Trace amines and their receptors. Pharmacol. Rev. 2018, 70,

549–620. [CrossRef] [PubMed]

Berry, M.D.; Gainetdinov, R.R.; Hoener, M.C.; Shahid, M. Pharmacology of human trace amine-associated

receptors: Therapeutic opportunities and challenges. Pharmacol. Ther. 2017, 180, 161–180. [CrossRef]

[PubMed]

7.

8.

Pei, Y.; Asif-Malik, A.; Canales, J.J. Trace amines and the trace amine-associated Receptor 1: Pharmacology,

neurochemistry, and clinical implications. Front. Neurosci. 2016, 10, 148. [CrossRef]

Leo, D.; Mus, L.; Espinoza, S.; Hoener, M.C.; Sotnikova, T.D.; Gainetdinov, R.R. TAAR1-mediated modulation

of presynaptic dopaminergic neurotransmission: Role of D2 dopamine autoreceptors. Neuropharmacology

2014, 81, 283–291. [CrossRef]

9.

Espinoza, S.; Ghisi, V.; Emanuele, M.; Leo, D.; Sukhanov, I.; Sotnikova, T.D.; Chieregatti, E.; Gainetdinov, R.R.

Postsynaptic D2 dopamine receptor supersensitivity in the striatum of mice lacking TAAR1. Neuropharmacology

2015, 93, 308–313. [CrossRef]

Pharmaceuticals 2020, 13, 391

24 of 26

10. Bräunig, J.; Dinter, J.; Höﬁg, C.; Paisdzior, S.; Szczepek, M.; Scheerer, P.; Rosowski, M.; Mittag, J.; Kleinau, G.;

Biebermann, H. The trace amine-associated receptor 1 agonist 3-iodothyronamine induces biased signaling

at the serotonin 1b receptor. Front. Pharmacol. 2018, 9, 222. [CrossRef]

11. Espinoza, S.; Lignani, G.; Caﬃno, L.; Maggi, S.; Sukhanov, I.; Leo, D.; Mus, L.; Emanuele, M.; Ronzitti, G.;

Harmeier, A.; et al. TAAR1 modulates cortical glutamate NMDA receptor function. Neuropsychopharmacology

2015, 40, 2217–2227. [CrossRef] [PubMed]

12. Xu, Z.; Li, Q. TAAR1 agonists. Cell Mol. Neurobiol. 2020, 40, 257–272. [CrossRef] [PubMed]

13. Cichero, E.; Tonelli, M. Targeting species-specific trace amine-associated receptor 1 ligands: To date perspective

of the rational drug design process. Future Med. Chem. 2017, 9, 1507–1527. [CrossRef] [PubMed]

14. Chiellini, G.; Nesi, G.; Digiacomo, M.; Malvasi, R.; Espinoza, S.; Sabatini, M.; Frascarelli, S.; Laurino, A.;

Cichero, E.; Macchia, M.; et al. Design, Synthesis, and evaluation of thyronamine analogues as novel potent

mouse trace amine associated receptor 1 (mTAAR1) agonists. J. Med. Chem. 2015, 58, 5096–5107. [CrossRef]

[PubMed]

15. Chiellini, G.; Nesi, G.; Sestito, S.; Chiarugi, S.; Runfola, M.; Espinoza, S.; Sabatini, M.; Bellusci, L.; Laurino, A.;

Cichero, E.; et al. Hit-to-lead optimization of mouse trace amine associated receptor 1 (mTAAR1) agonists

with a diphenylmethane-scaﬀold: Design, synthesis, and biological study. J. Med. Chem. 2016, 59, 9825–9836.

[CrossRef]

16. Regard, J.B.; Kataoka, H.; Cano, D.A.; Camerer, E.; Yin, L.; Zheng, Y.W.; Scanlan, T.S.; Hebrok, M.;

Coughlin, S.R. Probing cell type-speciﬁc functions of Gi in vivo identiﬁes GPCR regulators of insulin

secretion. J. Clin. Investig. 2007, 117, 4034–4043. [CrossRef]

17. Hoeﬁg, C.S.; Zucchi, R.; Köhrle, J. Thyronamines and derivatives: Physiological relevance, pharmacological

actions and future research directions. Thyroid 2016, 26, 1656–1673. [CrossRef]

18. Galley, G.; Beurier, A.; Décoret, G.; Goergler, A.; Hutter, R.; Mohr, S.; Pähler, A.; Schmid, P.; Türck, D.;

Unger, R.; et al. Discovery and characterization of 2-aminooxazolines as highly potent, selective, and orally

active TAAR1 agonists. ACS Med. Chem. Lett. 2015, 7, 192–197. [CrossRef]

19. Revel, F.G.; Moreau, J.L.; Gainetdinov, R.R.; Bradaia, A.; Sotnikova, T.D.; Mory, R.; Durkin, S.; Zbinden, K.G.;

Norcross, R.; Meyer, C.A.; et al. TAAR1 activation modulates monoaminergic neurotransmission, preventing

hyperdopaminergic and hypoglutamatergic activity. Proc. Natl. Acad. Sci. USA 2011, 108, 8485–8490.

[CrossRef]

20. Revel, F.G.; Moreau, J.L.; Gainetdinov, R.R.; Ferragud, A.; Velázquez-Sánchez, C.; Sotnikova, T.D.;

Morairty, S.R.; Harmeier, A.; Groebke Zbinden, K.; Norcross, R.D.; et al. Trace amine-associated receptor 1

partial agonism reveals novel paradigm for neuropsychiatric therapeutics. Biol. Psychiatry 2012, 72, 934–942.

[CrossRef]

21. Revel, F.G.; Moreau, J.L.; Pouzet, B.; Mory, R.; Bradaia, A.; Buchy, D.; Metzler, V.; Chaboz, S.; Groebke

Zbinden, K.; Galley, G.; et al. A new perspective for schizophrenia: TAAR1 agonists reveal antipsychotic- and

antidepressant-like activity, improve cognition and control body weight. Mol. Psychiatry 2013, 18, 543–556.

[CrossRef] [PubMed]

22. Hoﬀmann-La Roche, A.G.F. Combinations Comprising Atypical Antipsychotics and TAAR1 Agonists.

WO2012/016879 A1, 9 February 2012.

23. Hoﬀmann-La Roche, A.G.F. 5-Ethyl-4-Methyl-Pyrazole-3-Carboxamide Derivative Having Activity as

Agonist of TAAR. WO2017/157873 A1, 21 September 2017.

24. Michael, E.S.; Covic, L.; Kuliopulos, A. Trace amine-associated receptor 1 (TAAR1) promotes anti-diabetic

signaling in insulin-secreting cells. J. Biol. Chem. 2019, 294, 4401–4411. [CrossRef] [PubMed]

25. Raab, S.; Wang, H.; Uhles, S.; Cole, N.; Alvarez-Sanchez, R.; Künnecke, B.; Ullmer, C.; Matile, H.; Bedoucha, M.;

Norcross, R.D.; et al. Incretin-like eﬀects of small molecule trace amine-associated receptor 1 agonists.

Mol. Metab. 2016, 5, 47–56. [CrossRef] [PubMed]

26. Ferragud, A.; Howell, A.D.; Moore, C.F.; Ta, T.L.; Hoener, M.C.; Sabino, V.; Cottone, P. The trace amine-

associated receptor 1 agonist RO5256390 blocks compulsive, binge-like eating in rats. Neuropsychopharmacology

2017, 42, 1458–1470. [CrossRef] [PubMed]

Pharmaceuticals 2020, 13, 391

25 of 26

27. Dedic, N.; Jones, P.G.; Hopkins, S.C.; Lew, R.; Shao, L.; Campbell, J.E.; Spear, K.L.; Large, T.H.; Campbell, U.C.;

Hanania, T.; et al. SEP-363856, a novel psychotropic agent with a unique, non-D2 receptor mechanism of

action. J. Pharmacol. Exp. Ther. 2019, 371, 1–14. [CrossRef] [PubMed]

28. Koblan, K.S.; Kent, J.; Hopkins, S.C.; Krystal, J.H.; Cheng, H.; Goldman, R.; Loebel, A. A non-D2-

receptor-binding drug for the treatment of schizophrenia. N. Engl. J. Med. 2020, 382, 1497–1506. [CrossRef ]

29. Tonelli, M.; Espinoza, S.; Gainetdinov, R.R.; Cichero, E. Novel biguanide-based derivatives scouted as TAAR1

agonists: Synthesis, biological evaluation, ADME prediction and molecular docking studies. Eur. J. Med.

Chem. 2017, 127, 781–792. [CrossRef]

30. Cichero, E.; Espinoza, S.; Tonelli, M.; Franchini, S.; Gerasimov, A.S.; Sorbi, C.; Gainetdinov, R.R.;

Brasili, L.; Fossa, P. Homology modelling-driven study leading to the discovery of the ﬁrst mouse Trace

Amine-Associated Receptor 5 (TAAR5) antagonists. Med. Chem. Comm. 2016, 7, 353–364. [CrossRef]

31. Guariento, S.; Tonelli, M.; Espinoza, S.; Gerasimov, A.S.; Gainetdinov, R.R.; Cichero, E. Rational design,

chemical synthesis and biological evaluation of novel biguanides exploring species-speciﬁcity responsiveness

of TAAR1 agonists. Eur. J. Med. Chem. 2018, 146, 171–184. [CrossRef]

32. Guisado, O.; Martınez, S.; Pastor, J. Anovel, facilemethodologyforthesynthesisofN,N⁰-bis(tert-butoxycarbonyl)-

protected guanidines using polymer-supported carbodiimide. Tetrahedron Lett. 2002, 43, 7105–7109. [CrossRef ]

33. Guo, D.X.; Liu, Y.; Wang, N.; Hu, C.; Gong, P. Synthesis and antitumor activities of a new series of

4,5-dihydro-1H-thiochromeno[4,3-d]pyrimidine derivatives. Sci. China Chem. 2012, 55, 347–351. [CrossRef]

34. Bristol-Myers Squibb Pharma Company. Ligands for Imaging Cardiac Innervation. WO2008/83056, 10 July 2008

.

35. Protiva, M.; Rajšner, M.; Trcˇka, V.; Vaneˇcˇek, M.; Neˇmec, J.; Šediv , Z. 1-Aryl- and 1-(arylmethyl)-4-

ý

guanylpiperazines and other heterocyclic and alicyclic guanidine derivatives. Collect Czechoslov. Chem.

Commun. 1975, 40, 3904–3923. [CrossRef]

36. Ozawa, H.; Iwatsuki, K. Pharmacological properties of heterocyclic amidine derivatives. II. Pharmacological

studies of phenylguanylpiperazine derivatives. Chem. Pharm. Bull. 1968, 16, 2482–2487. [CrossRef] [PubMed]

37. Conroy, E.A.; Denton, J.J. Piperazine. III. 1-Heterocyclic-4-guanyl-, carbamyl-, and thiocarbamyl-piperazines.

J. Organic Chem. 1953, 18, 1489–1491. [CrossRef]

38. Ozawa, H.; Iwatsuki, K. Pharmacological properties of heterocyclic amidine derivatives. IV. The action of

1-phenyl-4-guanylpiperazine sulfate (PGP) on the concentration of norepinephrine in brain and heart of rat.

Yakugaku Zasshi 1971, 12, 1381–1383. [CrossRef]

39. Waterbeemd, H.; Gifford, E. ADMET in silico modelling: Towards prediction paradise? Nat. Rev. Drug Discov.

2003, 2, 192–204. [CrossRef]

40. Barak, L.S.; Salahpour, A.; Zhang, X.; Masri, B.; Sotnikova, T.D.; Ramsey, A.J.; Violin, J.D.; Lefkowitz, R.J.;

Caron, M.G.; Gainetdinov, R.R. Pharmacological characterization of membrane-expressed human trace

amine-associated receptor 1 (TAAR1) by a bioluminescence resonance energy transfer cAMP Biosensor.

Mol. Pharmacol. 2008, 74, 585–594. [CrossRef]

41. MOE: Chemical Computing Group Inc. Montreal. H3A 2R7 Canada. Available online: http://ww w .chemcomp.

com / (accessed on 14 November 2020).

42. Wolber, G.; Seidel, T.; Bendix, F.; Langer, T. Molecule-pharmacophore superpositioning and pattern matching

in computational drug design. Drug Discov. Today 2008, 13, 23–29. [CrossRef]

43. Khalid, S.; Hanif, R.; Jabeen, I.; Mansoor, Q.; Ismail, M. Pharmacophore modeling for identiﬁcation of

anti-IGF-1R drugs and in-vitro validation of fulvestrant as a potential inhibitor. PLoS ONE 2018, 13, e0196312.

[CrossRef]

44. Haidar, S.; Bouaziz, Z.; Marminon, C.; Laitinen, T.; Poso, A.; Le Borgne, M.; Jose, J. Development of

pharmacophore model for indeno[1,2-b]indoles as human protein kinase CK2 inhibitors and database mining.

Pharmaceuticals 2017, 10, 8. [CrossRef]

45. Cichero, E.; Menozzi, G.; Spallarossa, A.; Mosti, L.; Fossa, P. Exploring the binding features of rimonabant

analogues and acyclic CB1 antagonists: Docking studies and QSAR analysis. J. Mol. Model. 2008, 14,

1131–1145. [CrossRef] [PubMed]

Pharmaceuticals 2020, 13, 391

26 of 26

46. Tonelli, M.; Naesens, L.; Gazzarrini, S.; Santucci, M.; Cichero, E.; Tasso, B.; Moroni, A.; Costi, M.P.;

Loddo, R. Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with activity

against inﬂuenza virus and respiratory syncytial virus. Eur. J. Med. Chem. 2017, 135, 467–478. [CrossRef]

[PubMed]

47. Reese, E.A.; Norimatsu, Y.; Grandy, M.S.; Suchland, K.L.; Bunzow, J.R.; Grandy, D.K. Exploring

the determinants of trace amine-associated receptor 1’s functional selectivity for the stereoisomers of

amphetamine and methamphetamine. J. Med. Chem. 2014, 57, 378–390. [CrossRef] [PubMed]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional

aﬃliations.

©

2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access

article distributed under the terms and conditions of the Creative Commons Attribution

(CC BY) license (http://creativecommons.org /licenses /by /4.0 /).

Article Doi

Novel 1-amidino-4-phenylpiperazines as potent agonists at human taar1 receptor: Rational design, synthesis, biological evaluation and molecular docking studies

DOI: 10.3390/ph13110391

Source and publish data:

Authors:

Article abstract of DOI:10.3390/ph13110391

Full text of DOI:10.3390/ph13110391

Products guided by the article

R&D Labs maybe for 1361118-53-7

Relevant to this article

Hot Product